fix: thread app config through client and sync providers

.agent/skills/smoke-test/SKILL.md

@@ -59,7 +59,7 @@ smoke-test/
 2. **Check pnpm** - Package manager
 3. **Check uv** - Python package manager
 4. **Check nginx** - Reverse proxy
-5. **Check required ports** - Confirm that ports 2026, 3000, and 8001 are not occupied
+5. **Check required ports** - Confirm that ports 2026, 3000, 8001, and 2024 are not occupied
 
 **Docker mode environment check** (if Docker is selected):
 1. **Check whether Docker is installed** - Run `docker --version`
@@ -93,17 +93,17 @@ smoke-test/
 ### Phase 5: Service Health Check
 
 **Local mode health check**:
-1. **Check process status** - Confirm that Gateway, Frontend, and Nginx processes are all running
+1. **Check process status** - Confirm that LangGraph, Gateway, Frontend, and Nginx processes are all running
 2. **Check frontend service** - Visit `http://localhost:2026` and verify that the page loads
 3. **Check API Gateway** - Verify the `http://localhost:2026/health` endpoint
-4. **Check LangGraph-compatible API** - Verify the `/api/langgraph/*` route exposed by Gateway
+4. **Check LangGraph service** - Verify the availability of relevant endpoints
 5. **Frontend route smoke check** - Run `bash .agent/skills/smoke-test/scripts/frontend_check.sh` to verify key routes under `/workspace`
 
 **Docker mode health check** (when using Docker):
 1. **Check container status** - Run `docker ps` and confirm that all containers are running
 2. **Check frontend service** - Visit `http://localhost:2026` and verify that the page loads
 3. **Check API Gateway** - Verify the `http://localhost:2026/health` endpoint
-4. **Check LangGraph-compatible API** - Verify the `/api/langgraph/*` route exposed by Gateway
+4. **Check LangGraph service** - Verify the availability of relevant endpoints
 5. **Frontend route smoke check** - Run `bash .agent/skills/smoke-test/scripts/frontend_check.sh` to verify key routes under `/workspace`
 
 ### Optional Functional Verification
@@ -135,7 +135,7 @@ smoke-test/
 
 The following warnings can appear during smoke testing and do not block a successful result:
 - Feishu/Lark SSL errors in Gateway logs (certificate verification failure) can be ignored if that channel is not enabled
-- Warnings in Gateway logs about missing methods in the custom checkpointer, such as `adelete_for_runs` or `aprune`, do not affect the core functionality
+- Warnings in LangGraph logs about missing methods in the custom checkpointer, such as `adelete_for_runs` or `aprune`, do not affect the core functionality
 
 ## Key Tools
 

.agent/skills/smoke-test/references/SOP.md

@@ -138,6 +138,7 @@ This document describes the detailed operating steps for each phase of the DeerF
    lsof -i :2026  # Main port
    lsof -i :3000  # Frontend
    lsof -i :8001  # Gateway
+   lsof -i :2024  # LangGraph
    ```
 
 **Success Criteria**: All ports are free, or they are occupied only by DeerFlow-related processes.
@@ -257,7 +258,7 @@ This document describes the detailed operating steps for each phase of the DeerF
 **Steps**:
 1. Run `make dev-daemon` (background mode)
 
-**Description**: This command starts all services (Gateway embedded runtime, Frontend, Nginx).
+**Description**: This command starts all services (LangGraph, Gateway, Frontend, Nginx).
 
 **Notes**:
 - `make dev` runs in the foreground and stops with Ctrl+C
@@ -271,6 +272,7 @@ This document describes the detailed operating steps for each phase of the DeerF
 **Steps**:
 1. Wait 90-120 seconds for all services to start completely
 2. You can monitor startup progress by checking these log files:
+   - `logs/langgraph.log`
    - `logs/gateway.log`
    - `logs/frontend.log`
    - `logs/nginx.log`
@@ -314,10 +316,11 @@ This document describes the detailed operating steps for each phase of the DeerF
 **Steps**:
 1. Run the following command to check processes:
    ```bash
-   ps aux | grep -E "(uvicorn|next|nginx)" | grep -v grep
+   ps aux | grep -E "(langgraph|uvicorn|next|nginx)" | grep -v grep
    ```
 
 **Success Criteria**: Confirm that the following processes are running:
+- LangGraph (`langgraph dev`)
 - Gateway (`uvicorn app.gateway.app:app`)
 - Frontend (`next dev` or `next start`)
 - Nginx (`nginx`)
@@ -353,11 +356,10 @@ curl http://localhost:2026/health
 
 ---
 
-#### 5.1.4 Check LangGraph-compatible API
+#### 5.1.4 Check LangGraph Service
 
 **Steps**:
-1. Visit `http://localhost:2026/api/langgraph/assistants/lead_agent` to verify Gateway's LangGraph-compatible API route is reachable.
+1. Visit relevant LangGraph endpoints to verify availability
-2. A `401` response is acceptable when authentication is enabled and no session cookie is provided.
 
 ---
 
@@ -371,6 +373,7 @@ curl http://localhost:2026/health
    - `deer-flow-nginx`
    - `deer-flow-frontend`
    - `deer-flow-gateway`
+   - `deer-flow-langgraph` (if not in gateway mode)
 
 ---
 
@@ -403,11 +406,10 @@ curl http://localhost:2026/health
 
 ---
 
-#### 5.2.4 Check LangGraph-compatible API
+#### 5.2.4 Check LangGraph Service
 
 **Steps**:
-1. Visit `http://localhost:2026/api/langgraph/assistants/lead_agent` to verify Gateway's LangGraph-compatible API route is reachable.
+1. Visit relevant LangGraph endpoints to verify availability
-2. A `401` response is acceptable when authentication is enabled and no session cookie is provided.
 
 ---
 

.agent/skills/smoke-test/references/troubleshooting.md

@@ -254,6 +254,7 @@ Processes exit quickly after running `make dev-daemon`.
 **Solutions**:
 1. Check log files:
    ```bash
+   tail -f logs/langgraph.log
    tail -f logs/gateway.log
    tail -f logs/frontend.log
    tail -f logs/nginx.log
@@ -366,7 +367,24 @@ Errors appear in `gateway.log`.
    uv sync
    ```
 
-4. Confirm that the Gateway process is running normally.
+4. Confirm that the LangGraph service is running normally (if not in gateway mode)
+
+---
+
+### Issue: LangGraph Fails to Start
+
+**Symptoms**:
+Errors appear in `langgraph.log`.
+
+**Solutions**:
+1. Check LangGraph logs:
+   ```bash
+   tail -f logs/langgraph.log
+   ```
+
+2. Check config.yaml
+3. Check whether Python dependencies are complete
+4. Confirm that port 2024 is not occupied
 
 ---
 
@@ -501,7 +519,7 @@ Accessing `/health` returns an error or times out.
 
 2. Confirm that config.yaml exists and has valid formatting
 3. Check whether Python dependencies are complete
-4. Confirm that the Gateway process is running normally.
+4. Confirm that the LangGraph service is running normally
 
 **Solutions** (Docker mode):
 1. Check gateway container logs:
@@ -511,7 +529,7 @@ Accessing `/health` returns an error or times out.
 
 2. Confirm that config.yaml is mounted correctly
 3. Check whether Python dependencies are complete
-4. Confirm that the Gateway process is running normally.
+4. Confirm that the LangGraph service is running normally
 
 ---
 
@@ -521,7 +539,7 @@ Accessing `/health` returns an error or times out.
 
 #### View All Service Processes
 ```bash
-ps aux | grep -E "(uvicorn|next|nginx)" | grep -v grep
+ps aux | grep -E "(langgraph|uvicorn|next|nginx)" | grep -v grep
 ```
 
 #### View Service Logs
@@ -530,6 +548,7 @@ ps aux | grep -E "(uvicorn|next|nginx)" | grep -v grep
 tail -f logs/*.log
 
 # View specific service logs
+tail -f logs/langgraph.log
 tail -f logs/gateway.log
 tail -f logs/frontend.log
 tail -f logs/nginx.log

.agent/skills/smoke-test/scripts/check_local_env.sh

@@ -65,7 +65,7 @@ if ! command -v lsof >/dev/null 2>&1; then
     echo "  Install lsof and rerun this check"
     all_passed=false
 else
-    for port in 2026 3000 8001; do
+    for port in 2026 3000 8001 2024; do
         if lsof -i :$port >/dev/null 2>&1; then
             echo "⚠  Port $port is already in use:"
             lsof -i :$port | head -2

.agent/skills/smoke-test/scripts/deploy_local.sh

@@ -54,6 +54,7 @@ echo "=========================================="
 echo ""
 echo "🌐 Access URL: http://localhost:2026"
 echo "📋 View logs:"
+echo "   - logs/langgraph.log"
 echo "   - logs/gateway.log"
 echo "   - logs/frontend.log"
 echo "   - logs/nginx.log"

.agent/skills/smoke-test/scripts/health_check.sh

@@ -76,11 +76,12 @@ if [ "$mode" = "docker" ]; then
         all_passed=false
     fi
 else
-    summary_hint="logs/{gateway,frontend,nginx}.log"
+    summary_hint="logs/{langgraph,gateway,frontend,nginx}.log"
     print_step "1. Checking local service ports..."
     check_listen_port "Nginx" 2026
     check_listen_port "Frontend" 3000
     check_listen_port "Gateway" 8001
+    check_listen_port "LangGraph" 2024
 fi
 echo ""
 
@@ -103,8 +104,8 @@ else
 fi
 echo ""
 
-echo "5. Checking LangGraph-compatible Gateway API..."
+echo "5. Checking LangGraph service..."
-check_http_status "LangGraph-compatible Gateway API" "http://localhost:2026/api/langgraph/assistants/lead_agent" "200|401"
+check_http_status "LangGraph service" "http://localhost:2024/" "200|301|302|307|308|404"
 echo ""
 
 echo "=========================================="

.agent/skills/smoke-test/templates/report.docker.template.md

@@ -78,7 +78,7 @@
 - [x] Container status - {{status_containers}}
 - [x] Frontend service - {{status_frontend}}
 - [x] API Gateway - {{status_api_gateway}}
-- [x] LangGraph-compatible Gateway API - {{status_langgraph}}
+- [x] LangGraph service - {{status_langgraph}}
 
 **Phase Status**: {{stage5_status}}
 
@@ -147,6 +147,7 @@ Commit Message: {{git_commit_message}}
 | deer-flow-nginx | {{nginx_status}} | {{nginx_uptime}} |
 | deer-flow-frontend | {{frontend_status}} | {{frontend_uptime}} |
 | deer-flow-gateway | {{gateway_status}} | {{gateway_uptime}} |
+| deer-flow-langgraph | {{langgraph_status}} | {{langgraph_uptime}} |
 
 ---
 

.agent/skills/smoke-test/templates/report.local.template.md

@@ -80,7 +80,7 @@
 - [x] Process status - {{status_processes}}
 - [x] Frontend service - {{status_frontend}}
 - [x] API Gateway - {{status_api_gateway}}
-- [x] LangGraph-compatible Gateway API - {{status_langgraph}}
+- [x] LangGraph service - {{status_langgraph}}
 
 **Phase Status**: {{stage5_status}}
 
@@ -152,7 +152,7 @@ Commit Message: {{git_commit_message}}
 | Nginx | {{nginx_status}} | {{nginx_endpoint}} |
 | Frontend | {{frontend_status}} | {{frontend_endpoint}} |
 | Gateway | {{gateway_status}} | {{gateway_endpoint}} |
-| Gateway LangGraph API | {{langgraph_status}} | {{langgraph_endpoint}} |
+| LangGraph | {{langgraph_status}} | {{langgraph_endpoint}} |
 
 ---
 
@@ -166,7 +166,7 @@ Commit Message: {{git_commit_message}}
 
 ### If the Test Fails
 1. [ ] Review references/troubleshooting.md for common solutions
-2. [ ] Check local logs: `logs/{gateway,frontend,nginx}.log`
+2. [ ] Check local logs: `logs/{langgraph,gateway,frontend,nginx}.log`
 3. [ ] Verify configuration file format and content
 4. [ ] If needed, fully reset the environment: `make stop && make clean && make install && make dev-daemon`
 

.env.example

@@ -1,6 +1,3 @@
-# Serper API Key (Google Search) - https://serper.dev
-SERPER_API_KEY=your-serper-api-key
-
 # TAVILY API Key
 TAVILY_API_KEY=your-tavily-api-key
 
@@ -9,9 +6,8 @@ JINA_API_KEY=your-jina-api-key
 
 # InfoQuest API Key
 INFOQUEST_API_KEY=your-infoquest-api-key
-# Browser CORS allowlist for split-origin or port-forwarded deployments (comma-separated exact origins).
+# CORS Origins (comma-separated) - e.g., http://localhost:3000,http://localhost:3001
-# Leave unset when using the unified nginx endpoint, e.g. http://localhost:2026.
+# CORS_ORIGINS=http://localhost:3000
-# GATEWAY_CORS_ORIGINS=http://localhost:3000,http://127.0.0.1:3000
 
 # Optional:
 # FIRECRAWL_API_KEY=your-firecrawl-api-key
@@ -21,7 +17,6 @@ INFOQUEST_API_KEY=your-infoquest-api-key
 # DEEPSEEK_API_KEY=your-deepseek-api-key
 # NOVITA_API_KEY=your-novita-api-key  # OpenAI-compatible, see https://novita.ai
 # MINIMAX_API_KEY=your-minimax-api-key  # OpenAI-compatible, see https://platform.minimax.io
-# STEPFUN_API_KEY=your-stepfun-api-key  # OpenAI-compatible, see https://platform.stepfun.com
 # VLLM_API_KEY=your-vllm-api-key  # OpenAI-compatible
 # FEISHU_APP_ID=your-feishu-app-id
 # FEISHU_APP_SECRET=your-feishu-app-secret
@@ -50,19 +45,3 @@ INFOQUEST_API_KEY=your-infoquest-api-key
 
 # Set to "false" to disable Swagger UI, ReDoc, and OpenAPI schema in production
 # GATEWAY_ENABLE_DOCS=false
-
-# Shared internal Gateway auth token for multi-worker deployments.
-# `make up` generates and persists this automatically; set it manually only
-# when you run Gateway workers outside the bundled deploy script.
-# DEER_FLOW_INTERNAL_AUTH_TOKEN=your-shared-internal-token
-
-# ── Frontend SSR → Gateway wiring ─────────────────────────────────────────────
-# The Next.js server uses these to reach the Gateway during SSR (auth checks,
-# /api/* rewrites). They default to localhost values that match `make dev` and
-# `make start`, so most local users do not need to set them.
-#
-# Override only when the Gateway is not on localhost:8001 (e.g. when the
-# frontend and gateway run on different hosts, in containers with a service
-# alias, or behind a different port). docker-compose already sets these.
-# DEER_FLOW_INTERNAL_GATEWAY_BASE_URL=http://localhost:8001
-# DEER_FLOW_TRUSTED_ORIGINS=http://localhost:3000,http://localhost:2026

.github/ISSUE_TEMPLATE/bug-report.yml

@@ -1,159 +0,0 @@
-name: 🐛 Bug report
-description: Report something that isn't working so maintainers can reproduce and fix it.
-title: "[bug] "
-labels: ["bug"]
-body:
-  - type: markdown
-    attributes:
-      value: |
-        Thanks for taking the time to file a bug. A clear, reproducible report is the
-        single biggest factor in how fast it gets fixed.
-
-        Please fill in every required field — especially **reproduction steps** and **logs**.
-
-  - type: checkboxes
-    id: preflight
-    attributes:
-      label: Before you start
-      options:
-        - label: I searched [existing issues](https://github.com/bytedance/deer-flow/issues?q=is%3Aissue) and this is not a duplicate.
-          required: true
-        - label: I can reproduce this on the latest `main`.
-          required: false
-
-  - type: input
-    id: summary
-    attributes:
-      label: Problem summary
-      description: One sentence describing the bug.
-      placeholder: e.g. make dev fails to start the gateway service
-    validations:
-      required: true
-
-  - type: dropdown
-    id: areas
-    attributes:
-      label: Affected area(s)
-      description: Which part of DeerFlow does this touch? Select all that apply.
-      multiple: true
-      options:
-        - Frontend (UI / Next.js)
-        - Backend API (gateway / endpoints / SSE)
-        - Agents / LangGraph (graph, prompts, langgraph.json)
-        - Sandbox / Docker
-        - Skills
-        - MCP
-        - Config / setup (make, config.yaml, env)
-        - Docs
-        - Not sure
-    validations:
-      required: true
-
-  - type: textarea
-    id: actual
-    attributes:
-      label: What happened?
-      description: The actual behavior. Include the key error lines verbatim.
-      placeholder: When I do X, I expected Y but I got Z.
-    validations:
-      required: true
-
-  - type: textarea
-    id: expected
-    attributes:
-      label: Expected behavior
-      placeholder: What did you expect to happen instead?
-    validations:
-      required: true
-
-  - type: textarea
-    id: reproduce
-    attributes:
-      label: Steps to reproduce
-      description: Exact commands and sequence. Minimal steps that reliably reproduce the problem.
-      placeholder: |
-        1. make check
-        2. make install
-        3. make dev
-        4. ...
-    validations:
-      required: true
-
-  - type: textarea
-    id: logs
-    attributes:
-      label: Relevant logs
-      description: Paste key lines from logs (for example `logs/gateway.log`, `logs/frontend.log`). Redact secrets.
-      render: shell
-    validations:
-      required: true
-
-  - type: dropdown
-    id: run_mode
-    attributes:
-      label: How are you running DeerFlow?
-      options:
-        - Local (make dev)
-        - Docker (make docker-start)
-        - CI
-        - Other
-    validations:
-      required: true
-
-  - type: dropdown
-    id: os
-    attributes:
-      label: Operating system
-      options:
-        - macOS
-        - Linux
-        - Windows
-        - Other
-    validations:
-      required: true
-
-  - type: input
-    id: platform_details
-    attributes:
-      label: Platform details
-      description: Architecture and shell, if relevant.
-      placeholder: e.g. arm64, zsh
-
-  - type: input
-    id: python_version
-    attributes:
-      label: Python version
-      placeholder: e.g. Python 3.12.9
-
-  - type: input
-    id: node_version
-    attributes:
-      label: Node.js version
-      placeholder: e.g. v22.11.0
-
-  - type: input
-    id: pnpm_version
-    attributes:
-      label: pnpm version
-      placeholder: e.g. 10.26.2
-
-  - type: input
-    id: uv_version
-    attributes:
-      label: uv version
-      placeholder: e.g. 0.7.20
-
-  - type: textarea
-    id: git_info
-    attributes:
-      label: Git state
-      description: Output of `git branch --show-current` and the latest commit SHA.
-      placeholder: |
-        branch: feature/my-branch
-        commit: abcdef1
-
-  - type: textarea
-    id: additional
-    attributes:
-      label: Additional context
-      description: Screenshots, related issues, config snippets (redacted), or anything else that helps triage.

.github/ISSUE_TEMPLATE/config.yml

@@ -1,11 +0,0 @@
-blank_issues_enabled: false
-contact_links:
-  - name: 💬 Questions & usage help
-    url: https://github.com/bytedance/deer-flow/discussions/categories/q-a
-    about: "How do I use X? Why does Y behave like that? Ask in Discussions — it gets answered faster and stays searchable."
-  - name: 💡 Ideas & proposals
-    url: https://github.com/bytedance/deer-flow/discussions/categories/ideas
-    about: Have a half-formed idea? Float it in Discussions before opening a formal feature request.
-  - name: 🔒 Report a security vulnerability
-    url: https://github.com/bytedance/deer-flow/security/policy
-    about: Do not open a public issue for security problems. Follow the security policy instead.

.github/ISSUE_TEMPLATE/feature-request.yml

@@ -1,67 +0,0 @@
-name: 💡 Feature request
-description: Propose a new capability or an improvement to an existing one.
-title: "[feat] "
-labels: ["enhancement"]
-body:
-  - type: markdown
-    attributes:
-      value: |
-        Thanks for the suggestion. For non-trivial features, please open a
-        [Discussion](https://github.com/bytedance/deer-flow/discussions/categories/ideas)
-        first to align on scope before writing code.
-
-  - type: checkboxes
-    id: preflight
-    attributes:
-      label: Before you start
-      options:
-        - label: I searched [existing issues](https://github.com/bytedance/deer-flow/issues?q=is%3Aissue) and this is not a duplicate.
-          required: true
-
-  - type: textarea
-    id: problem
-    attributes:
-      label: Problem / motivation
-      description: What problem does this solve? What is painful today, or what does it unblock?
-      placeholder: "I'm always frustrated when ..."
-    validations:
-      required: true
-
-  - type: textarea
-    id: solution
-    attributes:
-      label: Proposed solution
-      description: Describe the change from a user's / caller's perspective.
-    validations:
-      required: true
-
-  - type: dropdown
-    id: areas
-    attributes:
-      label: Affected area(s)
-      description: Which part of DeerFlow would this touch? Select all that apply.
-      multiple: true
-      options:
-        - Frontend (UI / Next.js)
-        - Backend API (gateway / endpoints / SSE)
-        - Agents / LangGraph (graph, prompts, langgraph.json)
-        - Sandbox / Docker
-        - Skills
-        - MCP
-        - Config / setup
-        - Docs
-        - Not sure
-    validations:
-      required: true
-
-  - type: textarea
-    id: alternatives
-    attributes:
-      label: Alternatives considered
-      description: Other approaches you weighed and why you discarded them.
-
-  - type: textarea
-    id: additional
-    attributes:
-      label: Additional context
-      description: Mockups, links, related issues, or anything else that helps.

.github/ISSUE_TEMPLATE/runtime-information.yml

@@ -0,0 +1,128 @@
+name: Runtime Information
+description: Report runtime/environment details to help reproduce an issue.
+title: "[runtime] "
+labels:
+  - needs-triage
+body:
+  - type: markdown
+    attributes:
+      value: |
+        Thanks for sharing runtime details.
+        Complete this form so maintainers can quickly reproduce and diagnose the problem.
+
+  - type: input
+    id: summary
+    attributes:
+      label: Problem summary
+      description: Short summary of the issue.
+      placeholder: e.g. make dev fails to start gateway service
+    validations:
+      required: true
+
+  - type: textarea
+    id: expected
+    attributes:
+      label: Expected behavior
+      placeholder: What did you expect to happen?
+    validations:
+      required: true
+
+  - type: textarea
+    id: actual
+    attributes:
+      label: Actual behavior
+      placeholder: What happened instead? Include key error lines.
+    validations:
+      required: true
+
+  - type: dropdown
+    id: os
+    attributes:
+      label: Operating system
+      options:
+        - macOS
+        - Linux
+        - Windows
+        - Other
+    validations:
+      required: true
+
+  - type: input
+    id: platform_details
+    attributes:
+      label: Platform details
+      description: Add architecture and shell if relevant.
+      placeholder: e.g. arm64, zsh
+
+  - type: input
+    id: python_version
+    attributes:
+      label: Python version
+      placeholder: e.g. Python 3.12.9
+
+  - type: input
+    id: node_version
+    attributes:
+      label: Node.js version
+      placeholder: e.g. v23.11.0
+
+  - type: input
+    id: pnpm_version
+    attributes:
+      label: pnpm version
+      placeholder: e.g. 10.26.2
+
+  - type: input
+    id: uv_version
+    attributes:
+      label: uv version
+      placeholder: e.g. 0.7.20
+
+  - type: dropdown
+    id: run_mode
+    attributes:
+      label: How are you running DeerFlow?
+      options:
+        - Local (make dev)
+        - Docker (make docker-dev)
+        - CI
+        - Other
+    validations:
+      required: true
+
+  - type: textarea
+    id: reproduce
+    attributes:
+      label: Reproduction steps
+      description: Provide exact commands and sequence.
+      placeholder: |
+        1. make check
+        2. make install
+        3. make dev
+        4. ...
+    validations:
+      required: true
+
+  - type: textarea
+    id: logs
+    attributes:
+      label: Relevant logs
+      description: Paste key lines from logs (for example logs/gateway.log, logs/frontend.log).
+      render: shell
+    validations:
+      required: true
+
+  - type: textarea
+    id: git_info
+    attributes:
+      label: Git state
+      description: Share output of git branch and latest commit SHA.
+      placeholder: |
+        branch: feature/my-branch
+        commit: abcdef1
+
+  - type: textarea
+    id: additional
+    attributes:
+      label: Additional context
+      description: Add anything else that might help triage.

.github/labels.yml

@@ -1,119 +0,0 @@
-# Declarative label source of truth for DeerFlow.
-#
-# This file is the single source of truth for repository labels used by the
-# auto-labeling workflows (.github/workflows/pr-labeler.yml, pr-triage.yml,
-# issue-triage.yml). Auto-labelers can only apply labels that already exist,
-# so every label referenced by a workflow MUST be declared here.
-#
-# Apply with:  uv run --with pyyaml python scripts/sync_labels.py [--repo OWNER/NAME]
-# CI keeps it in sync via .github/workflows/label-sync.yml (runs on changes here).
-#
-# Sync is additive/update-only: it creates or updates the labels listed below
-# and never deletes labels that are not listed.
-#
-# Color = 6-digit hex without the leading '#'.
-
-labels:
-  # ── Type ─────────────────────────────────────────────────────────────────
-  # Mostly GitHub defaults; declared here so colors/descriptions stay stable
-  # and so issue templates can rely on them existing.
-  - name: bug
-    color: d73a4a
-    description: Something isn't working
-  - name: enhancement
-    color: a2eeef
-    description: New feature or request
-  - name: documentation
-    color: 0075ca
-    description: Improvements or additions to documentation
-  - name: question
-    color: d876e3
-    description: Further information is requested
-
-  # ── Area (auto, by changed paths — see .github/labeler.yml) ───────────────
-  # Mirrors the "Surface area" section of the pull request template.
-  - name: "area:frontend"
-    color: c5def5
-    description: Next.js frontend under frontend/
-  - name: "area:backend"
-    color: c5def5
-    description: Gateway / runtime / core backend under backend/
-  - name: "area:agents"
-    color: c5def5
-    description: Agents, subagents, graph wiring, prompts, langgraph.json
-  - name: "area:sandbox"
-    color: c5def5
-    description: Sandboxed execution and docker/
-  - name: "area:skills"
-    color: c5def5
-    description: Skills under skills/ or the skills harness
-  - name: "area:mcp"
-    color: c5def5
-    description: Model Context Protocol integration
-  - name: "area:ci"
-    color: c5def5
-    description: GitHub Actions, CI config, repo tooling
-  - name: "area:docs"
-    color: c5def5
-    description: Documentation and Markdown only
-  - name: "area:deps"
-    color: c5def5
-    description: Dependency manifests / lockfiles
-
-  # ── Size (auto, by additions + deletions — see pr-triage.yml) ─────────────
-  - name: "size/XS"
-    color: "009900"
-    description: PR changes < 20 lines
-  - name: "size/S"
-    color: 77bb00
-    description: PR changes 20-100 lines
-  - name: "size/M"
-    color: eebb00
-    description: PR changes 100-300 lines
-  - name: "size/L"
-    color: ee9900
-    description: PR changes 300-700 lines
-  - name: "size/XL"
-    color: ee5500
-    description: PR changes 700+ lines
-
-  # ── Risk (auto, by changed paths — see pr-triage.yml) ─────────────────────
-  - name: "risk:low"
-    color: 0e8a16
-    description: "Low risk: docs / i18n / assets only"
-  - name: "risk:medium"
-    color: fbca04
-    description: "Medium risk: regular code changes"
-  - name: "risk:high"
-    color: b60205
-    description: "High risk: backend API, agents, sandbox, auth, deps, CI"
-
-  # ── Priority (manual) ─────────────────────────────────────────────────────
-  - name: P0
-    color: b60205
-    description: Critical priority
-  - name: P1
-    color: d93f0b
-    description: Major priority
-  - name: P2
-    color: e99695
-    description: Normal priority
-
-  # ── Status (auto + manual) ────────────────────────────────────────────────
-  - name: needs-triage
-    color: fef2c0
-    description: Awaiting maintainer triage
-  - name: needs-validation
-    color: d4c5f9
-    description: Touches front/back contract surface; needs real-path validation
-  - name: skip-validation
-    color: cccccc
-    description: "Maintainer override: do not auto-add needs-validation on this PR"
-  - name: reviewing
-    color: 5319e7
-    description: A maintainer is reviewing this PR
-
-  # ── Contributor ───────────────────────────────────────────────────────────
-  - name: first-time-contributor
-    color: c2e0c6
-    description: First contribution to this repository — be welcoming

.github/pull_request_template.md

@@ -1,75 +0,0 @@
-<!-- Reference a related issue with #123. Use Fixes / Closes / Resolves to
-     auto-close it on merge. Delete this line if the PR doesn't reference an issue. -->
-Fixes #
-
-## Why
-
-<!-- Why are you opening this PR? Cover two things:
-       - The trigger — what made you write this? A bug you hit, a feature you need,
-         tech debt, or a prod issue?
-       - The pain being addressed — user-facing problem, or what it unblocks.
-     For non-trivial features, please open an issue/discussion first to align on
-     scope before writing code. -->
-
-
-## What changed
-
-<!-- Describe the change from a user's / caller's perspective, not as a code diff. e.g.:
-       - "Settings now has a 'Custom endpoint' field, off by default"
-       - "Backend /api/chat gains a `stream` flag, defaults to false"
-       - "Default model changed from X to Y — existing users notice on first run" -->
-
-
-## Surface area
-
-<!-- Check every box that applies. Reviewers use this to scope the review. -->
-
-- [ ] **Frontend UI** — page / component / setting / interaction under `frontend/`
-- [ ] **Backend API** — endpoint / SSE event / request-response shape under `backend/app`
-- [ ] **Agents / LangGraph** — agent node, graph wiring, `langgraph.json`, or prompt change
-- [ ] **Sandbox** — `docker/` or sandboxed execution
-- [ ] **Skills** — change under `skills/`
-- [ ] **Dependencies** — new/upgraded entry in `backend/pyproject.toml` or `frontend/package.json` (say what it buys us)
-- [ ] **Default behavior change** — changes existing behavior without the user opting in (default model, default setting, data shape)
-- [ ] **Docs / tests / CI only** — no runtime behavior change
-
-
-## Screenshots / Recording
-
-<!-- If you checked "Frontend UI", attach screenshots showing the entry point —
-     where users discover the change — not just the feature in isolation.
-     Before/after is best for behavior changes. Short GIFs welcome. -->
-
-
-## Bug fix verification
-
-<!-- Skip (delete) this section if this PR is not a bug fix.
-
-     Bugs should be encoded as a failing test that goes red before the fix.
-     Confirm:
-       - Test path that reproduces the bug:
-       - Did it go red on `main` and green on this branch? (yes / no)
-       - If a red test wasn't cheap to write, explain why and what you did instead. -->
-
-
-## Validation
-
-<!-- What you actually ran. Run at least the checks for the area you changed:
-       Backend:   cd backend  && make lint && make test
-       Frontend:  cd frontend && pnpm format && pnpm lint && pnpm typecheck && BETTER_AUTH_SECRET=local-dev-secret pnpm build && make test
-       Frontend E2E (if you touched frontend/): cd frontend && make test-e2e -->
-
-
-## AI assistance
-
-<!-- DeerFlow is an AI project — most PRs here use AI coding tools, and that's
-     welcome. Disclosing it just helps reviewers calibrate how closely to read the
-     diff. Please fill all three; don't delete the section. -->
-
-**Tool(s) used:** <!-- e.g. Claude Code, Cursor, GitHub Copilot, Codex, Windsurf, or "none" -->
-
-**How you used it:** <!-- e.g. "generated the module from a spec", "autocomplete only",
-     "AI wrote tests, I wrote the impl". A prompt or conversation link is great too. -->
-
-- [ ] I've read and understand every line of this change and take responsibility for it — it's not unreviewed AI output.
-

.github/workflows/backend-blocking-io-tests.yml

@@ -1,46 +0,0 @@
-name: Backend Blocking IO
-
-on:
-  push:
-    branches: ["main"]
-    paths:
-      - "backend/**"
-      - ".github/workflows/backend-blocking-io-tests.yml"
-  pull_request:
-    types: [opened, synchronize, reopened, ready_for_review]
-    paths:
-      - "backend/**"
-      - ".github/workflows/backend-blocking-io-tests.yml"
-
-concurrency:
-  group: blocking-io-${{ github.event.pull_request.number || github.ref }}
-  cancel-in-progress: true
-
-permissions:
-  contents: read
-
-jobs:
-  backend-blocking-io:
-    if: github.event_name != 'pull_request' || github.event.pull_request.draft == false
-    runs-on: ubuntu-latest
-    timeout-minutes: 10
-
-    steps:
-      - name: Checkout
-        uses: actions/checkout@v4
-
-      - name: Set up Python
-        uses: actions/setup-python@v5
-        with:
-          python-version: "3.12"
-
-      - name: Install uv
-        uses: astral-sh/setup-uv@v3
-
-      - name: Install backend dependencies
-        working-directory: backend
-        run: uv sync --group dev
-
-      - name: Run blocking IO regression tests
-        working-directory: backend
-        run: make test-blocking-io

.github/workflows/container.yaml

@@ -1,101 +0,0 @@
-name: Publish Containers
-
-on:
-  push:
-    tags:
-      - "v*"
-
-jobs:
-
-  backend-container:
-    runs-on: ubuntu-latest
-    permissions:
-      contents: read
-      packages: write
-      attestations: write
-      id-token: write
-    env:
-      REGISTRY: ghcr.io
-      IMAGE_NAME: ${{ github.repository }}-backend
-    steps:
-      - name: Checkout repository
-        uses: actions/checkout@v6
-      - name: Log in to the Container registry
-        uses: docker/login-action@74a5d142397b4f367a81961eba4e8cd7edddf772 #v3.4.0
-        with:
-          registry: ${{ env.REGISTRY }}
-          username: ${{ github.actor }}
-          password: ${{ secrets.GITHUB_TOKEN }}
-      - name: Extract metadata (tags, labels) for Docker
-        id: meta
-        uses: docker/metadata-action@902fa8ec7d6ecbf8d84d538b9b233a880e428804 #v5.7.0
-        with:
-          images: ${{ env.REGISTRY }}/${{ env.IMAGE_NAME }}
-          tags: |
-            type=ref,event=tag
-            type=ref,event=branch
-            type=sha
-            type=raw,value=latest,enable={{is_default_branch}}
-      - name: Build and push Docker image
-        id: push
-        uses: docker/build-push-action@263435318d21b8e681c14492fe198d362a7d2c83 #v6.18.0
-        with:
-          context: .
-          file: backend/Dockerfile
-          push: true
-          tags: ${{ steps.meta.outputs.tags }}
-          labels: ${{ steps.meta.outputs.labels }}
-
-      - name: Generate artifact attestation
-        uses: actions/attest-build-provenance@v2
-        with:
-          subject-name: ${{ env.REGISTRY }}/${{ env.IMAGE_NAME}}
-          subject-digest: ${{ steps.push.outputs.digest }}
-          push-to-registry: true
-
-  frontend-container:
-    runs-on: ubuntu-latest
-    permissions:
-      contents: read
-      packages: write
-      attestations: write
-      id-token: write
-    env:
-      REGISTRY: ghcr.io
-      IMAGE_NAME: ${{ github.repository }}-frontend
-
-    steps:
-      - name: Checkout repository
-        uses: actions/checkout@v6
-      - name: Log in to the Container registry
-        uses: docker/login-action@74a5d142397b4f367a81961eba4e8cd7edddf772 #v3.4.0
-        with:
-          registry: ${{ env.REGISTRY }}
-          username: ${{ github.actor }}
-          password: ${{ secrets.GITHUB_TOKEN }}
-      - name: Extract metadata (tags, labels) for Docker
-        id: meta
-        uses: docker/metadata-action@902fa8ec7d6ecbf8d84d538b9b233a880e428804 #v5.7.0
-        with:
-          images: ${{ env.REGISTRY }}/${{ env.IMAGE_NAME }}
-          tags: |
-            type=ref,event=tag
-            type=ref,event=branch
-            type=sha
-            type=raw,value=latest,enable={{is_default_branch}}
-      - name: Build and push Docker image
-        id: push
-        uses: docker/build-push-action@263435318d21b8e681c14492fe198d362a7d2c83 #v6.18.0
-        with:
-          context: .
-          file: frontend/Dockerfile
-          push: true
-          tags: ${{ steps.meta.outputs.tags }}
-          labels: ${{ steps.meta.outputs.labels }}
-
-      - name: Generate artifact attestation
-        uses: actions/attest-build-provenance@v2
-        with:
-          subject-name: ${{ env.REGISTRY }}/${{ env.IMAGE_NAME}}
-          subject-digest: ${{ steps.push.outputs.digest }}
-          push-to-registry: true

.github/workflows/label-sync.yml

@@ -1,38 +0,0 @@
-name: Label Sync
-
-# Keeps repository labels in sync with the declarative source of truth
-# (.github/labels.yml). Runs whenever that file changes on main, and can be
-# triggered manually. Additive/update-only — never deletes labels.
-
-on:
-  push:
-    branches: [main]
-    paths:
-      - ".github/labels.yml"
-      - "scripts/sync_labels.py"
-      - ".github/workflows/label-sync.yml"
-  workflow_dispatch:
-
-permissions:
-  contents: read
-  issues: write
-
-concurrency:
-  group: label-sync
-  cancel-in-progress: false
-
-jobs:
-  sync:
-    runs-on: ubuntu-latest
-    steps:
-      - name: Checkout
-        uses: actions/checkout@v6
-
-      - name: Install uv
-        uses: astral-sh/setup-uv@v7
-
-      - name: Sync labels
-        run: uv run --with pyyaml python scripts/sync_labels.py
-        env:
-          GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}
-          GH_REPO: ${{ github.repository }}

.github/workflows/replay-e2e.yml

@@ -1,108 +0,0 @@
-name: Replay E2E (front-back contract)
-
-# Guards the front-back contract via record/replay (no API key in CI):
-#   Layer 1 — backend golden: replay a recorded trace through the real gateway,
-#             assert the SSE event sequence matches the committed golden.
-#   Layer 2 — full-stack render: real Next.js frontend + real gateway (replay
-#             model) + Chromium; assert the replayed turns render in the browser.
-# Triggered by changes on EITHER side of the contract so a backend change can no
-# longer pass without the frontend-facing checks running.
-
-on:
-  push:
-    branches: ["main"]
-    paths:
-      - "frontend/**"
-      - "backend/app/gateway/**"
-      - "backend/packages/harness/**"
-      - "backend/tests/fixtures/replay/**"
-      - "backend/tests/replay_provider.py"
-      - "backend/tests/_replay_fixture.py"
-      - "backend/tests/seed_runs_router.py"
-      - "backend/tests/test_replay_golden.py"
-      - "backend/scripts/run_replay_gateway.py"
-      - ".github/workflows/replay-e2e.yml"
-  pull_request:
-    types: [opened, synchronize, reopened, ready_for_review]
-    paths:
-      - "frontend/**"
-      - "backend/app/gateway/**"
-      - "backend/packages/harness/**"
-      - "backend/tests/fixtures/replay/**"
-      - "backend/tests/replay_provider.py"
-      - "backend/tests/_replay_fixture.py"
-      - "backend/tests/seed_runs_router.py"
-      - "backend/tests/test_replay_golden.py"
-      - "backend/scripts/run_replay_gateway.py"
-      - ".github/workflows/replay-e2e.yml"
-
-concurrency:
-  group: replay-e2e-${{ github.event.pull_request.number || github.ref }}
-  cancel-in-progress: true
-
-permissions:
-  contents: read
-
-jobs:
-  backend-replay-golden:
-    name: Layer 1 — backend golden (no API key)
-    if: github.event_name != 'pull_request' || github.event.pull_request.draft == false
-    runs-on: ubuntu-latest
-    timeout-minutes: 15
-    steps:
-      - uses: actions/checkout@v6
-      - name: Set up Python
-        uses: actions/setup-python@v6
-        with:
-          python-version: "3.12"
-      - name: Install uv
-        uses: astral-sh/setup-uv@v7
-      - name: Install backend dependencies
-        working-directory: backend
-        run: uv sync --group dev
-      - name: Replay golden (backend SSE contract)
-        working-directory: backend
-        run: PYTHONPATH=. uv run pytest tests/test_replay_golden.py -v
-
-  fullstack-replay-render:
-    name: Layer 2 — full-stack render (no API key)
-    if: github.event_name != 'pull_request' || github.event.pull_request.draft == false
-    runs-on: ubuntu-latest
-    timeout-minutes: 25
-    steps:
-      - uses: actions/checkout@v6
-      - name: Set up Python
-        uses: actions/setup-python@v6
-        with:
-          python-version: "3.12"
-      - name: Install uv
-        uses: astral-sh/setup-uv@v7
-      - name: Install backend dependencies (replay gateway)
-        working-directory: backend
-        run: uv sync --group dev
-      - name: Setup Node.js
-        uses: actions/setup-node@v4
-        with:
-          node-version: "22"
-      - name: Enable Corepack
-        run: corepack enable
-      - name: Use pinned pnpm version
-        run: corepack prepare pnpm@10.26.2 --activate
-      - name: Install frontend dependencies
-        working-directory: frontend
-        run: pnpm install --frozen-lockfile
-      - name: Install Playwright Chromium
-        working-directory: frontend
-        run: npx playwright install chromium --with-deps
-      - name: Full-stack replay render (DOM assertions are the gate)
-        working-directory: frontend
-        run: pnpm exec playwright test -c playwright.real-backend.config.ts
-      - name: Upload report + render artifact
-        uses: actions/upload-artifact@v4
-        if: ${{ !cancelled() }}
-        with:
-          name: replay-render
-          path: |
-            frontend/playwright-report/
-            frontend/test-results/
-          retention-days: 7

.github/workflows/triage.yml

@@ -1,223 +0,0 @@
-name: Triage
-
-# One workflow for all event-driven PR/issue labeling. Replaces the former
-# pr-labeler / pr-triage / issue-triage workflows (and drops actions/labeler).
-#
-# Design notes:
-#   * All jobs are pure-metadata: they read changed-file lists / PR fields / the
-#     review payload via the API and write labels. PR code is NEVER checked out
-#     or executed, so pull_request_target is safe here.
-#   * Each job only reconciles labels in namespaces IT owns
-#     (area:* / size/* / risk:* / needs-validation). It never touches labels
-#     applied by maintainers or other tools (bug, priority, etc.). first-time-
-#     contributor and reviewing are add-only.
-#   * State is read LIVE (listFiles + listLabelsOnIssue) at run time, not from
-#     the (stale) event payload, so rapid synchronize events converge instead
-#     of thrashing.
-
-on:
-  pull_request_target:
-    types: [opened, synchronize, reopened, ready_for_review]
-  pull_request_review:
-    types: [submitted]
-  issues:
-    types: [opened]
-
-permissions:
-  contents: read
-  pull-requests: write
-  issues: write
-
-jobs:
-  # ── PR: area / size / risk / needs-validation / first-time ─────────────────
-  pr-labels:
-    if: github.event_name == 'pull_request_target' && github.event.pull_request.draft == false
-    runs-on: ubuntu-latest
-    concurrency:
-      group: triage-pr-${{ github.event.pull_request.number }}
-      cancel-in-progress: true
-    steps:
-      - name: Apply PR labels from live state
-        uses: actions/github-script@v8
-        with:
-          script: |
-            const pr = context.payload.pull_request;
-            const { owner, repo } = context.repo;
-            const num = pr.number;
-
-            // ---- live changed files ----
-            const files = await github.paginate(github.rest.pulls.listFiles, {
-              owner, repo, pull_number: num, per_page: 100,
-            });
-            const paths = files.map(f => f.filename);
-            const m = (re) => paths.some(p => re.test(p));
-
-            // ---- area: replaces .github/labeler.yml (path -> area) ----
-            const AREA_RULES = [
-              ['area:frontend', [/^frontend\//]],
-              ['area:backend',  [/^backend\/app\//, /^backend\/packages\/harness\/deerflow\/(runtime|persistence|config|tools|guardrails|tracing|models|utils|uploads)\//]],
-              ['area:agents',   [/^backend\/packages\/harness\/deerflow\/(agents|subagents|reflection)\//, /(^|\/)langgraph\.json$/, /^backend\/.*\/prompts\//]],
-              ['area:sandbox',  [/^docker\//, /^backend\/packages\/harness\/deerflow\/sandbox\//, /(^|\/)Dockerfile$/]],
-              ['area:skills',   [/^skills\//, /^backend\/packages\/harness\/deerflow\/skills\//, /^frontend\/src\/core\/skills\//]],
-              ['area:mcp',      [/^backend\/packages\/harness\/deerflow\/mcp\//, /^frontend\/src\/core\/mcp\//]],
-              ['area:ci',       [/^\.github\//, /^scripts\//]],
-              ['area:docs',     [/^docs\//, /\.mdx?$/]],
-              ['area:deps',     [/(^|\/)(pyproject\.toml|uv\.lock|package\.json|pnpm-lock\.yaml)$/]],
-            ];
-            const areaLabels = AREA_RULES
-              .filter(([, res]) => res.some(re => m(re)))
-              .map(([label]) => label);
-
-            // ---- size: additions+deletions, excluding lockfiles/snapshots ----
-            const EXCLUDE_SIZE = /(^|\/)(uv\.lock|pnpm-lock\.yaml|package-lock\.json)$|\.snap$/;
-            const churn = files
-              .filter(f => !EXCLUDE_SIZE.test(f.filename))
-              .reduce((s, f) => s + (f.additions || 0) + (f.deletions || 0), 0);
-            const sizeLabel =
-              churn < 20 ? 'size/XS' :
-              churn < 100 ? 'size/S' :
-              churn < 300 ? 'size/M' :
-              churn < 700 ? 'size/L' : 'size/XL';
-
-            // ---- risk ----
-            const docsOnly = paths.length > 0 && paths.every(p =>
-              /\.(md|mdx|txt)$/i.test(p) || p.startsWith('docs/') ||
-              /\.(png|jpe?g|gif|svg|webp|ico)$/i.test(p));
-            const highRisk =
-              m(/^backend\/app\/gateway\//) ||
-              m(/^backend\/packages\/harness\/deerflow\/(agents|subagents|sandbox)\//) ||
-              m(/(^|\/)langgraph\.json$/) ||
-              m(/(^|\/)(auth|authz|security)/i) ||
-              m(/(pyproject\.toml|uv\.lock|package\.json|pnpm-lock\.yaml)$/) ||
-              m(/^docker\//) ||
-              m(/^\.github\/workflows\//);
-            const riskLabel = docsOnly ? 'risk:low' : (highRisk ? 'risk:high' : 'risk:medium');
-
-            // ---- needs-validation: front/back contract surface ----
-            const contract =
-              m(/^backend\/app\/gateway\//) ||
-              m(/^backend\/packages\/harness\/deerflow\/(agents|subagents)\//) ||
-              m(/(^|\/)langgraph\.json$/) ||
-              m(/^frontend\/src\/core\/(api|threads|messages)\//);
-
-            // ---- live current labels (NOT the stale event payload) ----
-            const current = (await github.paginate(github.rest.issues.listLabelsOnIssue, {
-              owner, repo, issue_number: num, per_page: 100,
-            })).map(l => l.name);
-            const hasSkip = current.includes('skip-validation');
-
-            // Reconcile ONLY namespaces we own; never touch others.
-            const owned = (n) =>
-              n.startsWith('area:') || n.startsWith('size/') ||
-              n.startsWith('risk:') || n === 'needs-validation';
-            const desired = new Set([...areaLabels, sizeLabel, riskLabel]);
-            if (contract && !hasSkip) desired.add('needs-validation');
-
-            const toRemove = current.filter(n => owned(n) && !desired.has(n));
-            const toAdd = [...desired].filter(n => !current.includes(n));
-
-            // first-time-contributor: add-only, on opened, real users only.
-            if (context.payload.action === 'opened' &&
-                pr.user.type === 'User' &&
-                ['FIRST_TIME_CONTRIBUTOR', 'FIRST_TIMER'].includes(pr.author_association) &&
-                !current.includes('first-time-contributor')) {
-              toAdd.push('first-time-contributor');
-            }
-
-            for (const name of toRemove) {
-              try {
-                await github.rest.issues.removeLabel({ owner, repo, issue_number: num, name });
-              } catch (e) {
-                if (e.status !== 404) throw e;
-              }
-            }
-            if (toAdd.length) {
-              await github.rest.issues.addLabels({ owner, repo, issue_number: num, labels: toAdd });
-            }
-            core.info(`area=[${areaLabels.join(',')}] ${sizeLabel} ${riskLabel} churn=${churn} ` +
-              `validation=${desired.has('needs-validation')} ` +
-              `(+${toAdd.join(',') || '-'} / -${toRemove.join(',') || '-'})`);
-
-  # ── PR: reviewing label on a maintainer's human review ─────────────────────
-  reviewing:
-    if: github.event_name == 'pull_request_review'
-    runs-on: ubuntu-latest
-    concurrency:
-      group: triage-review-${{ github.event.pull_request.number }}
-      cancel-in-progress: false
-    steps:
-      - name: Add reviewing label for maintainer reviews
-        uses: actions/github-script@v8
-        with:
-          script: |
-            const { owner, repo } = context.repo;
-            const num = context.payload.pull_request.number;
-            const review = context.payload.review;
-            const assoc = review.author_association;     // payload field; no API call
-            const type = review.user && review.user.type;
-
-            // author_association is NONE for every automated reviewer
-            // (Copilot, CodeRabbit, Codex, Sourcery, ...), so this allowlist
-            // drops them all without a denylist — and never calls the
-            // collaborators API that 404s on "Copilot is not a user".
-            // user.type === 'User' guards the rare bot-added-as-collaborator case.
-            if (!['OWNER', 'MEMBER', 'COLLABORATOR'].includes(assoc) || type !== 'User') {
-              core.info(`reviewer ${review.user && review.user.login} assoc=${assoc} type=${type}; skipping.`);
-              return;
-            }
-
-            const labels = (await github.paginate(github.rest.issues.listLabelsOnIssue, {
-              owner, repo, issue_number: num, per_page: 100,
-            })).map(l => l.name);
-            if (labels.includes('reviewing')) {
-              core.info('Already labeled reviewing; skipping.');
-              return;
-            }
-            try {
-              await github.rest.issues.addLabels({
-                owner, repo, issue_number: num, labels: ['reviewing'],
-              });
-              core.info('Added "reviewing".');
-            } catch (e) {
-              if (e.status === 403) core.info('No permission to label (expected on some fork PRs).');
-              else throw e;
-            }
-
-  # ── Issue: needs-triage on every new issue ────────────────────────────────
-  issue-triage:
-    if: github.event_name == 'issues'
-    runs-on: ubuntu-latest
-    concurrency:
-      group: triage-issue-${{ github.event.issue.number }}
-      cancel-in-progress: false
-    steps:
-      - name: Add needs-triage label
-        uses: actions/github-script@v8
-        with:
-          script: |
-            const { owner, repo } = context.repo;
-            const issue_number = context.payload.issue.number;
-
-            // Read live labels (not the event payload) so labels added at creation
-            // time via the API or by another automation are seen — consistent with
-            // the live-state reads in the PR jobs above.
-            const current = (await github.paginate(github.rest.issues.listLabelsOnIssue, {
-              owner, repo, issue_number, per_page: 100,
-            })).map(l => l.name);
-            if (current.includes('needs-triage')) {
-              core.info('Issue already has needs-triage; nothing to do.');
-              return;
-            }
-            // Self-heal: create the label if it does not exist yet.
-            try {
-              await github.rest.issues.createLabel({
-                owner, repo, name: 'needs-triage', color: 'fef2c0',
-                description: 'Awaiting maintainer triage',
-              });
-            } catch (e) {
-              if (e.status !== 422) throw e; // 422 = already exists
-            }
-            await github.rest.issues.addLabels({
-              owner, repo, issue_number, labels: ['needs-triage'],
-            });
-            core.info(`Added needs-triage to #${issue_number}.`);

CONTRIBUTING.md

@@ -46,12 +46,12 @@ Docker provides a consistent, isolated environment with all dependencies pre-con
    All services will start with hot-reload enabled:
    - Frontend changes are automatically reloaded
    - Backend changes trigger automatic restart
-   - Gateway-hosted LangGraph-compatible runtime supports hot-reload
+   - LangGraph server supports hot-reload
 
 4. **Access the application**:
    - Web Interface: http://localhost:2026
    - API Gateway: http://localhost:2026/api/*
-   - LangGraph-compatible API: http://localhost:2026/api/langgraph/*
+   - LangGraph: http://localhost:2026/api/langgraph/*
 
 #### Docker Commands
 
@@ -94,7 +94,7 @@ Use these as practical starting points for development and review environments:
 If `make docker-init`, `make docker-start`, or `make docker-stop` fails on Linux with an error like below, your current user likely does not have permission to access the Docker daemon socket:
 
 ```text
-unable to get image 'deer-flow-gateway': permission denied while trying to connect to the Docker daemon socket at unix:///var/run/docker.sock
+unable to get image 'deer-flow-dev-langgraph': permission denied while trying to connect to the Docker daemon socket at unix:///var/run/docker.sock
 ```
 
 Recommended fix: add your current user to the `docker` group so Docker commands work without `sudo`.
@@ -131,8 +131,9 @@ Host Machine
 Docker Compose (deer-flow-dev)
   ├→ nginx (port 2026) ← Reverse proxy
   ├→ web (port 3000) ← Frontend with hot-reload
-  ├→ gateway (port 8001) ← Gateway API + LangGraph-compatible runtime with hot-reload
+  ├→ api (port 8001) ← Gateway API with hot-reload
-  └→ provisioner (optional, port 8002) ← Started only in provisioner/K8s sandbox mode
+   ├→ langgraph (port 2024) ← LangGraph server with hot-reload
+   └→ provisioner (optional, port 8002) ← Started only in provisioner/K8s sandbox mode
 ```
 
 **Benefits of Docker Development**:
@@ -183,13 +184,17 @@ Required tools:
 
 If you need to start services individually:
 
-1. **Start backend service**:
+1. **Start backend services**:
    ```bash
-   # Terminal 1: Start Gateway API + embedded agent runtime (port 8001)
+   # Terminal 1: Start LangGraph Server (port 2024)
    cd backend
    make dev
 
-   # Terminal 2: Start Frontend (port 3000)
+   # Terminal 2: Start Gateway API (port 8001)
+   cd backend
+   make gateway
+
+   # Terminal 3: Start Frontend (port 3000)
    cd frontend
    pnpm dev
    ```
@@ -207,10 +212,10 @@ If you need to start services individually:
 
 The nginx configuration provides:
 - Unified entry point on port 2026
-- Rewrites `/api/langgraph/*` to Gateway's LangGraph-compatible API (8001)
+- Routes `/api/langgraph/*` to LangGraph Server (2024)
 - Routes other `/api/*` endpoints to Gateway API (8001)
 - Routes non-API requests to Frontend (3000)
-- Same-origin API routing; split-origin or port-forwarded browser clients should use the Gateway `GATEWAY_CORS_ORIGINS` allowlist
+- Centralized CORS handling
 - SSE/streaming support for real-time agent responses
 - Optimized timeouts for long-running operations
 
@@ -230,8 +235,8 @@ deer-flow/
 │       └── nginx.local.conf # Nginx config for local dev
 ├── backend/                 # Backend application
 │   ├── src/
-│   │   ├── gateway/        # Gateway API and LangGraph-compatible runtime (port 8001)
+│   │   ├── gateway/        # Gateway API (port 8001)
-│   │   ├── agents/         # LangGraph agent runtime used by Gateway
+│   │   ├── agents/         # LangGraph agents (port 2024)
 │   │   ├── mcp/            # Model Context Protocol integration
 │   │   ├── skills/         # Skills system
 │   │   └── sandbox/        # Sandbox execution
@@ -251,7 +256,8 @@ Browser
   ↓
 Nginx (port 2026) ← Unified entry point
   ├→ Frontend (port 3000) ← / (non-API requests)
-  └→ Gateway API (port 8001) ← /api/* and /api/langgraph/* (LangGraph-compatible agent interactions)
+  ├→ Gateway API (port 8001) ← /api/models, /api/mcp, /api/skills, /api/threads/*/artifacts
+  └→ LangGraph Server (port 2024) ← /api/langgraph/* (agent interactions)
 ```
 
 ## Development Workflow
@@ -287,21 +293,6 @@ Nginx (port 2026) ← Unified entry point
    git push origin feature/your-feature-name
    ```
 
-## AI assistance disclosure
-
-DeerFlow is an AI project and we welcome AI-assisted contributions. To help
-reviewers calibrate how closely to read a change, **every pull request must
-complete the "AI assistance" section of the
-[PR template](.github/pull_request_template.md)**:
-
-- which tool(s) you used (or `none`),
-- how you used them, and
-- a confirmation that a human has read, understands, and takes responsibility
-  for the change.
-
-Please don't delete the section. PRs that ignore it may be asked to fill it in
-before review.
-
 ## Testing
 
 ```bash

Makefile

@@ -1,6 +1,6 @@
 # DeerFlow - Unified Development Environment
 
-.PHONY: help config config-upgrade check install setup doctor detect-thread-boundaries detect-blocking-io dev dev-daemon start start-daemon stop up down clean docker-init docker-start docker-stop docker-logs docker-logs-frontend docker-logs-gateway
+.PHONY: help config config-upgrade check install setup doctor dev dev-daemon start start-daemon stop up down clean docker-init docker-start docker-stop docker-logs docker-logs-frontend docker-logs-gateway
 
 BASH ?= bash
 BACKEND_UV_RUN = cd backend && uv run
@@ -23,8 +23,6 @@ help:
 	@echo "  make config          - Generate local config files (aborts if config already exists)"
 	@echo "  make config-upgrade  - Merge new fields from config.example.yaml into config.yaml"
 	@echo "  make check           - Check if all required tools are installed"
-	@echo "  make detect-thread-boundaries - Inventory async/thread boundary points"
-	@echo "  make detect-blocking-io        - Inventory blocking IO that may block the backend event loop"
 	@echo "  make install         - Install all dependencies (frontend + backend + pre-commit hooks)"
 	@echo "  make setup-sandbox   - Pre-pull sandbox container image (recommended)"
 	@echo "  make dev             - Start all services in development mode (with hot-reloading)"
@@ -53,12 +51,6 @@ setup:
 doctor:
 	@$(BACKEND_UV_RUN) python ../scripts/doctor.py
 
-detect-thread-boundaries:
-	@$(PYTHON) ./scripts/detect_thread_boundaries.py
-
-detect-blocking-io:
-	@$(MAKE) -C backend detect-blocking-io
-
 config:
 	@$(PYTHON) ./scripts/configure.py
 
@@ -89,7 +81,36 @@ install:
 
 # Pre-pull sandbox Docker image (optional but recommended)
 setup-sandbox:
-	@$(RUN_WITH_GIT_BASH) ./scripts/setup-sandbox.sh
+	@echo "=========================================="
+	@echo "  Pre-pulling Sandbox Container Image"
+	@echo "=========================================="
+	@echo ""
+	@IMAGE=$$(grep -A 20 "# sandbox:" config.yaml 2>/dev/null | grep "image:" | awk '{print $$2}' | head -1); \
+	if [ -z "$$IMAGE" ]; then \
+		IMAGE="enterprise-public-cn-beijing.cr.volces.com/vefaas-public/all-in-one-sandbox:latest"; \
+		echo "Using default image: $$IMAGE"; \
+	else \
+		echo "Using configured image: $$IMAGE"; \
+	fi; \
+	echo ""; \
+	if command -v container >/dev/null 2>&1 && [ "$$(uname)" = "Darwin" ]; then \
+		echo "Detected Apple Container on macOS, pulling image..."; \
+		container image pull "$$IMAGE" || echo "⚠ Apple Container pull failed, will try Docker"; \
+	fi; \
+	if command -v docker >/dev/null 2>&1; then \
+		echo "Pulling image using Docker..."; \
+		if docker pull "$$IMAGE"; then \
+			echo ""; \
+			echo "✓ Sandbox image pulled successfully"; \
+		else \
+			echo ""; \
+			echo "⚠ Failed to pull sandbox image (this is OK for local sandbox mode)"; \
+		fi; \
+	else \
+		echo "✗ Neither Docker nor Apple Container is available"; \
+		echo "  Please install Docker: https://docs.docker.com/get-docker/"; \
+		exit 1; \
+	fi
 
 # Start all services in development mode (with hot-reloading)
 dev:
@@ -119,6 +140,7 @@ stop:
 clean: stop
 	@echo "Cleaning up..."
 	@-rm -rf backend/.deer-flow 2>/dev/null || true
+	@-rm -rf backend/.langgraph_api 2>/dev/null || true
 	@-rm -rf logs/*.log 2>/dev/null || true
 	@echo "✓ Cleanup complete"
 

README.md

@@ -245,8 +245,6 @@ make down   # Stop and remove containers
 
 Access: http://localhost:2026
 
-The unified nginx endpoint is same-origin by default and does not emit browser CORS headers. If you run a split-origin or port-forwarded browser client, set `GATEWAY_CORS_ORIGINS` to comma-separated exact origins such as `http://localhost:3000`; the Gateway then applies the CORS allowlist and matching CSRF origin checks.
-
 See [CONTRIBUTING.md](CONTRIBUTING.md) for detailed Docker development guide.
 
 #### Option 2: Local Development
@@ -546,15 +544,6 @@ LANGFUSE_BASE_URL=https://cloud.langfuse.com
 
 If you are using a self-hosted Langfuse instance, set `LANGFUSE_BASE_URL` to your deployment URL.
 
-**Trace correlation fields.** Every agent run is annotated with Langfuse's reserved trace attributes so the Sessions and Users pages light up automatically:
-
-- `session_id` = LangGraph `thread_id` — groups every trace of the same conversation
-- `user_id` = effective user from `get_effective_user_id()` (falls back to `default` in no-auth mode)
-- `trace_name` = assistant id (defaults to `lead-agent`)
-- `tags` = `[env:<DEER_FLOW_ENV>, model:<model_name>]` (omitted when not set)
-
-These are injected into `RunnableConfig.metadata` at the graph invocation root for both the gateway path (`runtime/runs/worker.py::run_agent`) and the embedded path (`client.py::DeerFlowClient.stream`), so any LangChain-compatible callback can read them. Set `DEER_FLOW_ENV` (or `ENVIRONMENT`) to tag traces by deployment environment.
-
 #### Using Both Providers
 
 If both LangSmith and Langfuse are enabled, DeerFlow attaches both tracing callbacks and reports the same model activity to both systems.
@@ -585,8 +574,6 @@ A standard Agent Skill is a structured capability module — a Markdown file tha
 
 Skills are loaded progressively — only when the task needs them, not all at once. This keeps the context window lean and makes DeerFlow work well even with token-sensitive models.
 
-Users can explicitly activate an enabled skill for a single turn by starting the request with `/skill-name`, for example `/data-analysis analyze uploads/foo.csv`. DeerFlow loads that skill's `SKILL.md` as hidden current-turn context while leaving the base prompt limited to skill metadata. Slash activation respects disabled skills, custom-agent skill whitelists, and existing channel commands such as `/new` and `/help`.
-
 When you install `.skill` archives through the Gateway, DeerFlow accepts standard optional frontmatter metadata such as `version`, `author`, and `compatibility` instead of rejecting otherwise valid external skills.
 
 Tools follow the same philosophy. DeerFlow comes with a core toolset — web search, web fetch, file operations, bash execution — and supports custom tools via MCP servers and Python functions. Swap anything. Add anything.
@@ -639,7 +626,7 @@ See [`skills/public/claude-to-deerflow/SKILL.md`](skills/public/claude-to-deerfl
 
 Complex tasks rarely fit in a single pass. DeerFlow decomposes them.
 
-The lead agent can spawn sub-agents on the fly — each with its own scoped context, tools, and termination conditions. Sub-agents run in parallel when possible, report back structured results, and the lead agent synthesizes everything into a coherent output. When token usage tracking is enabled, completed sub-agent usage is attributed back to the dispatching step.
+The lead agent can spawn sub-agents on the fly — each with its own scoped context, tools, and termination conditions. Sub-agents run in parallel when possible, report back structured results, and the lead agent synthesizes everything into a coherent output.
 
 This is how DeerFlow handles tasks that take minutes to hours: a research task might fan out into a dozen sub-agents, each exploring a different angle, then converge into a single report — or a website — or a slide deck with generated visuals. One harness, many hands.
 
@@ -742,12 +729,6 @@ DeerFlow has key high-privilege capabilities including **system command executio
 We welcome contributions! Please see [CONTRIBUTING.md](CONTRIBUTING.md) for development setup, workflow, and guidelines.
 
 Regression coverage includes Docker sandbox mode detection and provisioner kubeconfig-path handling tests in `backend/tests/`.
-Backend blocking-IO diagnostics are available from the repository root with
-`make detect-blocking-io`: it statically scans backend business code for
-blocking IO that may run on the backend event loop, prints a concise summary,
-and writes complete JSON findings to `.deer-flow/blocking-io-findings.json`.
-The JSON includes compact review records with `priority`, `location`,
-`blocking_call`, `event_loop_exposure`, `reason`, and `code`.
 Gateway artifact serving now forces active web content types (`text/html`, `application/xhtml+xml`, `image/svg+xml`) to download as attachments instead of inline rendering, reducing XSS risk for generated artifacts.
 
 ## License

README_fr.md

@@ -228,7 +228,7 @@ make down   # Stop and remove containers
 ```
 
 > [!NOTE]
-> Le runtime d'agent s'exécute actuellement dans la Gateway. nginx réécrit `/api/langgraph/*` vers l'API compatible LangGraph servie par la Gateway.
+> Le serveur d'agents LangGraph fonctionne actuellement via `langgraph dev` (le serveur CLI open source).
 
 Accès : http://localhost:2026
 
@@ -296,8 +296,8 @@ DeerFlow peut recevoir des tâches depuis des applications de messagerie. Les ca
 
 ```yaml
 channels:
-  # LangGraph-compatible Gateway API base URL (default: http://localhost:8001/api)
+  # LangGraph Server URL (default: http://localhost:2024)
-  langgraph_url: http://localhost:8001/api
+  langgraph_url: http://localhost:2024
   # Gateway API URL (default: http://localhost:8001)
   gateway_url: http://localhost:8001
 

README_ja.md

@@ -181,7 +181,7 @@ make down   # コンテナを停止して削除
 ```
 
 > [!NOTE]
-> Agentランタイムは現在Gateway内で実行されます。`/api/langgraph/*`はnginxによってGatewayのLangGraph-compatible APIへ書き換えられます。
+> LangGraphエージェントサーバーは現在`langgraph dev`（オープンソースCLIサーバー）経由で実行されます。
 
 アクセス: http://localhost:2026
 
@@ -249,8 +249,8 @@ DeerFlowはメッセージングアプリからのタスク受信をサポート
 
 ```yaml
 channels:
-  # LangGraph-compatible Gateway API base URL（デフォルト: http://localhost:8001/api）
+  # LangGraphサーバーURL（デフォルト: http://localhost:2024）
-  langgraph_url: http://localhost:8001/api
+  langgraph_url: http://localhost:2024
   # Gateway API URL（デフォルト: http://localhost:8001）
   gateway_url: http://localhost:8001
 

README_zh.md

@@ -184,7 +184,7 @@ make down   # 停止并移除容器
 ```
 
 > [!NOTE]
-> 当前 Agent 运行时嵌入在 Gateway 中运行，`/api/langgraph/*` 会由 nginx 重写到 Gateway 的 LangGraph-compatible API。
+> 当前 LangGraph agent server 通过开源 CLI 服务 `langgraph dev` 运行。
 
 访问地址：http://localhost:2026
 
@@ -254,8 +254,8 @@ DeerFlow 支持从即时通讯应用接收任务。只要配置完成，对应
 
 ```yaml
 channels:
-  # LangGraph-compatible Gateway API base URL（默认：http://localhost:8001/api）
+  # LangGraph Server URL（默认：http://localhost:2024）
-  langgraph_url: http://localhost:8001/api
+  langgraph_url: http://localhost:2024
   # Gateway API URL（默认：http://localhost:8001）
   gateway_url: http://localhost:8001
 

backend/.gitignore

@@ -24,10 +24,5 @@ config.yaml
 # Langgraph
 .langgraph_api
 
-# Sandbox runtime working dir — pre-created and excluded from uvicorn reload
-# (scripts/serve.sh, docker/dev-entrypoint.sh). Anchored so it does not match
-# the source package backend/packages/harness/deerflow/sandbox/.
-/sandbox/
-
 # Claude Code settings
 .claude/settings.local.json

backend/CLAUDE.md

@@ -88,57 +88,18 @@ make stop       # Stop all services
 
 **Backend directory** (for backend development only):
 ```bash
-make install            # Install backend dependencies
+make install    # Install backend dependencies
-make dev                # Run Gateway API with reload (port 8001)
+make dev        # Run Gateway API with reload (port 8001)
-make gateway            # Run Gateway API only (port 8001)
+make gateway    # Run Gateway API only (port 8001)
-make test               # Run all backend tests
+make test       # Run all backend tests
-make test-blocking-io   # Run strict Blockbuster runtime gate on tests/blocking_io/
+make lint       # Lint with ruff
-make lint               # Lint with ruff
+make format     # Format code with ruff
-make format             # Format code with ruff
 ```
 
-The `detect-blocking-io` target parses `app/`, `packages/harness/deerflow/`,
-and `scripts/` with AST. By default it reports only blocking IO candidates that
-are inside async code, reachable from async code in the same file, or reachable
-from sync-only `AgentMiddleware` before/after hooks that LangGraph can execute
-on the async graph path. It prints a concise summary and writes complete JSON
-findings to `.deer-flow/blocking-io-findings.json` at the repository root
-(both `make detect-blocking-io` from the repo root and `cd backend && make
-detect-blocking-io` resolve to the same repo-root path). JSON findings include
-`priority`, `location`, `blocking_call`, `event_loop_exposure`, `reason`, and
-`code` for model-assisted or manual review. `priority` is a deterministic
-review ordering from operation type, not proof of a bug. Bare-name same-file
-calls are resolved by function name, so duplicate helper names in one file can
-conservatively over-report async reachability. It is intentionally
-informational and is not run from CI in this round.
-
 Regression tests related to Docker/provisioner behavior:
 - `tests/test_docker_sandbox_mode_detection.py` (mode detection from `config.yaml`)
 - `tests/test_provisioner_kubeconfig.py` (kubeconfig file/directory handling)
 
-Blocking-IO runtime gate (`tests/blocking_io/`):
-- Wraps every item under `tests/blocking_io/` with a strict Blockbuster
-  context scoped to `app.*` and `deerflow.*` (see
-  `tests/support/detectors/blocking_io_runtime.py`). Any sync blocking IO
-  call whose stack passes through DeerFlow business code while running on
-  the asyncio event loop raises `BlockingError` and fails the test.
-- Regression anchors live there: `test_skills_load.py` (locks the
-  `asyncio.to_thread` offload around `LocalSkillStorage.load_skills`, fix
-  for #1917); `test_sqlite_lifespan.py` (locks the offload around
-  SQLite path resolution plus `ensure_sqlite_parent_dir`, fix for #1912);
-  `test_jsonl_run_event_store.py` (locks `JsonlRunEventStore`'s async
-  API offloading its file IO via `asyncio.to_thread`, fix #3084); and
-  `test_uploads_middleware.py` (locks `UploadsMiddleware.abefore_agent`
-  offloading the uploads-directory scan off the event loop).
-- `test_gate_smoke.py` is a meta-test asserting the gate actually catches
-  unoffloaded blocking IO and that the `@pytest.mark.allow_blocking_io`
-  opt-out works.
-- Coverage boundary: the gate only sees code that test execution actually
-  touches. Static AST coverage is a separate concern (out of scope for
-  this PR).
-- CI: runs on every PR via `.github/workflows/backend-blocking-io-tests.yml`,
-  hard-fail.
-
 Boundary check (harness → app import firewall):
 - `tests/test_harness_boundary.py` — ensures `packages/harness/deerflow/` never imports from `app.*`
 
@@ -192,7 +153,7 @@ from deerflow.config import get_app_config
 
 ### Middleware Chain
 
-Lead-agent middlewares are assembled in strict append order across `packages/harness/deerflow/agents/middlewares/tool_error_handling_middleware.py` (`build_lead_runtime_middlewares`) and `packages/harness/deerflow/agents/lead_agent/agent.py` (`build_middlewares`):
+Lead-agent middlewares are assembled in strict append order across `packages/harness/deerflow/agents/middlewares/tool_error_handling_middleware.py` (`build_lead_runtime_middlewares`) and `packages/harness/deerflow/agents/lead_agent/agent.py` (`_build_middlewares`):
 
 1. **ThreadDataMiddleware** - Creates per-thread directories under the user's isolation scope (`backend/.deer-flow/users/{user_id}/threads/{thread_id}/user-data/{workspace,uploads,outputs}`); resolves `user_id` via `get_effective_user_id()` (falls back to `"default"` in no-auth mode); Web UI thread deletion now follows LangGraph thread removal with Gateway cleanup of the local thread directory
 2. **UploadsMiddleware** - Tracks and injects newly uploaded files into conversation
@@ -202,17 +163,16 @@ Lead-agent middlewares are assembled in strict append order across `packages/har
 6. **GuardrailMiddleware** - Pre-tool-call authorization via pluggable `GuardrailProvider` protocol (optional, if `guardrails.enabled` in config). Evaluates each tool call and returns error ToolMessage on deny. Three provider options: built-in `AllowlistProvider` (zero deps), OAP policy providers (e.g. `aport-agent-guardrails`), or custom providers. See [docs/GUARDRAILS.md](docs/GUARDRAILS.md) for setup, usage, and how to implement a provider.
 7. **SandboxAuditMiddleware** - Audits sandboxed shell/file operations for security logging before tool execution continues
 8. **ToolErrorHandlingMiddleware** - Converts tool exceptions into error `ToolMessage`s so the run can continue instead of aborting
-9. **SkillActivationMiddleware** - Detects strict `/skill-name task` syntax on the latest real user message, resolves only enabled and runtime-allowed skills, reads `SKILL.md` from trusted skill storage, injects the skill body as hidden current-turn model context, and records a `middleware:skill_activation` audit event with skill name, category, path, and content hash
+9. **SummarizationMiddleware** - Context reduction when approaching token limits (optional, if enabled)
-10. **SummarizationMiddleware** - Context reduction when approaching token limits (optional, if enabled)
+10. **TodoListMiddleware** - Task tracking with `write_todos` tool (optional, if plan_mode)
-11. **TodoListMiddleware** - Task tracking with `write_todos` tool (optional, if plan_mode)
+11. **TokenUsageMiddleware** - Records token usage metrics when token tracking is enabled (optional)
-12. **TokenUsageMiddleware** - Records token usage metrics when token tracking is enabled (optional); subagent usage is cached by `tool_call_id` only while token usage is enabled and merged back into the dispatching AIMessage by message position rather than message id
+12. **TitleMiddleware** - Auto-generates thread title after first complete exchange and normalizes structured message content before prompting the title model
-13. **TitleMiddleware** - Auto-generates thread title after first complete exchange and normalizes structured message content before prompting the title model
+13. **MemoryMiddleware** - Queues conversations for async memory update (filters to user + final AI responses)
-14. **MemoryMiddleware** - Queues conversations for async memory update (filters to user + final AI responses)
+14. **ViewImageMiddleware** - Injects base64 image data before LLM call (conditional on vision support)
-15. **ViewImageMiddleware** - Injects base64 image data before LLM call (conditional on vision support)
+15. **DeferredToolFilterMiddleware** - Hides deferred tool schemas from the bound model until tool search is enabled (optional)
-16. **DeferredToolFilterMiddleware** - Hides deferred (MCP) tool schemas from the bound model using a build-time deferred-name set + catalog hash, reading per-thread promotions from `ThreadState.promoted` (hash-scoped, no ContextVar); a tool becomes bound on subsequent turns after `tool_search` returns its schema (optional, if `tool_search.enabled`)
+16. **SubagentLimitMiddleware** - Truncates excess `task` tool calls from model response to enforce `MAX_CONCURRENT_SUBAGENTS` limit (optional, if `subagent_enabled`)
-17. **SubagentLimitMiddleware** - Truncates excess `task` tool calls from model response to enforce `MAX_CONCURRENT_SUBAGENTS` limit (optional, if `subagent_enabled`)
+17. **LoopDetectionMiddleware** - Detects repeated tool-call loops; hard-stop responses clear both structured `tool_calls` and raw provider tool-call metadata before forcing a final text answer
-18. **LoopDetectionMiddleware** - Detects repeated tool-call loops; hard-stop responses clear both structured `tool_calls` and raw provider tool-call metadata before forcing a final text answer
+18. **ClarificationMiddleware** - Intercepts `ask_clarification` tool calls, interrupts via `Command(goto=END)` (must be last)
-19. **ClarificationMiddleware** - Intercepts `ask_clarification` tool calls, interrupts via `Command(goto=END)` (must be last)
 
 ### Configuration System
 
@@ -224,10 +184,6 @@ Setup: Copy `config.example.yaml` to `config.yaml` in the **project root** direc
 
 **Config Caching**: `get_app_config()` caches the parsed config, but automatically reloads it when the resolved config path changes or the file's mtime increases. This keeps Gateway and LangGraph reads aligned with `config.yaml` edits without requiring a manual process restart.
 
-**Config Hot-Reload Boundary**: Gateway dependencies route through `get_app_config()` on every request, so per-run fields like `models[*].max_tokens`, `summarization.*`, `title.*`, `memory.*`, `subagents.*`, `tools[*]`, and the agent system prompt pick up `config.yaml` edits on the next message. `AppConfig` is intentionally **not** cached on `app.state` — `lifespan()` keeps a local `startup_config` variable for one-shot bootstrap work and passes it to `langgraph_runtime(app, startup_config)`.
-
-Infrastructure fields are **restart-required**. The authoritative list lives in `packages/harness/deerflow/config/reload_boundary.py::STARTUP_ONLY_FIELDS` and is mirrored by the standardised `"startup-only:"` prefix on the corresponding `Field(description=...)` in `AppConfig`, so IDE hover on those fields surfaces the reason inline (no need to context-switch into this table). Currently registered: `database`, `checkpointer`, `run_events`, `stream_bridge`, `sandbox`, `log_level`, `channels`. Adding a new restart-required field requires updating the registry; drift is pinned by `tests/test_reload_boundary.py`.
-
 Configuration priority:
 1. Explicit `config_path` argument
 2. `DEER_FLOW_CONFIG_PATH` environment variable
@@ -251,8 +207,6 @@ Configuration priority:
 
 FastAPI application on port 8001 with health check at `GET /health`. Set `GATEWAY_ENABLE_DOCS=false` to disable `/docs`, `/redoc`, and `/openapi.json` in production (default: enabled).
 
-CORS is same-origin by default when requests enter through nginx on port 2026. Split-origin or port-forwarded browser clients must opt in with `GATEWAY_CORS_ORIGINS` (comma-separated exact origins); Gateway `CORSMiddleware` and `CSRFMiddleware` both read that variable so browser CORS and auth-origin checks stay aligned.
-
 **Routers**:
 
 | Router | Endpoints |
@@ -264,39 +218,32 @@ CORS is same-origin by default when requests enter through nginx on port 2026. S
 | **Uploads** (`/api/threads/{id}/uploads`) | `POST /` - upload files (auto-converts PDF/PPT/Excel/Word); `GET /list` - list; `DELETE /{filename}` - delete |
 | **Threads** (`/api/threads/{id}`) | `DELETE /` - remove DeerFlow-managed local thread data after LangGraph thread deletion; unexpected failures are logged server-side and return a generic 500 detail |
 | **Artifacts** (`/api/threads/{id}/artifacts`) | `GET /{path}` - serve artifacts; active content types (`text/html`, `application/xhtml+xml`, `image/svg+xml`) are always forced as download attachments to reduce XSS risk; `?download=true` still forces download for other file types |
-| **Suggestions** (`/api/threads/{id}/suggestions`) | `POST /` - generate follow-up questions; rich list/block model content is normalized and inline reasoning (`<think>...</think>`, including unclosed/truncated blocks from reasoning models like MiniMax-M3) is stripped before JSON parsing |
+| **Suggestions** (`/api/threads/{id}/suggestions`) | `POST /` - generate follow-up questions; rich list/block model content is normalized before JSON parsing |
 | **Thread Runs** (`/api/threads/{id}/runs`) | `POST /` - create background run; `POST /stream` - create + SSE stream; `POST /wait` - create + block; `GET /` - list runs; `GET /{rid}` - run details; `POST /{rid}/cancel` - cancel; `GET /{rid}/join` - join SSE; `GET /{rid}/messages` - paginated messages `{data, has_more}`; `GET /{rid}/events` - full event stream; `GET /../messages` - thread messages with feedback; `GET /../token-usage` - aggregate tokens |
 | **Feedback** (`/api/threads/{id}/runs/{rid}/feedback`) | `PUT /` - upsert feedback; `DELETE /` - delete user feedback; `POST /` - create feedback; `GET /` - list feedback; `GET /stats` - aggregate stats; `DELETE /{fid}` - delete specific |
 | **Runs** (`/api/runs`) | `POST /stream` - stateless run + SSE; `POST /wait` - stateless run + block; `GET /{rid}/messages` - paginated messages by run_id `{data, has_more}` (cursor: `after_seq`/`before_seq`); `GET /{rid}/feedback` - list feedback by run_id |
 
-**RunManager / RunStore contract**:
+Proxied through nginx: `/api/langgraph/*` → LangGraph, all other `/api/*` → Gateway.
-- `RunManager.get()` is async; direct callers must `await` it.
-- When a persistent `RunStore` is configured, `get()` and `list_by_thread()` hydrate historical runs from the store. In-memory records win for the same `run_id` so task, abort, and stream-control state stays attached to active local runs.
-- `cancel()` and `create_or_reject(..., multitask_strategy="interrupt"|"rollback")` persist interrupted status through `RunStore.update_status()`, matching normal `set_status()` transitions.
-- Store-only hydrated runs are readable history. If the current worker has no in-memory task/control state for that run, cancellation APIs can return 409 because this worker cannot stop the task.
-- `POST /wait` (both thread-scoped and `/api/runs/wait`) drains the stream bridge via `wait_for_run_completion()` instead of bare `await record.task`, so it honours the run's `on_disconnect` setting and cancels the background run on real client disconnect rather than returning a stale checkpoint (issue #3265).
-
-Proxied through nginx: `/api/langgraph/*` → Gateway LangGraph-compatible runtime, all other `/api/*` → Gateway REST APIs.
 
 ### Sandbox System (`packages/harness/deerflow/sandbox/`)
 
 **Interface**: Abstract `Sandbox` with `execute_command`, `read_file`, `write_file`, `list_dir`
-**Provider Pattern**: `SandboxProvider` with `acquire`, `acquire_async`, `get`, `release` lifecycle. Async agent/tool paths call async sandbox lifecycle hooks so Docker sandbox creation, discovery, cross-process locking, readiness polling, and release stay off the event loop.
+**Provider Pattern**: `SandboxProvider` with `acquire`, `get`, `release` lifecycle
 **Implementations**:
-- `LocalSandboxProvider` - Local filesystem execution. `acquire(thread_id)` returns a per-thread `LocalSandbox` (id `local:{thread_id}`) whose `path_mappings` resolve `/mnt/user-data/{workspace,uploads,outputs}` and `/mnt/acp-workspace` to that thread's host directories, so the public `Sandbox` API honours the `/mnt/user-data` contract uniformly with AIO. `acquire()` / `acquire(None)` keeps the legacy generic singleton (id `local`) for callers without a thread context. Per-thread sandboxes are held in an LRU cache (default 256 entries) guarded by a `threading.Lock`.
+- `LocalSandboxProvider` - Singleton local filesystem execution with path mappings
 - `AioSandboxProvider` (`packages/harness/deerflow/community/`) - Docker-based isolation
 
 **Virtual Path System**:
 - Agent sees: `/mnt/user-data/{workspace,uploads,outputs}`, `/mnt/skills`
 - Physical: `backend/.deer-flow/users/{user_id}/threads/{thread_id}/user-data/...`, `deer-flow/skills/`
-- Translation: `LocalSandboxProvider` builds per-thread `PathMapping`s for the user-data prefixes at acquire time; `tools.py` keeps `replace_virtual_path()` / `replace_virtual_paths_in_command()` as a defense-in-depth layer (and for path validation). AIO has the directories volume-mounted at the same virtual paths inside its container, so both implementations accept `/mnt/user-data/...` natively.
+- Translation: `replace_virtual_path()` / `replace_virtual_paths_in_command()`
-- Detection: `is_local_sandbox()` accepts both `sandbox_id == "local"` (legacy / no-thread) and `sandbox_id.startswith("local:")` (per-thread)
+- Detection: `is_local_sandbox()` checks `sandbox_id == "local"`
 
 **Sandbox Tools** (in `packages/harness/deerflow/sandbox/tools.py`):
 - `bash` - Execute commands with path translation and error handling
 - `ls` - Directory listing (tree format, max 2 levels)
 - `read_file` - Read file contents with optional line range
-- `write_file` - Write/append to files, creates directories; overwrites by default and exposes the `append` argument in the model-facing schema for end-of-file writes
+- `write_file` - Write/append to files, creates directories
 - `str_replace` - Substring replacement (single or all occurrences); same-path serialization is scoped to `(sandbox.id, path)` so isolated sandboxes do not contend on identical virtual paths inside one process
 
 ### Subagent System (`packages/harness/deerflow/subagents/`)
@@ -306,7 +253,6 @@ Proxied through nginx: `/api/langgraph/*` → Gateway LangGraph-compatible runti
 **Concurrency**: `MAX_CONCURRENT_SUBAGENTS = 3` enforced by `SubagentLimitMiddleware` (truncates excess tool calls in `after_model`), 15-minute timeout
 **Flow**: `task()` tool → `SubagentExecutor` → background thread → poll 5s → SSE events → result
 **Events**: `task_started`, `task_running`, `task_completed`/`task_failed`/`task_timed_out`
-**Deferred MCP tools** (if `tool_search.enabled`): `SubagentExecutor._build_initial_state` assembles deferral after policy filtering via the shared `assemble_deferred_tools` (fail-closed), appends the `tool_search` tool, injects the `<available-deferred-tools>` section into the subagent's `SystemMessage`, and threads the setup to `_create_agent`, which attaches `DeferredToolFilterMiddleware` through `build_subagent_runtime_middlewares(deferred_setup=...)`. Subagents thus withhold full MCP schemas until promotion, same as the lead agent; each task run gets a fresh `ThreadState` so promotion is isolated per run
 
 ### Tool System (`packages/harness/deerflow/tools/`)
 
@@ -317,10 +263,8 @@ Proxied through nginx: `/api/langgraph/*` → Gateway LangGraph-compatible runti
    - `present_files` - Make output files visible to user (only `/mnt/user-data/outputs`)
    - `ask_clarification` - Request clarification (intercepted by ClarificationMiddleware → interrupts)
    - `view_image` - Read image as base64 (added only if model supports vision)
-   - `setup_agent` - Bootstrap-only: persist a brand-new custom agent's `SOUL.md` and `config.yaml`. Bound only when `is_bootstrap=True`.
-   - `update_agent` - Custom-agent-only: persist self-updates to the current agent's `SOUL.md` / `config.yaml` from inside a normal chat (partial update + atomic write). Bound when `agent_name` is set and `is_bootstrap=False`.
 4. **Subagent tool** (if enabled):
-   - `task` - Delegate to subagent (description, prompt, subagent_type)
+   - `task` - Delegate to subagent (description, prompt, subagent_type, max_turns)
 
 **Community tools** (`packages/harness/deerflow/community/`):
 - `tavily/` - Web search (5 results default) and web fetch (4KB limit)
@@ -341,7 +285,7 @@ Proxied through nginx: `/api/langgraph/*` → Gateway LangGraph-compatible runti
 - **Cache invalidation**: Detects config file changes via mtime comparison
 - **Transports**: stdio (command-based), SSE, HTTP
 - **OAuth (HTTP/SSE)**: Supports token endpoint flows (`client_credentials`, `refresh_token`) with automatic token refresh + Authorization header injection
-- **Runtime updates**: Gateway API saves to extensions_config.json; the Gateway-embedded runtime detects changes via mtime
+- **Runtime updates**: Gateway API saves to extensions_config.json; LangGraph detects via mtime
 
 ### Skills System (`packages/harness/deerflow/skills/`)
 
@@ -349,7 +293,6 @@ Proxied through nginx: `/api/langgraph/*` → Gateway LangGraph-compatible runti
 - **Format**: Directory with `SKILL.md` (YAML frontmatter: name, description, license, allowed-tools)
 - **Loading**: `load_skills()` recursively scans `skills/{public,custom}` for `SKILL.md`, parses metadata, and reads enabled state from extensions_config.json
 - **Injection**: Enabled skills listed in agent system prompt with container paths
-- **Slash activation**: `/skill-name task` loads that enabled skill's `SKILL.md` for the current model call only. The resolver rejects leading whitespace, missing separators, reserved channel commands (`/new`, `/help`, `/bootstrap`, `/status`, `/models`, `/memory`), disabled skills, and skills outside a custom agent's whitelist.
 - **Installation**: `POST /api/skills/install` extracts .skill ZIP archive to custom/ directory
 
 ### Model Factory (`packages/harness/deerflow/models/factory.py`)
@@ -369,7 +312,7 @@ Proxied through nginx: `/api/langgraph/*` → Gateway LangGraph-compatible runti
 
 ### IM Channels System (`app/channels/`)
 
-Bridges external messaging platforms (Feishu, Slack, Telegram, DingTalk) to the DeerFlow agent via Gateway's LangGraph-compatible API.
+Bridges external messaging platforms (Feishu, Slack, Telegram, DingTalk) to the DeerFlow agent via the LangGraph Server.
 
 
 **Architecture**: Channels communicate with Gateway through the `langgraph-sdk` HTTP client (same as the frontend), ensuring threads are created and managed server-side. The internal SDK client injects process-local internal auth plus a matching CSRF cookie/header pair so Gateway accepts state-changing thread/run requests from channel workers without relying on browser session cookies.
@@ -411,11 +354,10 @@ Bridges external messaging platforms (Feishu, Slack, Telegram, DingTalk) to the
 **Per-User Isolation**:
 - Memory is stored per-user at `{base_dir}/users/{user_id}/memory.json`
 - Per-agent per-user memory at `{base_dir}/users/{user_id}/agents/{agent_name}/memory.json`
-- Custom agent definitions (`SOUL.md` + `config.yaml`) are also per-user at `{base_dir}/users/{user_id}/agents/{agent_name}/`. The legacy shared layout `{base_dir}/agents/{agent_name}/` remains read-only fallback for unmigrated installations
 - `user_id` is resolved via `get_effective_user_id()` from `deerflow.runtime.user_context`
 - In no-auth mode, `user_id` defaults to `"default"` (constant `DEFAULT_USER_ID`)
 - Absolute `storage_path` in config opts out of per-user isolation
-- **Migration**: Run `PYTHONPATH=. python scripts/migrate_user_isolation.py` to move legacy `memory.json`, `threads/`, and `agents/` into per-user layout. Supports `--dry-run` (preview changes) and `--user-id USER_ID` (assign unowned legacy data to a user, defaults to `default`).
+- **Migration**: Run `PYTHONPATH=. python scripts/migrate_user_isolation.py` to move legacy `memory.json` and `threads/` into per-user layout; supports `--dry-run`
 
 **Data Structure** (stored in `{base_dir}/users/{user_id}/memory.json`):
 - **User Context**: `workContext`, `personalContext`, `topOfMind` (1-3 sentence summaries)
@@ -444,24 +386,6 @@ Focused regression coverage for the updater lives in `backend/tests/test_memory_
 - `resolve_variable(path)` - Import module and return variable (e.g., `module.path:variable_name`)
 - `resolve_class(path, base_class)` - Import and validate class against base class
 
-### Tracing System (`packages/harness/deerflow/tracing/`)
-
-LangSmith and Langfuse are both supported. The wiring lives in two layers:
-
-- `factory.py::build_tracing_callbacks()` — returns the LangChain `CallbackHandler` list for the providers currently enabled via env vars (`LANGSMITH_TRACING`, `LANGFUSE_TRACING`, etc.). The handlers are attached at the **graph invocation root** for in-graph runs (`make_lead_agent` and `DeerFlowClient.stream` both append them to `config["callbacks"]` before invoking the graph) so a single run produces one trace with all node / LLM / tool calls as child spans. Standalone callers — anything that invokes a model outside such a graph (e.g. `MemoryUpdater`) — keep `create_chat_model`'s default `attach_tracing=True`, which falls back to model-level callback attachment.
-- `metadata.py::build_langfuse_trace_metadata()` — builds the Langfuse-reserved trace attributes for `RunnableConfig.metadata`. The Langfuse v4 `langchain.CallbackHandler` lifts these onto the root trace (see its `_parse_langfuse_trace_attributes`), but only when it sees `on_chain_start(parent_run_id=None)` — which is why the callbacks have to live at the graph root, not the model.
-
-**Trace-attribute injection points**: both `runtime/runs/worker.py::run_agent` (gateway path) and `client.py::DeerFlowClient.stream` (embedded path) merge the metadata into `config["metadata"]` right before constructing the graph. Caller-supplied keys win via `setdefault`, so an external `session_id` override is preserved. Field mapping:
-
-| Langfuse field         | Source                                       |
-|-----------------------|----------------------------------------------|
-| `langfuse_session_id` | LangGraph `thread_id`                         |
-| `langfuse_user_id`    | `get_effective_user_id()` (`default` in no-auth) |
-| `langfuse_trace_name` | `RunRecord.assistant_id` / client `agent_name` (defaults to `lead-agent`) |
-| `langfuse_tags`       | `env:<DEER_FLOW_ENV>` + `model:<model_name>`  |
-
-Returns `{}` when Langfuse is not in the enabled providers — LangSmith-only deployments are unaffected. Set `DEER_FLOW_ENV` (or `ENVIRONMENT`) to tag traces by deployment environment. Tests live in `tests/test_tracing_factory.py`, `tests/test_tracing_metadata.py`, `tests/test_worker_langfuse_metadata.py`, and `tests/test_client_langfuse_metadata.py`.
-
 ### Config Schema
 
 **`config.yaml`** key sections:
@@ -495,7 +419,7 @@ Both can be modified at runtime via Gateway API endpoints or `DeerFlowClient` me
   - `"messages-tuple"` — per-chunk update: for AI text this is a **delta** (concat per `id` to rebuild the full message); tool calls and tool results are emitted once each
   - `"custom"` — forwarded from `StreamWriter`
   - `"end"` — stream finished (carries cumulative `usage` counted once per message id)
-- Agent created lazily via `create_agent()` + `build_middlewares()`, same as `make_lead_agent`
+- Agent created lazily via `create_agent()` + `_build_middlewares()`, same as `make_lead_agent`
 - Supports `checkpointer` parameter for state persistence across turns
 - `reset_agent()` forces agent recreation (e.g. after memory or skill changes)
 - See [docs/STREAMING.md](docs/STREAMING.md) for the full design: why Gateway and DeerFlowClient are parallel paths, LangGraph's `stream_mode` semantics, the per-id dedup invariants, and regression testing strategy
@@ -593,7 +517,6 @@ Multi-file upload with automatic document conversion:
 - Rejects directory inputs before copying so uploads stay all-or-nothing
 - Reuses one conversion worker per request when called from an active event loop
 - Files stored in thread-isolated directories
-- Duplicate filenames in a single upload request are auto-renamed with `_N` suffixes so later files do not truncate earlier files
 - Agent receives uploaded file list via `UploadsMiddleware`
 
 See [docs/FILE_UPLOAD.md](docs/FILE_UPLOAD.md) for details.

backend/CONTRIBUTING.md

@@ -56,8 +56,11 @@ export OPENAI_API_KEY="your-api-key"
 ### Run the Development Server
 
 ```bash
-# Gateway API + embedded agent runtime
+# Terminal 1: LangGraph server
 make dev
+
+# Terminal 2: Gateway API
+make gateway
 ```
 
 ## Project Structure

backend/Dockerfile

@@ -50,12 +50,6 @@ COPY backend ./backend
 RUN --mount=type=cache,target=/root/.cache/uv \
     sh -c "cd backend && UV_INDEX_URL=${UV_INDEX_URL:-https://pypi.org/simple} uv sync ${UV_EXTRAS:+--extra $UV_EXTRAS}"
 
-# UTF-8 locale prevents UnicodeEncodeError on Chinese/emoji content in minimal
-# containers where locale configuration may be missing and the default encoding is not UTF-8.
-ENV LANG=C.UTF-8
-ENV LC_ALL=C.UTF-8
-ENV PYTHONIOENCODING=utf-8
-
 # ── Stage 2: Dev ──────────────────────────────────────────────────────────────
 # Retains compiler toolchain from builder so startup-time `uv sync` can build
 # source distributions in development containers.
@@ -64,7 +58,7 @@ FROM builder AS dev
 # Install Docker CLI (for DooD: allows starting sandbox containers via host Docker socket)
 COPY --from=docker:cli /usr/local/bin/docker /usr/local/bin/docker
 
-EXPOSE 8001
+EXPOSE 8001 2024
 
 CMD ["sh", "-c", "cd backend && PYTHONPATH=. uv run uvicorn app.gateway.app:app --host 0.0.0.0 --port 8001"]
 
@@ -72,10 +66,6 @@ CMD ["sh", "-c", "cd backend && PYTHONPATH=. uv run uvicorn app.gateway.app:app
 # Clean image without build-essential — reduces size (~200 MB) and attack surface.
 FROM python:3.12-slim-bookworm
 
-ENV LANG=C.UTF-8
-ENV LC_ALL=C.UTF-8
-ENV PYTHONIOENCODING=utf-8
-
 # Copy Node.js runtime from builder (provides npx for MCP servers)
 COPY --from=builder /usr/bin/node /usr/bin/node
 COPY --from=builder /usr/lib/node_modules /usr/lib/node_modules
@@ -94,8 +84,8 @@ WORKDIR /app
 # Copy backend with pre-built virtualenv from builder
 COPY --from=builder /app/backend ./backend
 
-# Expose Gateway API port.
+# Expose ports (gateway: 8001, langgraph: 2024)
-EXPOSE 8001
+EXPOSE 8001 2024
 
 # Default command (can be overridden in docker-compose)
 CMD ["sh", "-c", "cd backend && PYTHONPATH=. uv run --no-sync uvicorn app.gateway.app:app --host 0.0.0.0 --port 8001"]

backend/Makefile

@@ -2,16 +2,13 @@ install:
 	uv sync
 
 dev:
-	PYTHONPATH=. PYTHONIOENCODING=utf-8 PYTHONUTF8=1 uv run uvicorn app.gateway.app:app --host 0.0.0.0 --port 8001 --reload
+	PYTHONPATH=. uv run uvicorn app.gateway.app:app --host 0.0.0.0 --port 8001 --reload
 
 gateway:
-	PYTHONPATH=. PYTHONIOENCODING=utf-8 PYTHONUTF8=1 uv run uvicorn app.gateway.app:app --host 0.0.0.0 --port 8001
+	PYTHONPATH=. uv run uvicorn app.gateway.app:app --host 0.0.0.0 --port 8001
 
 test:
-	PYTHONPATH=. PYTHONIOENCODING=utf-8 PYTHONUTF8=1 uv run pytest tests/ -v
+	PYTHONPATH=. uv run pytest tests/ -v
-
-test-blocking-io:
-	PYTHONPATH=. PYTHONIOENCODING=utf-8 PYTHONUTF8=1 uv run pytest tests/blocking_io -q --tb=short
 
 lint:
 	uvx ruff check .
@@ -19,6 +16,3 @@ lint:
 
 format:
 	uvx ruff check . --fix && uvx ruff format .
-
-detect-blocking-io:
-	@PYTHONPATH=. PYTHONIOENCODING=utf-8 PYTHONUTF8=1 uv run python ../scripts/detect_blocking_io_static.py --output ../.deer-flow/blocking-io-findings.json

backend/README.md

@@ -11,26 +11,31 @@ DeerFlow is a LangGraph-based AI super agent with sandbox execution, persistent
                         │          Nginx (Port 2026)           │
                         │      Unified reverse proxy           │
                         └───────┬──────────────────┬───────────┘
-                                │
+                                │                  │
-            /api/langgraph/*    │    /api/* (other)
+              /api/langgraph/*  │                  │  /api/* (other)
-            rewritten to /api/* │
+                                ▼                  ▼
-                                ▼
+               ┌────────────────────┐  ┌────────────────────────┐
-               ┌────────────────────────────────────────┐
+               │ LangGraph Server   │  │   Gateway API (8001)   │
-               │        Gateway API (8001)              │
+               │    (Port 2024)     │  │   FastAPI REST         │
-               │        FastAPI REST + agent runtime    │
+               │                    │  │                        │
-               │                                        │
+               │ ┌────────────────┐ │  │ Models, MCP, Skills,   │
-               │ Models, MCP, Skills, Memory, Uploads,  │
+               │ │  Lead Agent    │ │  │ Memory, Uploads,       │
-               │ Artifacts, Threads, Runs, Streaming    │
+               │ │  ┌──────────┐  │ │  │ Artifacts              │
-               │                                        │
+               │ │  │Middleware│  │ │  └────────────────────────┘
-               │ ┌────────────────────────────────────┐ │
+               │ │  │  Chain   │  │ │
-               │ │ Lead Agent                         │ │
+               │ │  └──────────┘  │ │
-               │ │ Middleware Chain, Tools, Subagents │ │
+               │ │  ┌──────────┐  │ │
-               │ └────────────────────────────────────┘ │
+               │ │  │  Tools   │  │ │
-               └────────────────────────────────────────┘
+               │ │  └──────────┘  │ │
+               │ │  ┌──────────┐  │ │
+               │ │  │Subagents │  │ │
+               │ │  └──────────┘  │ │
+               │ └────────────────┘ │
+               └────────────────────┘
 ```
 
 **Request Routing** (via Nginx):
-- `/api/langgraph/*` → Gateway LangGraph-compatible API - agent interactions, threads, streaming
+- `/api/langgraph/*` → LangGraph Server - agent interactions, threads, streaming
 - `/api/*` (other) → Gateway API - models, MCP, skills, memory, artifacts, uploads, thread-local cleanup
 - `/` (non-API) → Frontend - Next.js web interface
 
@@ -69,12 +74,12 @@ Middlewares execute in strict order, each handling a specific concern:
 Per-thread isolated execution with virtual path translation:
 
 - **Abstract interface**: `execute_command`, `read_file`, `write_file`, `list_dir`
-- **Providers**: `LocalSandboxProvider` (filesystem) and `AioSandboxProvider` (Docker, in community/). Async runtime paths use async sandbox lifecycle hooks so startup, readiness polling, and release do not block the event loop.
+- **Providers**: `LocalSandboxProvider` (filesystem) and `AioSandboxProvider` (Docker, in community/)
 - **Virtual paths**: `/mnt/user-data/{workspace,uploads,outputs}` → thread-specific physical directories
 - **Skills path**: `/mnt/skills` → `deer-flow/skills/` directory
 - **Skills loading**: Recursively discovers nested `SKILL.md` files under `skills/{public,custom}` and preserves nested container paths
 - **File-write safety**: `str_replace` serializes read-modify-write per `(sandbox.id, path)` so isolated sandboxes keep concurrency even when virtual paths match
-- **Tools**: `bash`, `ls`, `read_file`, `write_file`, `str_replace` (`write_file` overwrites by default and exposes `append` for end-of-file writes; `bash` is disabled by default when using `LocalSandboxProvider`; use `AioSandboxProvider` for isolated shell access)
+- **Tools**: `bash`, `ls`, `read_file`, `write_file`, `str_replace` (`bash` is disabled by default when using `LocalSandboxProvider`; use `AioSandboxProvider` for isolated shell access)
 
 ### Subagent System
 
@@ -119,7 +124,7 @@ FastAPI application providing REST endpoints for frontend integration:
 | `POST /api/memory/reload` | Force memory reload |
 | `GET /api/memory/config` | Memory configuration |
 | `GET /api/memory/status` | Combined config + data |
-| `POST /api/threads/{id}/uploads` | Upload files (auto-converts PDF/PPT/Excel/Word to Markdown, rejects directory paths, auto-renames duplicate filenames in one request) |
+| `POST /api/threads/{id}/uploads` | Upload files (auto-converts PDF/PPT/Excel/Word to Markdown, rejects directory paths) |
 | `GET /api/threads/{id}/uploads/list` | List uploaded files |
 | `DELETE /api/threads/{id}` | Delete DeerFlow-managed local thread data after LangGraph thread deletion; unexpected failures are logged server-side and return a generic 500 detail |
 | `GET /api/threads/{id}/artifacts/{path}` | Serve generated artifacts |
@@ -188,7 +193,7 @@ export OPENAI_API_KEY="your-api-key-here"
 **Full Application** (from project root):
 
 ```bash
-make dev  # Starts Gateway + Frontend + Nginx
+make dev  # Starts LangGraph + Gateway + Frontend + Nginx
 ```
 
 Access at: http://localhost:2026
@@ -196,11 +201,14 @@ Access at: http://localhost:2026
 **Backend Only** (from backend directory):
 
 ```bash
-# Gateway API + embedded agent runtime
+# Terminal 1: LangGraph server
 make dev
+
+# Terminal 2: Gateway API
+make gateway
 ```
 
-Direct access: Gateway at http://localhost:8001
+Direct access: LangGraph at http://localhost:2024, Gateway at http://localhost:8001
 
 ---
 
@@ -236,16 +244,12 @@ backend/
 │   └── utils/                  # Utilities
 ├── docs/                       # Documentation
 ├── tests/                      # Test suite
-├── langgraph.json              # LangGraph graph registry for tooling/Studio compatibility
+├── langgraph.json              # LangGraph server configuration
 ├── pyproject.toml              # Python dependencies
 ├── Makefile                    # Development commands
 └── Dockerfile                  # Container build
 ```
 
-`langgraph.json` is not the default service entrypoint.  The scripts and Docker
-deployments run the Gateway embedded runtime; the file is kept for LangGraph
-tooling, Studio, or direct LangGraph Server compatibility.
-
 ---
 
 ## Configuration
@@ -358,11 +362,10 @@ If a provider is explicitly enabled but required credentials are missing, or the
 
 ```bash
 make install    # Install dependencies
-make dev        # Run Gateway API + embedded agent runtime (port 8001)
+make dev        # Run LangGraph server (port 2024)
-make gateway    # Run Gateway API without reload (port 8001)
+make gateway    # Run Gateway API (port 8001)
 make lint       # Run linter (ruff)
 make format     # Format code (ruff)
-make detect-blocking-io  # Inventory blocking IO that may block the backend event loop
 ```
 
 ### Code Style
@@ -379,18 +382,6 @@ make detect-blocking-io  # Inventory blocking IO that may block the backend even
 uv run pytest
 ```
 
-`make detect-blocking-io` statically scans backend business code for blocking
-IO that may run on the backend event loop and is not test-coverage-bound. It
-prints a concise summary for human review and writes complete JSON findings to
-`.deer-flow/blocking-io-findings.json` at the repository root (regardless of
-whether the target is invoked from the repo root or from `backend/`). JSON
-findings include both broad IO category and review-oriented fields such as
-`priority`, `location`, `blocking_call`, `event_loop_exposure`, `reason`, and
-`code`. `priority` is a deterministic review ordering from the operation type,
-not proof of a bug. Bare-name same-file calls are resolved by function name,
-so duplicate helper names in one file can conservatively over-report async
-reachability.
-
 ---
 
 ## Technology Stack

backend/app/channels/commands.py

@@ -18,10 +18,3 @@ KNOWN_CHANNEL_COMMANDS: frozenset[str] = frozenset(
         "/help",
     }
 )
-
-
-def is_known_channel_command(text: str) -> bool:
-    """Return whether text starts with a registered channel control command."""
-    if not text.startswith("/"):
-        return False
-    return text.split(maxsplit=1)[0].lower() in KNOWN_CHANNEL_COMMANDS

backend/app/channels/dingtalk.py

@@ -14,7 +14,7 @@ from typing import Any
 import httpx
 
 from app.channels.base import Channel
-from app.channels.commands import is_known_channel_command
+from app.channels.commands import KNOWN_CHANNEL_COMMANDS
 from app.channels.message_bus import InboundMessage, InboundMessageType, MessageBus, OutboundMessage, ResolvedAttachment
 
 logger = logging.getLogger(__name__)
@@ -59,7 +59,9 @@ def _normalize_allowed_users(allowed_users: Any) -> set[str]:
 
 
 def _is_dingtalk_command(text: str) -> bool:
-    return is_known_channel_command(text)
+    if not text.startswith("/"):
+        return False
+    return text.split(maxsplit=1)[0].lower() in KNOWN_CHANNEL_COMMANDS
 
 
 def _extract_text_from_rich_text(rich_text_list: list) -> str:

backend/app/channels/discord.py

@@ -3,14 +3,11 @@
 from __future__ import annotations
 
 import asyncio
-import json
 import logging
 import threading
-from pathlib import Path
 from typing import Any
 
 from app.channels.base import Channel
-from app.channels.commands import is_known_channel_command
 from app.channels.message_bus import InboundMessageType, MessageBus, OutboundMessage, ResolvedAttachment
 
 logger = logging.getLogger(__name__)
@@ -24,12 +21,6 @@ class DiscordChannel(Channel):
     Configuration keys (in ``config.yaml`` under ``channels.discord``):
         - ``bot_token``: Discord Bot token.
         - ``allowed_guilds``: (optional) List of allowed Discord guild IDs. Empty = allow all.
-        - ``mention_only``: (optional) If true, only respond when the bot is mentioned.
-        - ``allowed_channels``: (optional) List of channel IDs where messages are always accepted
-          (even when mention_only is true). Use for channels where you want the bot to respond
-          without mentions. Empty = mention_only applies everywhere.
-        - ``thread_mode``: (optional) If true, group a channel conversation into a thread.
-          Default: same as ``mention_only``.
     """
 
     def __init__(self, bus: MessageBus, config: dict[str, Any]) -> None:
@@ -41,29 +32,6 @@ class DiscordChannel(Channel):
                 self._allowed_guilds.add(int(guild_id))
             except (TypeError, ValueError):
                 continue
-        self._mention_only: bool = bool(config.get("mention_only", False))
-        self._thread_mode: bool = config.get("thread_mode", self._mention_only)
-        self._allowed_channels: set[str] = set()
-        for channel_id in config.get("allowed_channels", []):
-            self._allowed_channels.add(str(channel_id))
-
-        # Session tracking: channel_id -> Discord thread_id (in-memory, persisted to JSON).
-        # Uses a dedicated JSON file separate from ChannelStore, which maps IM
-        # conversations to DeerFlow thread IDs — a different concern.
-        self._active_threads: dict[str, str] = {}
-        # Reverse-lookup set for O(1) thread ID checks (avoids O(n) scan of _active_threads.values()).
-        self._active_thread_ids: set[str] = set()
-        # Lock protecting _active_threads and the JSON file from concurrent access.
-        # _run_client (Discord loop thread) and the main thread both read/write.
-        self._thread_store_lock = threading.Lock()
-        store = config.get("channel_store")
-        if store is not None:
-            self._thread_store_path = store._path.parent / "discord_threads.json"
-        else:
-            self._thread_store_path = Path.home() / ".deer-flow" / "channels" / "discord_threads.json"
-
-        # Typing indicator management
-        self._typing_tasks: dict[str, asyncio.Task] = {}
 
         self._client = None
         self._thread: threading.Thread | None = None
@@ -107,56 +75,12 @@ class DiscordChannel(Channel):
 
         self._thread = threading.Thread(target=self._run_client, daemon=True)
         self._thread.start()
-        self._load_active_threads()
         logger.info("Discord channel started")
 
-    def _load_active_threads(self) -> None:
-        """Restore Discord thread mappings from the dedicated JSON file on startup."""
-        with self._thread_store_lock:
-            try:
-                if not self._thread_store_path.exists():
-                    logger.debug("[Discord] no thread mappings file at %s", self._thread_store_path)
-                    return
-                data = json.loads(self._thread_store_path.read_text())
-                self._active_threads.clear()
-                self._active_thread_ids.clear()
-                for channel_id, thread_id in data.items():
-                    self._active_threads[channel_id] = thread_id
-                    self._active_thread_ids.add(thread_id)
-                if self._active_threads:
-                    logger.info("[Discord] restored %d thread mappings from %s", len(self._active_threads), self._thread_store_path)
-            except Exception:
-                logger.exception("[Discord] failed to load thread mappings")
-
-    def _save_thread(self, channel_id: str, thread_id: str) -> None:
-        """Persist a Discord thread mapping to the dedicated JSON file."""
-        with self._thread_store_lock:
-            try:
-                data: dict[str, str] = {}
-                if self._thread_store_path.exists():
-                    data = json.loads(self._thread_store_path.read_text())
-                old_id = data.get(channel_id)
-                data[channel_id] = thread_id
-                # Update reverse-lookup set
-                if old_id:
-                    self._active_thread_ids.discard(old_id)
-                self._active_thread_ids.add(thread_id)
-                self._thread_store_path.parent.mkdir(parents=True, exist_ok=True)
-                self._thread_store_path.write_text(json.dumps(data, indent=2))
-            except Exception:
-                logger.exception("[Discord] failed to save thread mapping for channel %s", channel_id)
-
     async def stop(self) -> None:
         self._running = False
         self.bus.unsubscribe_outbound(self._on_outbound)
 
-        # Cancel all active typing indicator tasks
-        for target_id, task in list(self._typing_tasks.items()):
-            if not task.done():
-                task.cancel()
-            logger.debug("[Discord] cancelled typing task for target %s", target_id)
-        self._typing_tasks.clear()
-
         if self._client and self._discord_loop and self._discord_loop.is_running():
             close_future = asyncio.run_coroutine_threadsafe(self._client.close(), self._discord_loop)
             try:
@@ -176,10 +100,6 @@ class DiscordChannel(Channel):
         logger.info("Discord channel stopped")
 
     async def send(self, msg: OutboundMessage) -> None:
-        # Stop typing indicator once we're sending the response
-        stop_future = asyncio.run_coroutine_threadsafe(self._stop_typing(msg.chat_id, msg.thread_ts), self._discord_loop)
-        await asyncio.wrap_future(stop_future)
-
         target = await self._resolve_target(msg)
         if target is None:
             logger.error("[Discord] target not found for chat_id=%s thread_ts=%s", msg.chat_id, msg.thread_ts)
@@ -191,9 +111,6 @@ class DiscordChannel(Channel):
             await asyncio.wrap_future(send_future)
 
     async def send_file(self, msg: OutboundMessage, attachment: ResolvedAttachment) -> bool:
-        stop_future = asyncio.run_coroutine_threadsafe(self._stop_typing(msg.chat_id, msg.thread_ts), self._discord_loop)
-        await asyncio.wrap_future(stop_future)
-
         target = await self._resolve_target(msg)
         if target is None:
             logger.error("[Discord] target not found for file upload chat_id=%s thread_ts=%s", msg.chat_id, msg.thread_ts)
@@ -213,41 +130,6 @@ class DiscordChannel(Channel):
             logger.exception("[Discord] failed to upload file: %s", attachment.filename)
             return False
 
-    async def _start_typing(self, channel, chat_id: str, thread_ts: str | None = None) -> None:
-        """Starts a loop to send periodic typing indicators."""
-        target_id = thread_ts or chat_id
-        if target_id in self._typing_tasks:
-            return  # Already typing for this target
-
-        async def _typing_loop():
-            try:
-                while True:
-                    try:
-                        await channel.trigger_typing()
-                    except Exception:
-                        pass
-                    await asyncio.sleep(10)
-            except asyncio.CancelledError:
-                pass
-
-        task = asyncio.create_task(_typing_loop())
-        self._typing_tasks[target_id] = task
-
-    async def _stop_typing(self, chat_id: str, thread_ts: str | None = None) -> None:
-        """Stops the typing loop for a specific target."""
-        target_id = thread_ts or chat_id
-        task = self._typing_tasks.pop(target_id, None)
-        if task and not task.done():
-            task.cancel()
-            logger.debug("[Discord] stopped typing indicator for target %s", target_id)
-
-    async def _add_reaction(self, message) -> None:
-        """Add a checkmark reaction to acknowledge the message was received."""
-        try:
-            await message.add_reaction("✅")
-        except Exception:
-            logger.debug("[Discord] failed to add reaction to message %s", message.id, exc_info=True)
-
     async def _on_message(self, message) -> None:
         if not self._running or not self._client:
             return
@@ -270,145 +152,17 @@ class DiscordChannel(Channel):
         if self._discord_module is None:
             return
 
-        # Determine whether the bot is mentioned in this message
-        user = self._client.user if self._client else None
-        if user:
-            bot_mention = user.mention  # <@ID>
-            alt_mention = f"<@!{user.id}>"  # <@!ID> (ping variant)
-            standard_mention = f"<@{user.id}>"
-        else:
-            bot_mention = None
-            alt_mention = None
-            standard_mention = ""
-        has_mention = (bot_mention and bot_mention in message.content) or (alt_mention and alt_mention in message.content) or (standard_mention and standard_mention in message.content)
-
-        # Strip mention from text for processing
-        if has_mention:
-            text = text.replace(bot_mention or "", "").replace(alt_mention or "", "").replace(standard_mention or "", "").strip()
-            # Don't return early if text is empty — still process the mention (e.g., create thread)
-
-        # --- Determine thread/channel routing and typing target ---
-        thread_id = None
-        chat_id = None
-        typing_target = None  # The Discord object to type into
-
         if isinstance(message.channel, self._discord_module.Thread):
-            # --- Message already inside a thread ---
+            chat_id = str(message.channel.parent_id or message.channel.id)
-            thread_obj = message.channel
+            thread_id = str(message.channel.id)
-            thread_id = str(thread_obj.id)
-            chat_id = str(thread_obj.parent_id or thread_obj.id)
-            typing_target = thread_obj
-
-            # If this is a known active thread, process normally
-            if thread_id in self._active_thread_ids:
-                msg_type = InboundMessageType.COMMAND if is_known_channel_command(text) else InboundMessageType.CHAT
-                inbound = self._make_inbound(
-                    chat_id=chat_id,
-                    user_id=str(message.author.id),
-                    text=text,
-                    msg_type=msg_type,
-                    thread_ts=thread_id,
-                    metadata={
-                        "guild_id": str(guild.id) if guild else None,
-                        "channel_id": str(message.channel.id),
-                        "message_id": str(message.id),
-                    },
-                )
-                inbound.topic_id = thread_id
-                self._publish(inbound)
-                # Start typing indicator in the thread
-                if typing_target:
-                    asyncio.create_task(self._start_typing(typing_target, chat_id, thread_id))
-                asyncio.create_task(self._add_reaction(message))
-                return
-
-            # Thread not tracked (orphaned) — create new thread and handle below
-            logger.debug("[Discord] message in orphaned thread %s, will create new thread", thread_id)
-            thread_id = None
-            typing_target = None
-
-        # At this point we're guaranteed to be in a channel, not a thread
-        # (the Thread case is handled above). Apply mention_only for all
-        # non-thread messages — no special case needed.
-        channel_id = str(message.channel.id)
-
-        # Check if there's an active thread for this channel
-        if channel_id in self._active_threads:
-            # respect mention_only: if enabled, only process messages that mention the bot
-            # (unless the channel is in allowed_channels)
-            # Messages within a thread are always allowed through (continuation).
-            # At this code point we know the message is in a channel, not a thread
-            # (Thread case handled above), so always apply the check.
-            if self._mention_only and not has_mention and channel_id not in self._allowed_channels:
-                logger.debug("[Discord] skipping no-@ message in channel %s (not in thread)", channel_id)
-                return
-            # mention_only + fresh @ → create new thread instead of routing to existing one
-            if self._mention_only and has_mention:
-                thread_obj = await self._create_thread(message)
-                if thread_obj is not None:
-                    target_thread_id = str(thread_obj.id)
-                    self._active_threads[channel_id] = target_thread_id
-                    self._save_thread(channel_id, target_thread_id)
-                    thread_id = target_thread_id
-                    chat_id = channel_id
-                    typing_target = thread_obj
-                    logger.info("[Discord] created new thread %s in channel %s on mention (replacing existing thread)", target_thread_id, channel_id)
-                else:
-                    logger.info("[Discord] thread creation failed in channel %s, falling back to channel replies", channel_id)
-                    thread_id = channel_id
-                    chat_id = channel_id
-                    typing_target = message.channel
-            else:
-                # Existing session → route to the existing thread
-                target_thread_id = self._active_threads[channel_id]
-                logger.debug("[Discord] routing message in channel %s to existing thread %s", channel_id, target_thread_id)
-                thread_id = target_thread_id
-                chat_id = channel_id
-                typing_target = await self._get_channel_or_thread(target_thread_id)
-        elif self._mention_only and not has_mention and channel_id not in self._allowed_channels:
-            # Not mentioned and not in an allowed channel → skip
-            logger.debug("[Discord] skipping message without mention in channel %s", channel_id)
-            return
-        elif self._mention_only and has_mention:
-            # First mention in this channel → create thread
-            thread_obj = await self._create_thread(message)
-            if thread_obj is not None:
-                target_thread_id = str(thread_obj.id)
-                self._active_threads[channel_id] = target_thread_id
-                self._save_thread(channel_id, target_thread_id)
-                thread_id = target_thread_id
-                chat_id = channel_id
-                typing_target = thread_obj  # Type into the new thread
-                logger.info("[Discord] created thread %s in channel %s for user %s", target_thread_id, channel_id, message.author.display_name)
-            else:
-                # Fallback: thread creation failed (disabled/permissions), reply in channel
-                logger.info("[Discord] thread creation failed in channel %s, falling back to channel replies", channel_id)
-                thread_id = channel_id
-                chat_id = channel_id
-                typing_target = message.channel  # Type into the channel
-        elif self._thread_mode:
-            # thread_mode but mention_only is False → create thread anyway for conversation grouping
-            thread_obj = await self._create_thread(message)
-            if thread_obj is None:
-                # Thread creation failed (disabled/permissions), fall back to channel replies
-                logger.info("[Discord] thread creation failed in channel %s, falling back to channel replies", channel_id)
-                thread_id = channel_id
-                chat_id = channel_id
-                typing_target = message.channel  # Type into the channel
-            else:
-                target_thread_id = str(thread_obj.id)
-                self._active_threads[channel_id] = target_thread_id
-                self._save_thread(channel_id, target_thread_id)
-                thread_id = target_thread_id
-                chat_id = channel_id
-                typing_target = thread_obj  # Type into the new thread
         else:
-            # No threading — reply directly in channel
+            thread = await self._create_thread(message)
-            thread_id = channel_id
+            if thread is None:
-            chat_id = channel_id
+                return
-            typing_target = message.channel  # Type into the channel
+            chat_id = str(message.channel.id)
+            thread_id = str(thread.id)
 
-        msg_type = InboundMessageType.COMMAND if is_known_channel_command(text) else InboundMessageType.CHAT
+        msg_type = InboundMessageType.COMMAND if text.startswith("/") else InboundMessageType.CHAT
         inbound = self._make_inbound(
             chat_id=chat_id,
             user_id=str(message.author.id),
@@ -423,15 +177,6 @@ class DiscordChannel(Channel):
         )
         inbound.topic_id = thread_id
 
-        # Start typing indicator in the correct target (thread or channel)
-        if typing_target:
-            asyncio.create_task(self._start_typing(typing_target, chat_id, thread_id))
-
-        self._publish(inbound)
-        asyncio.create_task(self._add_reaction(message))
-
-    def _publish(self, inbound) -> None:
-        """Publish an inbound message to the main event loop."""
         if self._main_loop and self._main_loop.is_running():
             future = asyncio.run_coroutine_threadsafe(self.bus.publish_inbound(inbound), self._main_loop)
             future.add_done_callback(lambda f: logger.exception("[Discord] publish_inbound failed", exc_info=f.exception()) if f.exception() else None)
@@ -453,40 +198,14 @@ class DiscordChannel(Channel):
 
     async def _create_thread(self, message):
         try:
-            if self._discord_module is None:
-                return None
-
-            # Only TextChannel (type 0) and NewsChannel (type 10) support threads
-            channel_type = message.channel.type
-            if channel_type not in (
-                self._discord_module.ChannelType.text,
-                self._discord_module.ChannelType.news,
-            ):
-                logger.info(
-                    "[Discord] channel type %s (%s) does not support threads",
-                    channel_type.value,
-                    channel_type.name,
-                )
-                return None
-
             thread_name = f"deerflow-{message.author.display_name}-{message.id}"[:100]
             return await message.create_thread(name=thread_name)
-        except self._discord_module.errors.HTTPException as exc:
-            if exc.code == 50024:
-                logger.info(
-                    "[Discord] cannot create thread in channel %s (error code 50024): %s",
-                    message.channel.id,
-                    channel_type.name if (channel_type := message.channel.type) else "unknown",
-                )
-            else:
-                logger.exception(
-                    "[Discord] failed to create thread for message=%s (HTTPException %s)",
-                    message.id,
-                    exc.code,
-                )
-            return None
         except Exception:
             logger.exception("[Discord] failed to create thread for message=%s (threads may be disabled or missing permissions)", message.id)
+            try:
+                await message.channel.send("Could not create a thread for your message. Please check that threads are enabled in this channel.")
+            except Exception:
+                pass
             return None
 
     async def _resolve_target(self, msg: OutboundMessage):

backend/app/channels/feishu.py

@@ -7,30 +7,22 @@ import json
 import logging
 import re
 import threading
-import time
 from typing import Any, Literal
 
 from app.channels.base import Channel
-from app.channels.commands import is_known_channel_command
+from app.channels.commands import KNOWN_CHANNEL_COMMANDS
-from app.channels.message_bus import (
+from app.channels.message_bus import InboundMessage, InboundMessageType, MessageBus, OutboundMessage, ResolvedAttachment
-    PENDING_CLARIFICATION_METADATA_KEY,
-    RESOLVED_FROM_PENDING_CLARIFICATION_METADATA_KEY,
-    InboundMessage,
-    InboundMessageType,
-    MessageBus,
-    OutboundMessage,
-    ResolvedAttachment,
-)
 from deerflow.config.paths import VIRTUAL_PATH_PREFIX, get_paths
 from deerflow.runtime.user_context import get_effective_user_id
 from deerflow.sandbox.sandbox_provider import get_sandbox_provider
 
 logger = logging.getLogger(__name__)
-PENDING_CLARIFICATION_TTL_SECONDS = 30 * 60
 
 
 def _is_feishu_command(text: str) -> bool:
-    return is_known_channel_command(text)
+    if not text.startswith("/"):
+        return False
+    return text.split(maxsplit=1)[0].lower() in KNOWN_CHANNEL_COMMANDS
 
 
 class FeishuChannel(Channel):
@@ -64,7 +56,6 @@ class FeishuChannel(Channel):
         self._background_tasks: set[asyncio.Task] = set()
         self._running_card_ids: dict[str, str] = {}
         self._running_card_tasks: dict[str, asyncio.Task] = {}
-        self._pending_clarifications: dict[tuple[str, str], list[dict[str, Any]]] = {}
         self._CreateFileRequest = None
         self._CreateFileRequestBody = None
         self._CreateImageRequest = None
@@ -72,16 +63,6 @@ class FeishuChannel(Channel):
         self._GetMessageResourceRequest = None
         self._thread_lock = threading.Lock()
 
-    @staticmethod
-    def _non_empty_str(value: Any) -> str | None:
-        if isinstance(value, str) and value.strip():
-            return value.strip()
-        return None
-
-    @staticmethod
-    def _pending_key(chat_id: str, user_id: str) -> tuple[str, str]:
-        return (chat_id, user_id)
-
     @property
     def supports_streaming(self) -> bool:
         return True
@@ -550,25 +531,18 @@ class FeishuChannel(Channel):
                         "[Feishu] failed to patch running card %s, falling back to final reply",
                         running_card_id,
                     )
-                    fallback_card_id = await self._reply_card(source_message_id, msg.text)
+                    await self._reply_card(source_message_id, msg.text)
-                    self._remember_thread_mapping(msg, source_message_id, fallback_card_id)
-                    self._remember_pending_clarification(msg, fallback_card_id)
                 else:
-                    self._remember_thread_mapping(msg, source_message_id, running_card_id)
-                    self._remember_pending_clarification(msg, running_card_id)
                     logger.info("[Feishu] running card updated: source=%s card=%s", source_message_id, running_card_id)
             elif msg.is_final:
-                final_card_id = await self._reply_card(source_message_id, msg.text)
+                await self._reply_card(source_message_id, msg.text)
-                self._remember_thread_mapping(msg, source_message_id, final_card_id)
-                self._remember_pending_clarification(msg, final_card_id)
             elif awaited_running_card_task:
                 logger.warning(
                     "[Feishu] running card task finished without message_id for source=%s, skipping duplicate non-final creation",
                     source_message_id,
                 )
             else:
-                created_card_id = await self._ensure_running_card(source_message_id, msg.text)
+                await self._ensure_running_card(source_message_id, msg.text)
-                self._remember_thread_mapping(msg, source_message_id, created_card_id)
 
             if msg.is_final:
                 self._running_card_ids.pop(source_message_id, None)
@@ -579,129 +553,6 @@ class FeishuChannel(Channel):
 
     # -- internal ----------------------------------------------------------
 
-    def _remember_thread_mapping(self, msg: OutboundMessage, *topic_ids: str | None) -> None:
-        store = self.config.get("channel_store")
-        if store is None or not msg.thread_id:
-            return
-
-        metadata_topic_ids = [
-            msg.metadata.get("message_id"),
-            msg.metadata.get("root_id"),
-            msg.metadata.get("parent_id"),
-            msg.metadata.get("thread_id"),
-            msg.metadata.get("topic_id"),
-        ]
-        user_id = ""
-        raw_user_id = msg.metadata.get("user_id")
-        if isinstance(raw_user_id, str):
-            user_id = raw_user_id
-
-        seen: set[str] = set()
-        for topic_id in [*topic_ids, *metadata_topic_ids]:
-            topic_id = self._non_empty_str(topic_id)
-            if not topic_id or topic_id in seen:
-                continue
-            seen.add(topic_id)
-            try:
-                store.set_thread_id(
-                    self.name,
-                    msg.chat_id,
-                    msg.thread_id,
-                    topic_id=topic_id,
-                    user_id=user_id,
-                )
-            except Exception:
-                logger.exception("[Feishu] failed to remember thread mapping for topic_id=%s", topic_id)
-
-    def _remember_pending_clarification(self, msg: OutboundMessage, card_message_id: str | None) -> None:
-        if not msg.is_final or msg.metadata.get(PENDING_CLARIFICATION_METADATA_KEY) is not True:
-            return
-
-        user_id = self._non_empty_str(msg.metadata.get("user_id"))
-        topic_id = self._non_empty_str(msg.metadata.get("topic_id"))
-        source_message_id = self._non_empty_str(msg.thread_ts) or self._non_empty_str(msg.metadata.get("message_id"))
-        if not (user_id and topic_id and msg.thread_id and source_message_id and card_message_id):
-            return
-
-        key = self._pending_key(msg.chat_id, user_id)
-        pending = {
-            "thread_id": msg.thread_id,
-            "topic_id": topic_id,
-            "source_message_id": source_message_id,
-            "card_message_id": card_message_id,
-            "created_at": time.time(),
-        }
-        with self._thread_lock:
-            # Plain-message clarification continuity is a short-lived in-memory
-            # hint; explicit Feishu replies are still covered by persisted
-            # message-id mappings.
-            self._pending_clarifications.setdefault(key, []).append(pending)
-        logger.info(
-            "[Feishu] pending clarification remembered: chat_id=%s user_id=%s topic_id=%s thread_id=%s",
-            msg.chat_id,
-            user_id,
-            topic_id,
-            msg.thread_id,
-        )
-
-    def _consume_pending_clarification(self, chat_id: str, user_id: str) -> dict[str, Any] | None:
-        key = self._pending_key(chat_id, user_id)
-        with self._thread_lock:
-            pending_items = self._pending_clarifications.get(key)
-            if not pending_items:
-                return None
-
-            now = time.time()
-            while pending_items:
-                pending = pending_items.pop(0)
-                created_at = pending.get("created_at")
-                if isinstance(created_at, (int, float)) and now - created_at <= PENDING_CLARIFICATION_TTL_SECONDS:
-                    if pending_items:
-                        self._pending_clarifications[key] = pending_items
-                    else:
-                        self._pending_clarifications.pop(key, None)
-                    return pending
-                logger.info("[Feishu] pending clarification expired: chat_id=%s user_id=%s", chat_id, user_id)
-
-            self._pending_clarifications.pop(key, None)
-            return None
-
-    def _ensure_pending_thread_mapping(self, chat_id: str, user_id: str, pending: dict[str, Any]) -> None:
-        store = self.config.get("channel_store")
-        topic_id = self._non_empty_str(pending.get("topic_id"))
-        thread_id = self._non_empty_str(pending.get("thread_id"))
-        if store is None or not topic_id or not thread_id:
-            return
-        try:
-            store.set_thread_id(self.name, chat_id, thread_id, topic_id=topic_id, user_id=user_id)
-        except Exception:
-            logger.exception("[Feishu] failed to restore pending clarification mapping for topic_id=%s", topic_id)
-
-    def _resolve_topic_id(
-        self,
-        chat_id: str,
-        msg_id: str,
-        *,
-        root_id: str | None,
-        parent_id: str | None,
-        thread_id: str | None,
-    ) -> tuple[str, bool]:
-        store = self.config.get("channel_store")
-        candidates = [root_id, parent_id, thread_id]
-
-        if store is not None:
-            for candidate in candidates:
-                candidate = self._non_empty_str(candidate)
-                if not candidate:
-                    continue
-                try:
-                    if store.get_thread_id(self.name, chat_id, topic_id=candidate):
-                        return candidate, True
-                except Exception:
-                    logger.exception("[Feishu] failed to resolve stored topic mapping for topic_id=%s", candidate)
-
-        return root_id or msg_id, False
-
     @staticmethod
     def _log_future_error(fut, name: str, msg_id: str) -> None:
         """Callback for run_coroutine_threadsafe futures to surface errors."""
@@ -742,9 +593,7 @@ class FeishuChannel(Channel):
 
             # root_id is set when the message is a reply within a Feishu thread.
             # Use it as topic_id so all replies share the same DeerFlow thread.
-            root_id = self._non_empty_str(getattr(message, "root_id", None))
+            root_id = getattr(message, "root_id", None) or None
-            parent_id = self._non_empty_str(getattr(message, "parent_id", None))
-            feishu_thread_id = self._non_empty_str(getattr(message, "thread_id", None))
 
             # Parse message content
             content = json.loads(message.content)
@@ -805,12 +654,10 @@ class FeishuChannel(Channel):
             text = text.strip()
 
             logger.info(
-                "[Feishu] parsed message: chat_id=%s, msg_id=%s, root_id=%s, parent_id=%s, thread_id=%s, sender=%s, text=%r",
+                "[Feishu] parsed message: chat_id=%s, msg_id=%s, root_id=%s, sender=%s, text=%r",
                 chat_id,
                 msg_id,
                 root_id,
-                parent_id,
-                feishu_thread_id,
                 sender_id,
                 text[:100] if text else "",
             )
@@ -826,24 +673,8 @@ class FeishuChannel(Channel):
             else:
                 msg_type = InboundMessageType.CHAT
 
-            # Prefer any platform message id that already maps to a DeerFlow
+            # topic_id: use root_id for replies (same topic), msg_id for new messages (new topic)
-            # thread. This keeps replies to bot clarification cards in the
+            topic_id = root_id or msg_id
-            # original conversation even when Feishu reports the card as root.
-            topic_id, resolved_from_stored_mapping = self._resolve_topic_id(
-                chat_id,
-                msg_id,
-                root_id=root_id,
-                parent_id=parent_id,
-                thread_id=feishu_thread_id,
-            )
-            resolved_from_pending = False
-            if msg_type == InboundMessageType.CHAT and not resolved_from_stored_mapping:
-                pending = self._consume_pending_clarification(chat_id, sender_id)
-                pending_topic_id = self._non_empty_str(pending.get("topic_id")) if pending else None
-                if pending_topic_id:
-                    topic_id = pending_topic_id
-                    self._ensure_pending_thread_mapping(chat_id, sender_id, pending)
-                    resolved_from_pending = True
 
             inbound = self._make_inbound(
                 chat_id=chat_id,
@@ -852,15 +683,7 @@ class FeishuChannel(Channel):
                 msg_type=msg_type,
                 thread_ts=msg_id,
                 files=files_list,
-                metadata={
+                metadata={"message_id": msg_id, "root_id": root_id},
-                    "message_id": msg_id,
-                    "root_id": root_id,
-                    "parent_id": parent_id,
-                    "thread_id": feishu_thread_id,
-                    "topic_id": topic_id,
-                    "user_id": sender_id,
-                    RESOLVED_FROM_PENDING_CLARIFICATION_METADATA_KEY: resolved_from_pending,
-                },
             )
             inbound.topic_id = topic_id
 

backend/app/channels/manager.py

@@ -8,7 +8,6 @@ import mimetypes
 import re
 import time
 from collections.abc import Awaitable, Callable, Mapping
-from dataclasses import dataclass
 from pathlib import Path
 from typing import Any
 
@@ -16,24 +15,11 @@ import httpx
 from langgraph_sdk.errors import ConflictError
 
 from app.channels.commands import KNOWN_CHANNEL_COMMANDS
-from app.channels.message_bus import (
+from app.channels.message_bus import InboundMessage, InboundMessageType, MessageBus, OutboundMessage, ResolvedAttachment
-    PENDING_CLARIFICATION_METADATA_KEY,
-    InboundMessage,
-    InboundMessageType,
-    MessageBus,
-    OutboundMessage,
-    ResolvedAttachment,
-)
 from app.channels.store import ChannelStore
 from app.gateway.csrf_middleware import CSRF_COOKIE_NAME, CSRF_HEADER_NAME, generate_csrf_token
 from app.gateway.internal_auth import create_internal_auth_headers
-from deerflow.config.agents_config import load_agent_config
-from deerflow.config.paths import make_safe_user_id
 from deerflow.runtime.user_context import get_effective_user_id
-from deerflow.skills.slash import parse_slash_skill_reference
-from deerflow.skills.storage import get_or_new_skill_storage
-from deerflow.skills.storage.skill_storage import SkillStorage
-from deerflow.utils.messages import ORIGINAL_USER_CONTENT_KEY
 
 logger = logging.getLogger(__name__)
 
@@ -130,16 +116,6 @@ class InvalidChannelSessionConfigError(ValueError):
     """Raised when IM channel session overrides contain invalid agent config."""
 
 
-class SlashSkillCommandResolutionError(RuntimeError):
-    """Raised when IM slash-skill command resolution cannot complete safely."""
-
-
-@dataclass(frozen=True, slots=True)
-class _SlashSkillCommandResolution:
-    route_to_chat: bool = False
-    failure_message: str | None = None
-
-
 def _is_thread_busy_error(exc: BaseException | None) -> bool:
     if exc is None:
         return False
@@ -179,6 +155,7 @@ def _extract_response_text(result: dict | list) -> str:
     Handles special cases:
     - Regular AI text responses
     - Clarification interrupts (``ask_clarification`` tool messages)
+    - AI messages with tool_calls but no text content
     """
     if isinstance(result, list):
         messages = result
@@ -197,8 +174,6 @@ def _extract_response_text(result: dict | list) -> str:
 
         # Stop at the last human message — anything before it is a previous turn
         if msg_type == "human":
-            if _is_hidden_human_control_message(msg):
-                continue
             break
 
         # Check for tool messages from ask_clarification (interrupt case)
@@ -226,54 +201,6 @@ def _extract_response_text(result: dict | list) -> str:
     return ""
 
 
-def _messages_from_result(result: dict | list) -> list[Any]:
-    if isinstance(result, list):
-        return result
-    if isinstance(result, dict):
-        messages = result.get("messages", [])
-        if isinstance(messages, list):
-            return messages
-    return []
-
-
-def _current_turn_messages(result: dict | list) -> list[dict[str, Any]]:
-    messages = _messages_from_result(result)
-    current_turn: list[dict[str, Any]] = []
-    for msg in reversed(messages):
-        if not isinstance(msg, dict):
-            continue
-        if msg.get("type") == "human":
-            break
-        current_turn.append(msg)
-    current_turn.reverse()
-    return current_turn
-
-
-def _has_current_turn_clarification(result: dict | list) -> bool:
-    """Return True only when the current turn's final result is clarification."""
-    for msg in reversed(_current_turn_messages(result)):
-        msg_type = msg.get("type")
-        if msg_type == "tool":
-            return msg.get("name") == "ask_clarification"
-        if msg_type == "ai":
-            content = msg.get("content")
-            if isinstance(content, str):
-                if content:
-                    return False
-            elif content:
-                return False
-            if msg.get("tool_calls"):
-                return False
-    return False
-
-
-def _response_metadata(base_metadata: dict[str, Any], *, pending_clarification: bool = False) -> dict[str, Any]:
-    metadata = _slim_metadata(base_metadata)
-    if pending_clarification:
-        metadata[PENDING_CLARIFICATION_METADATA_KEY] = True
-    return metadata
-
-
 def _extract_text_content(content: Any) -> str:
     """Extract text from a streaming payload content field."""
     if isinstance(content, str):
@@ -387,8 +314,6 @@ def _extract_artifacts(result: dict | list) -> list[str]:
             continue
         # Stop at the last human message — anything before it is a previous turn
         if msg.get("type") == "human":
-            if _is_hidden_human_control_message(msg):
-                continue
             break
         # Look for AI messages with present_files tool calls
         if msg.get("type") == "ai":
@@ -401,18 +326,6 @@ def _extract_artifacts(result: dict | list) -> list[str]:
     return artifacts
 
 
-def _is_hidden_human_control_message(msg: Mapping[str, Any]) -> bool:
-    """Return whether a human message is an internal control message hidden from UI."""
-    if msg.get("type") != "human":
-        return False
-
-    additional_kwargs = msg.get("additional_kwargs")
-    if not isinstance(additional_kwargs, Mapping):
-        return False
-
-    return additional_kwargs.get("hide_from_ui") is True
-
-
 def _format_artifact_text(artifacts: list[str]) -> str:
     """Format artifact paths into a human-readable text block listing filenames."""
     import posixpath
@@ -426,46 +339,6 @@ def _format_artifact_text(artifacts: list[str]) -> str:
 _OUTPUTS_VIRTUAL_PREFIX = "/mnt/user-data/outputs/"
 
 
-def _unknown_command_reply(command: str | None = None) -> str:
-    available = " | ".join(sorted(KNOWN_CHANNEL_COMMANDS))
-    if command:
-        return f"Unknown command: /{command}. Available commands: {available}"
-    return f"Unknown command. Available commands: {available}"
-
-
-def _human_input_message(content: str, *, original_content: str | None = None) -> dict[str, Any]:
-    message: dict[str, Any] = {"role": "human", "content": content}
-    if original_content is not None and original_content != content:
-        message["additional_kwargs"] = {ORIGINAL_USER_CONTENT_KEY: original_content}
-    return message
-
-
-def _resolve_slash_skill_command(
-    text: str,
-    available_skills: set[str] | None = None,
-    storage: SkillStorage | Callable[[], SkillStorage] | None = None,
-) -> _SlashSkillCommandResolution | None:
-    reference = parse_slash_skill_reference(text)
-    if reference is None:
-        return None
-    try:
-        resolved_storage = storage() if callable(storage) else storage or get_or_new_skill_storage()
-        skills = resolved_storage.load_skills(enabled_only=False)
-
-        skill = next((candidate for candidate in skills if candidate.name == reference.name), None)
-        if skill is None:
-            return None
-        if not skill.enabled:
-            return _SlashSkillCommandResolution(failure_message=f"Skill `/{reference.name}` is installed but disabled. Enable it before using slash activation.")
-        if available_skills is not None and reference.name not in available_skills:
-            return _SlashSkillCommandResolution(failure_message=f"Skill `/{reference.name}` is not available for this agent.")
-
-        return _SlashSkillCommandResolution(route_to_chat=True)
-    except Exception as exc:
-        logger.exception("[Manager] failed to resolve slash skill command")
-        raise SlashSkillCommandResolutionError("Failed to resolve slash skill command. Please check the skill configuration.") from exc
-
-
 def _resolve_attachments(thread_id: str, artifacts: list[str]) -> list[ResolvedAttachment]:
     """Resolve virtual artifact paths to host filesystem paths with metadata.
 
@@ -547,13 +420,7 @@ async def _ingest_inbound_files(thread_id: str, msg: InboundMessage) -> list[dic
     if not msg.files:
         return []
 
-    from deerflow.uploads.manager import (
+    from deerflow.uploads.manager import claim_unique_filename, ensure_uploads_dir, normalize_filename
-        UnsafeUploadPathError,
-        claim_unique_filename,
-        ensure_uploads_dir,
-        normalize_filename,
-        write_upload_file_no_symlink,
-    )
 
     uploads_dir = ensure_uploads_dir(thread_id)
     seen_names = {entry.name for entry in uploads_dir.iterdir() if entry.is_file()}
@@ -604,10 +471,7 @@ async def _ingest_inbound_files(thread_id: str, msg: InboundMessage) -> list[dic
 
             dest = uploads_dir / safe_name
             try:
-                dest = write_upload_file_no_symlink(uploads_dir, safe_name, data)
+                dest.write_bytes(data)
-            except UnsafeUploadPathError:
-                logger.warning("[Manager] skipping inbound file with unsafe destination: %s", safe_name)
-                continue
             except Exception:
                 logger.exception("[Manager] failed to write inbound file: %s", dest)
                 continue
@@ -680,7 +544,6 @@ class ChannelManager:
         self._default_session = _as_dict(default_session)
         self._channel_sessions = dict(channel_sessions or {})
         self._client = None  # lazy init — langgraph_sdk async client
-        self._skill_storage: SkillStorage | None = None
         self._csrf_token = generate_csrf_token()
         self._semaphore: asyncio.Semaphore | None = None
         self._running = False
@@ -717,31 +580,12 @@ class ChannelManager:
             user_layer.get("config"),
         )
 
-        configurable = run_config.get("configurable")
-        if isinstance(configurable, Mapping):
-            configurable = dict(configurable)
-        else:
-            configurable = {}
-        run_config["configurable"] = configurable
-        # Pin channel-triggered runs to the root graph namespace so follow-up
-        # turns continue from the same conversation checkpoint.
-        configurable["checkpoint_ns"] = ""
-        configurable["thread_id"] = thread_id
-
-        # ``user_id`` drives user-scoped filesystem buckets that only accept
-        # ``[A-Za-z0-9_-]``, so normalize the channel id and keep the raw value
-        # under ``channel_user_id`` for platform-facing lookups.
-        run_context_identity: dict[str, Any] = {"thread_id": thread_id}
-        if msg.user_id:
-            run_context_identity["user_id"] = make_safe_user_id(msg.user_id)
-            run_context_identity["channel_user_id"] = msg.user_id
-
         run_context = _merge_dicts(
             DEFAULT_RUN_CONTEXT,
             self._default_session.get("context"),
             channel_layer.get("context"),
             user_layer.get("context"),
-            run_context_identity,
+            {"thread_id": thread_id},
         )
 
         # Custom agents are implemented as lead_agent + agent_name context.
@@ -753,21 +597,6 @@ class ChannelManager:
 
         return assistant_id, run_config, run_context
 
-    def _resolve_available_skill_names(self, msg: InboundMessage) -> set[str] | None:
-        thread_id = self.store.get_thread_id(msg.channel_name, msg.chat_id, topic_id=msg.topic_id) or ""
-        _, _, run_context = self._resolve_run_params(msg, thread_id)
-        if run_context.get("is_bootstrap"):
-            return {"bootstrap"}
-
-        agent_name = run_context.get("agent_name")
-        if not isinstance(agent_name, str) or not agent_name.strip():
-            return None
-
-        agent_config = load_agent_config(_normalize_custom_agent_name(agent_name))
-        if agent_config and agent_config.skills is not None:
-            return set(agent_config.skills)
-        return None
-
     # -- LangGraph SDK client (lazy) ----------------------------------------
 
     def _get_client(self):
@@ -785,11 +614,6 @@ class ChannelManager:
             )
         return self._client
 
-    def _get_skill_storage(self) -> SkillStorage:
-        if self._skill_storage is None:
-            self._skill_storage = get_or_new_skill_storage()
-        return self._skill_storage
-
     # -- lifecycle ---------------------------------------------------------
 
     async def start(self) -> None:
@@ -859,14 +683,6 @@ class ChannelManager:
                     exc,
                 )
                 await self._send_error(msg, str(exc))
-            except SlashSkillCommandResolutionError as exc:
-                logger.warning(
-                    "Slash skill command resolution failed for %s (chat=%s): %s",
-                    msg.channel_name,
-                    msg.chat_id,
-                    exc,
-                )
-                await self._send_error(msg, str(exc))
             except Exception:
                 logger.exception(
                     "Error handling message from %s (chat=%s)",
@@ -921,11 +737,9 @@ class ChannelManager:
         if extra_context:
             run_context.update(extra_context)
 
-        original_text = msg.text
         uploaded = await _ingest_inbound_files(thread_id, msg)
         if uploaded:
             msg.text = f"{_format_uploaded_files_block(uploaded)}\n\n{msg.text}".strip()
-        human_message = _human_input_message(msg.text, original_content=original_text)
 
         if self._channel_supports_streaming(msg.channel_name):
             await self._handle_streaming_chat(
@@ -935,30 +749,19 @@ class ChannelManager:
                 assistant_id,
                 run_config,
                 run_context,
-                human_message,
             )
             return
 
         logger.info("[Manager] invoking runs.wait(thread_id=%s, text=%r)", thread_id, msg.text[:100])
-        try:
+        result = await client.runs.wait(
-            result = await client.runs.wait(
+            thread_id,
-                thread_id,
+            assistant_id,
-                assistant_id,
+            input={"messages": [{"role": "human", "content": msg.text}]},
-                input={"messages": [human_message]},
+            config=run_config,
-                config=run_config,
+            context=run_context,
-                context=run_context,
+        )
-                multitask_strategy="reject",
-            )
-        except Exception as exc:
-            if _is_thread_busy_error(exc):
-                logger.warning("[Manager] thread busy (concurrent run rejected): thread_id=%s", thread_id)
-                await self._send_error(msg, THREAD_BUSY_MESSAGE)
-                return
-            else:
-                raise
 
         response_text = _extract_response_text(result)
-        pending_clarification = _has_current_turn_clarification(result)
         artifacts = _extract_artifacts(result)
 
         logger.info(
@@ -984,7 +787,7 @@ class ChannelManager:
             artifacts=artifacts,
             attachments=attachments,
             thread_ts=msg.thread_ts,
-            metadata=_response_metadata(msg.metadata, pending_clarification=pending_clarification),
+            metadata=_slim_metadata(msg.metadata),
         )
         logger.info("[Manager] publishing outbound message to bus: channel=%s, chat_id=%s", msg.channel_name, msg.chat_id)
         await self.bus.publish_outbound(outbound)
@@ -997,7 +800,6 @@ class ChannelManager:
         assistant_id: str,
         run_config: dict[str, Any],
         run_context: dict[str, Any],
-        human_message: dict[str, Any],
     ) -> None:
         logger.info("[Manager] invoking runs.stream(thread_id=%s, text=%r)", thread_id, msg.text[:100])
 
@@ -1013,7 +815,7 @@ class ChannelManager:
             async for chunk in client.runs.stream(
                 thread_id,
                 assistant_id,
-                input={"messages": [human_message]},
+                input={"messages": [{"role": "human", "content": msg.text}]},
                 config=run_config,
                 context=run_context,
                 stream_mode=["messages-tuple", "values"],
@@ -1047,7 +849,7 @@ class ChannelManager:
                         text=latest_text,
                         is_final=False,
                         thread_ts=msg.thread_ts,
-                        metadata=_response_metadata(msg.metadata),
+                        metadata=_slim_metadata(msg.metadata),
                     )
                 )
                 last_published_text = latest_text
@@ -1061,7 +863,6 @@ class ChannelManager:
         finally:
             result = last_values if last_values is not None else {"messages": [{"type": "ai", "content": latest_text}]}
             response_text = _extract_response_text(result)
-            pending_clarification = _has_current_turn_clarification(result)
             artifacts = _extract_artifacts(result)
             response_text, attachments = _prepare_artifact_delivery(thread_id, response_text, artifacts)
 
@@ -1093,27 +894,18 @@ class ChannelManager:
                     attachments=attachments,
                     is_final=True,
                     thread_ts=msg.thread_ts,
-                    metadata=_response_metadata(msg.metadata, pending_clarification=pending_clarification),
+                    metadata=_slim_metadata(msg.metadata),
                 )
             )
 
     # -- command handling --------------------------------------------------
 
     async def _handle_command(self, msg: InboundMessage) -> None:
-        raw_text = msg.text
+        text = msg.text.strip()
-        text = raw_text.strip()
         parts = text.split(maxsplit=1)
-        reply: str | None = None
+        command = parts[0].lower().lstrip("/")
-        if not parts:
-            command = None
-            reply = _unknown_command_reply()
-        else:
-            command = parts[0].lower().removeprefix("/")
 
-        if reply is None and not raw_text.startswith("/"):
+        if command == "bootstrap":
-            reply = _unknown_command_reply(command)
-
-        if reply is None and command == "bootstrap":
             from dataclasses import replace as _dc_replace
 
             chat_text = parts[1] if len(parts) > 1 else "Initialize workspace"
@@ -1121,7 +913,7 @@ class ChannelManager:
             await self._handle_chat(chat_msg, extra_context={"is_bootstrap": True})
             return
 
-        if reply is None and command == "new":
+        if command == "new":
             # Create a new thread through Gateway
             client = self._get_client()
             thread = await client.threads.create()
@@ -1134,14 +926,14 @@ class ChannelManager:
                 user_id=msg.user_id,
             )
             reply = "New conversation started."
-        elif reply is None and command == "status":
+        elif command == "status":
             thread_id = self.store.get_thread_id(msg.channel_name, msg.chat_id, topic_id=msg.topic_id)
             reply = f"Active thread: {thread_id}" if thread_id else "No active conversation."
-        elif reply is None and command == "models":
+        elif command == "models":
             reply = await self._fetch_gateway("/api/models", "models")
-        elif reply is None and command == "memory":
+        elif command == "memory":
             reply = await self._fetch_gateway("/api/memory", "memory")
-        elif reply is None and command == "help":
+        elif command == "help":
             reply = (
                 "Available commands:\n"
                 "/bootstrap — Start a bootstrap session (enables agent setup)\n"
@@ -1149,32 +941,16 @@ class ChannelManager:
                 "/status — Show current thread info\n"
                 "/models — List available models\n"
                 "/memory — Show memory status\n"
-                "/<skill-name> <task> — Activate an enabled skill for one turn\n"
                 "/help — Show this help"
             )
-        elif reply is None:
+        else:
-            slash_resolution = await asyncio.to_thread(
+            available = " | ".join(sorted(KNOWN_CHANNEL_COMMANDS))
-                lambda: _resolve_slash_skill_command(
+            reply = f"Unknown command: /{command}. Available commands: {available}"
-                    raw_text,
-                    self._resolve_available_skill_names(msg),
-                    self._get_skill_storage,
-                )
-            )
-            if slash_resolution and slash_resolution.failure_message:
-                reply = slash_resolution.failure_message
-            elif slash_resolution and slash_resolution.route_to_chat:
-                from dataclasses import replace as _dc_replace
-
-                chat_msg = _dc_replace(msg, msg_type=InboundMessageType.CHAT)
-                await self._handle_chat(chat_msg)
-                return
-            else:
-                reply = _unknown_command_reply(command)
 
         outbound = OutboundMessage(
             channel_name=msg.channel_name,
             chat_id=msg.chat_id,
-            thread_id=self.store.get_thread_id(msg.channel_name, msg.chat_id, topic_id=msg.topic_id) or "",
+            thread_id=self.store.get_thread_id(msg.channel_name, msg.chat_id) or "",
             text=reply,
             thread_ts=msg.thread_ts,
             metadata=_slim_metadata(msg.metadata),
@@ -1187,11 +963,7 @@ class ChannelManager:
 
         try:
             async with httpx.AsyncClient() as http:
-                resp = await http.get(
+                resp = await http.get(f"{self._gateway_url}{path}", timeout=10)
-                    f"{self._gateway_url}{path}",
-                    timeout=10,
-                    headers=create_internal_auth_headers(),
-                )
                 resp.raise_for_status()
                 data = resp.json()
         except Exception:
@@ -1212,7 +984,7 @@ class ChannelManager:
         outbound = OutboundMessage(
             channel_name=msg.channel_name,
             chat_id=msg.chat_id,
-            thread_id=self.store.get_thread_id(msg.channel_name, msg.chat_id, topic_id=msg.topic_id) or "",
+            thread_id=self.store.get_thread_id(msg.channel_name, msg.chat_id) or "",
             text=error_text,
             thread_ts=msg.thread_ts,
             metadata=_slim_metadata(msg.metadata),

backend/app/channels/message_bus.py

@@ -13,9 +13,6 @@ from typing import Any
 
 logger = logging.getLogger(__name__)
 
-PENDING_CLARIFICATION_METADATA_KEY = "pending_clarification"
-RESOLVED_FROM_PENDING_CLARIFICATION_METADATA_KEY = "resolved_from_pending_clarification"
-
 
 # ---------------------------------------------------------------------------
 # Message types

backend/app/channels/service.py

@@ -167,8 +167,6 @@ class ChannelService:
             return False
 
         try:
-            config = dict(config)
-            config["channel_store"] = self.store
             channel = channel_cls(bus=self.bus, config=config)
             self._channels[name] = channel
             await channel.start()

backend/app/channels/slack.py

@@ -9,7 +9,6 @@ from typing import Any
 from markdown_to_mrkdwn import SlackMarkdownConverter
 
 from app.channels.base import Channel
-from app.channels.commands import is_known_channel_command
 from app.channels.message_bus import InboundMessageType, MessageBus, OutboundMessage, ResolvedAttachment
 
 logger = logging.getLogger(__name__)
@@ -33,20 +32,6 @@ def _normalize_allowed_users(allowed_users: Any) -> set[str]:
     return {str(user_id) for user_id in values if str(user_id)}
 
 
-def _strip_leading_slack_bot_mention(text: str, bot_user_id: str | None) -> str:
-    if not bot_user_id:
-        return text
-    if not text.startswith("<@"):
-        return text
-    end = text.find(">")
-    if end <= 2:
-        return text
-    mentioned_user_id = text[2:end].split("|", 1)[0].lstrip("!")
-    if mentioned_user_id != bot_user_id:
-        return text
-    return text[end + 1 :].lstrip()
-
-
 class SlackChannel(Channel):
     """Slack IM channel using Socket Mode (WebSocket, no public IP).
 
@@ -64,8 +49,6 @@ class SlackChannel(Channel):
         self._web_client = None
         self._loop: asyncio.AbstractEventLoop | None = None
         self._allowed_users = _normalize_allowed_users(config.get("allowed_users", []))
-        configured_bot_user_id = config.get("bot_user_id")
-        self._bot_user_id = str(configured_bot_user_id).lstrip("@") if configured_bot_user_id else None
 
     async def start(self) -> None:
         if self._running:
@@ -89,17 +72,6 @@ class SlackChannel(Channel):
             return
 
         self._web_client = WebClient(token=bot_token)
-        if self._bot_user_id is None:
-            try:
-                auth_info = await asyncio.to_thread(self._web_client.auth_test)
-                user_id = auth_info.get("user_id") if isinstance(auth_info, dict) else None
-                if user_id is None:
-                    auth_get = getattr(auth_info, "get", None)
-                    user_id = auth_get("user_id") if callable(auth_get) else None
-                if isinstance(user_id, str) and user_id:
-                    self._bot_user_id = user_id
-            except Exception:
-                logger.warning("[Slack] failed to resolve bot user id; app mention text may include the bot mention", exc_info=True)
         self._socket_client = SocketModeClient(
             app_token=app_token,
             web_client=self._web_client,
@@ -238,12 +210,6 @@ class SlackChannel(Channel):
             if event_type != "events_api":
                 return
 
-            if self._bot_user_id is None:
-                authorization = next((item for item in req.payload.get("authorizations", []) if isinstance(item, dict)), None)
-                user_id = authorization.get("user_id") if authorization else None
-                if isinstance(user_id, str) and user_id:
-                    self._bot_user_id = user_id
-
             event = req.payload.get("event", {})
             etype = event.get("type", "")
 
@@ -267,15 +233,13 @@ class SlackChannel(Channel):
             return
 
         text = event.get("text", "").strip()
-        if event.get("type") == "app_mention":
-            text = _strip_leading_slack_bot_mention(text, self._bot_user_id)
         if not text:
             return
 
         channel_id = event.get("channel", "")
         thread_ts = event.get("thread_ts") or event.get("ts", "")
 
-        if is_known_channel_command(text):
+        if text.startswith("/"):
             msg_type = InboundMessageType.COMMAND
         else:
             msg_type = InboundMessageType.CHAT

backend/app/channels/telegram.py

@@ -60,17 +60,12 @@ class TelegramChannel(Channel):
 
         # Command handlers
         app.add_handler(CommandHandler("start", self._cmd_start))
-        app.add_handler(CommandHandler("bootstrap", self._cmd_generic))
         app.add_handler(CommandHandler("new", self._cmd_generic))
         app.add_handler(CommandHandler("status", self._cmd_generic))
         app.add_handler(CommandHandler("models", self._cmd_generic))
         app.add_handler(CommandHandler("memory", self._cmd_generic))
         app.add_handler(CommandHandler("help", self._cmd_generic))
 
-        # Slash skill commands are dynamic and cannot all be pre-registered
-        # with Telegram, so route unknown slash commands through chat handling.
-        app.add_handler(MessageHandler(filters.TEXT & filters.COMMAND, self._on_text))
-
         # General message handler
         app.add_handler(MessageHandler(filters.TEXT & ~filters.COMMAND, self._on_text))
 
@@ -233,33 +228,6 @@ class TelegramChannel(Channel):
             return True
         return user_id in self._allowed_users
 
-    def _get_bot_username(self, context) -> str | None:
-        bot = getattr(context, "bot", None)
-        username = getattr(bot, "username", None)
-        if not username and self._application is not None:
-            username = getattr(getattr(self._application, "bot", None), "username", None)
-        return str(username) if username else None
-
-    @staticmethod
-    def _strip_bot_username_from_leading_command(text: str, bot_username: str | None) -> str:
-        username = (bot_username or "").lstrip("@").lower()
-        if not username or not text.startswith("/"):
-            return text
-
-        parts = text.split(maxsplit=1)
-        command_token = parts[0]
-        if "@" not in command_token:
-            return text
-
-        command_name, addressed_username = command_token[1:].rsplit("@", 1)
-        if not command_name or addressed_username.lower() != username:
-            return text
-
-        normalized = f"/{command_name}"
-        if len(parts) > 1:
-            normalized = f"{normalized} {parts[1]}"
-        return normalized
-
     async def _cmd_start(self, update, context) -> None:
         """Handle /start command."""
         if not self._check_user(update.effective_user.id):
@@ -275,7 +243,7 @@ class TelegramChannel(Channel):
         if not self._check_user(update.effective_user.id):
             return
 
-        text = self._strip_bot_username_from_leading_command(update.message.text.strip(), self._get_bot_username(context))
+        text = update.message.text
         chat_id = str(update.effective_chat.id)
         user_id = str(update.effective_user.id)
         msg_id = str(update.message.message_id)
@@ -311,7 +279,7 @@ class TelegramChannel(Channel):
         if not self._check_user(update.effective_user.id):
             return
 
-        text = self._strip_bot_username_from_leading_command(update.message.text.strip(), self._get_bot_username(context))
+        text = update.message.text.strip()
         if not text:
             return
 

backend/app/channels/wechat.py

@@ -22,7 +22,6 @@ from cryptography.hazmat.primitives import padding
 from cryptography.hazmat.primitives.ciphers import Cipher, algorithms, modes
 
 from app.channels.base import Channel
-from app.channels.commands import is_known_channel_command
 from app.channels.message_bus import InboundMessageType, MessageBus, OutboundMessage, ResolvedAttachment
 
 logger = logging.getLogger(__name__)
@@ -621,7 +620,7 @@ class WechatChannel(Channel):
             chat_id=chat_id,
             user_id=chat_id,
             text=text,
-            msg_type=InboundMessageType.COMMAND if is_known_channel_command(text) else InboundMessageType.CHAT,
+            msg_type=InboundMessageType.COMMAND if text.startswith("/") else InboundMessageType.CHAT,
             thread_ts=thread_ts,
             files=files,
             metadata={

backend/app/channels/wecom.py

@@ -8,7 +8,6 @@ from collections.abc import Awaitable, Callable
 from typing import Any, cast
 
 from app.channels.base import Channel
-from app.channels.commands import is_known_channel_command
 from app.channels.message_bus import (
     InboundMessageType,
     MessageBus,
@@ -271,7 +270,7 @@ class WeComChannel(Channel):
 
         user_id = (body.get("from") or {}).get("userid")
 
-        inbound_type = InboundMessageType.COMMAND if is_known_channel_command(text) else InboundMessageType.CHAT
+        inbound_type = InboundMessageType.COMMAND if text.startswith("/") else InboundMessageType.CHAT
         inbound = self._make_inbound(
             chat_id=user_id,  # keep user's conversation in memory
             user_id=user_id,

backend/app/gateway/app.py

@@ -1,5 +1,6 @@
 import asyncio
 import logging
+import os
 from collections.abc import AsyncGenerator
 from contextlib import asynccontextmanager
 
@@ -8,7 +9,7 @@ from fastapi.middleware.cors import CORSMiddleware
 
 from app.gateway.auth_middleware import AuthMiddleware
 from app.gateway.config import get_gateway_config
-from app.gateway.csrf_middleware import CSRFMiddleware, get_configured_cors_origins
+from app.gateway.csrf_middleware import CSRFMiddleware
 from app.gateway.deps import langgraph_runtime
 from app.gateway.routers import (
     agents,
@@ -62,7 +63,7 @@ async def _ensure_admin_user(app: FastAPI) -> None:
 
     Subsequent boots (admin already exists):
       - Runs the one-time "no-auth → with-auth" orphan thread migration for
-        existing LangGraph thread metadata that has no user_id.
+        existing LangGraph thread metadata that has no owner_id.
 
     No SQL persistence migration is needed: the four user_id columns
     (threads_meta, runs, run_events, feedback) only come into existence
@@ -161,16 +162,10 @@ async def _migrate_orphaned_threads(store, admin_user_id: str) -> int:
 async def lifespan(app: FastAPI) -> AsyncGenerator[None, None]:
     """Application lifespan handler."""
 
-    # Load config and check necessary environment variables at startup.
+    # Load config and check necessary environment variables at startup
-    # `startup_config` is a local snapshot used only for one-shot bootstrap
-    # work (logging level, langgraph_runtime engines, channels). Request-time
-    # config resolution always routes through `get_app_config()` in
-    # `app/gateway/deps.py::get_config()` so `config.yaml` edits become
-    # visible without a process restart. We deliberately do NOT cache this
-    # snapshot on `app.state` to keep that contract enforceable.
     try:
-        startup_config = get_app_config()
+        app.state.config = get_app_config()
-        apply_logging_level(startup_config.log_level)
+        apply_logging_level(app.state.config.log_level)
         logger.info("Configuration loaded successfully")
     except Exception as e:
         error_msg = f"Failed to load configuration during gateway startup: {e}"
@@ -179,30 +174,11 @@ async def lifespan(app: FastAPI) -> AsyncGenerator[None, None]:
     config = get_gateway_config()
     logger.info(f"Starting API Gateway on {config.host}:{config.port}")
 
-    # Pre-warm tiktoken encoding cache so the first memory-injection request
-    # never blocks on the BPE data download (which hits an OpenAI/Azure URL
-    # that may be unreachable in restricted networks — see issue #3402).
-    try:
-        from deerflow.agents.memory.prompt import warm_tiktoken_cache
-
-        warmed = await asyncio.wait_for(
-            asyncio.to_thread(warm_tiktoken_cache),
-            timeout=5,
-        )
-        if warmed:
-            logger.info("tiktoken encoding cache warmed successfully")
-        else:
-            logger.warning("tiktoken encoding cache warm-up failed; token counting will use character-based fallback")
-    except TimeoutError:
-        logger.warning("tiktoken encoding cache warm-up timed out; token counting will use character-based fallback")
-    except Exception:
-        logger.warning("tiktoken warm-up skipped", exc_info=True)
-
     # Initialize LangGraph runtime components (StreamBridge, RunManager, checkpointer, store)
-    async with langgraph_runtime(app, startup_config):
+    async with langgraph_runtime(app):
         logger.info("LangGraph runtime initialised")
 
-        # Check admin bootstrap state and migrate orphan threads after admin exists.
+        # Ensure admin user exists (auto-create on first boot)
         # Must run AFTER langgraph_runtime so app.state.store is available for thread migration
         await _ensure_admin_user(app)
 
@@ -210,7 +186,7 @@ async def lifespan(app: FastAPI) -> AsyncGenerator[None, None]:
         try:
             from app.channels.service import start_channel_service
 
-            channel_service = await start_channel_service(startup_config)
+            channel_service = await start_channel_service(app.state.config)
             logger.info("Channel service started: %s", channel_service.get_status())
         except Exception:
             logger.exception("No IM channels configured or channel service failed to start")
@@ -243,9 +219,7 @@ def create_app() -> FastAPI:
         Configured FastAPI application instance.
     """
     config = get_gateway_config()
-    docs_url = "/docs" if config.enable_docs else None
+    docs_kwargs = {"docs_url": "/docs", "redoc_url": "/redoc", "openapi_url": "/openapi.json"} if config.enable_docs else {"docs_url": None, "redoc_url": None, "openapi_url": None}
-    redoc_url = "/redoc" if config.enable_docs else None
-    openapi_url = "/openapi.json" if config.enable_docs else None
 
     app = FastAPI(
         title="DeerFlow API Gateway",
@@ -265,14 +239,12 @@ API Gateway for DeerFlow - A LangGraph-based AI agent backend with sandbox execu
 
 ### Architecture
 
-LangGraph-compatible requests are routed through nginx to this gateway.
+LangGraph requests are handled by nginx reverse proxy.
-This gateway provides runtime endpoints for agent runs plus custom endpoints for models, MCP configuration, skills, and artifacts.
+This gateway provides custom endpoints for models, MCP configuration, skills, and artifacts.
         """,
         version="0.1.0",
         lifespan=lifespan,
-        docs_url=docs_url,
+        **docs_kwargs,
-        redoc_url=redoc_url,
-        openapi_url=openapi_url,
         openapi_tags=[
             {
                 "name": "models",
@@ -335,18 +307,25 @@ This gateway provides runtime endpoints for agent runs plus custom endpoints for
     # CSRF: Double Submit Cookie pattern for state-changing requests
     app.add_middleware(CSRFMiddleware)
 
-    # CORS: the unified nginx endpoint is same-origin by default. Split-origin
+    # CORS: when GATEWAY_CORS_ORIGINS is set (dev without nginx), add CORS middleware.
-    # browser clients must opt in with this explicit Gateway allowlist so CORS
+    # In production, nginx handles CORS and no middleware is needed.
-    # and CSRF origin checks share the same source of truth.
+    cors_origins_env = os.environ.get("GATEWAY_CORS_ORIGINS", "")
-    cors_origins = sorted(get_configured_cors_origins())
+    if cors_origins_env:
-    if cors_origins:
+        cors_origins = [o.strip() for o in cors_origins_env.split(",") if o.strip()]
-        app.add_middleware(
+        # Validate: wildcard origin with credentials is a security misconfiguration
-            CORSMiddleware,
+        for origin in cors_origins:
-            allow_origins=cors_origins,
+            if origin == "*":
-            allow_credentials=True,
+                logger.error("GATEWAY_CORS_ORIGINS contains wildcard '*' with allow_credentials=True. This is a security misconfiguration — browsers will reject the response. Use explicit scheme://host:port origins instead.")
-            allow_methods=["*"],
+                cors_origins = [o for o in cors_origins if o != "*"]
-            allow_headers=["*"],
+                break
-        )
+        if cors_origins:
+            app.add_middleware(
+                CORSMiddleware,
+                allow_origins=cors_origins,
+                allow_credentials=True,
+                allow_methods=["*"],
+                allow_headers=["*"],
+            )
 
     # Include routers
     # Models API is mounted at /api/models
@@ -395,7 +374,7 @@ This gateway provides runtime endpoints for agent runs plus custom endpoints for
     app.include_router(runs.router)
 
     @app.get("/health", tags=["health"])
-    async def health_check() -> dict[str, str]:
+    async def health_check() -> dict:
         """Health check endpoint.
 
         Returns:

backend/app/gateway/auth/config.py

@@ -8,8 +8,6 @@ from pydantic import BaseModel, Field
 
 logger = logging.getLogger(__name__)
 
-_SECRET_FILE = ".jwt_secret"
-
 
 class AuthConfig(BaseModel):
     """JWT and auth-related configuration. Parsed once at startup.
@@ -32,32 +30,6 @@ class AuthConfig(BaseModel):
 _auth_config: AuthConfig | None = None
 
 
-def _load_or_create_secret() -> str:
-    """Load persisted JWT secret from ``{base_dir}/.jwt_secret``, or generate and persist a new one."""
-    from deerflow.config.paths import get_paths
-
-    paths = get_paths()
-    secret_file = paths.base_dir / _SECRET_FILE
-
-    try:
-        if secret_file.exists():
-            secret = secret_file.read_text(encoding="utf-8").strip()
-            if secret:
-                return secret
-    except OSError as exc:
-        raise RuntimeError(f"Failed to read JWT secret from {secret_file}. Set AUTH_JWT_SECRET explicitly or fix DEER_FLOW_HOME/base directory permissions so DeerFlow can read its persisted auth secret.") from exc
-
-    secret = secrets.token_urlsafe(32)
-    try:
-        secret_file.parent.mkdir(parents=True, exist_ok=True)
-        fd = os.open(secret_file, os.O_WRONLY | os.O_CREAT | os.O_TRUNC, 0o600)
-        with os.fdopen(fd, "w", encoding="utf-8") as fh:
-            fh.write(secret)
-    except OSError as exc:
-        raise RuntimeError(f"Failed to persist JWT secret to {secret_file}. Set AUTH_JWT_SECRET explicitly or fix DEER_FLOW_HOME/base directory permissions so DeerFlow can store a stable auth secret.") from exc
-    return secret
-
-
 def get_auth_config() -> AuthConfig:
     """Get the global AuthConfig instance. Parses from env on first call."""
     global _auth_config
@@ -67,11 +39,11 @@ def get_auth_config() -> AuthConfig:
         load_dotenv()
         jwt_secret = os.environ.get("AUTH_JWT_SECRET")
         if not jwt_secret:
-            jwt_secret = _load_or_create_secret()
+            jwt_secret = secrets.token_urlsafe(32)
             os.environ["AUTH_JWT_SECRET"] = jwt_secret
             logger.warning(
-                "⚠ AUTH_JWT_SECRET is not set — using an auto-generated secret "
+                "⚠ AUTH_JWT_SECRET is not set — using an auto-generated ephemeral secret. "
-                "persisted to .jwt_secret. Sessions will survive restarts. "
+                "Sessions will be invalidated on restart. "
                 "For production, add AUTH_JWT_SECRET to your .env file: "
                 'python -c "import secrets; print(secrets.token_urlsafe(32))"'
             )

backend/app/gateway/auth/models.py

@@ -28,7 +28,7 @@ class User(BaseModel):
     oauth_id: str | None = Field(None, description="User ID from OAuth provider")
 
     # Auth lifecycle
-    needs_setup: bool = Field(default=False, description="True when a reset account must complete setup")
+    needs_setup: bool = Field(default=False, description="True for auto-created admin until setup completes")
     token_version: int = Field(default=0, description="Incremented on password change to invalidate old JWTs")
 
 

backend/app/gateway/config.py

@@ -8,6 +8,7 @@ class GatewayConfig(BaseModel):
 
     host: str = Field(default="0.0.0.0", description="Host to bind the gateway server")
     port: int = Field(default=8001, description="Port to bind the gateway server")
+    cors_origins: list[str] = Field(default_factory=lambda: ["http://localhost:3000"], description="Allowed CORS origins")
     enable_docs: bool = Field(default=True, description="Enable Swagger/ReDoc/OpenAPI endpoints")
 
 
@@ -18,9 +19,11 @@ def get_gateway_config() -> GatewayConfig:
     """Get gateway config, loading from environment if available."""
     global _gateway_config
     if _gateway_config is None:
+        cors_origins_str = os.getenv("CORS_ORIGINS", "http://localhost:3000")
         _gateway_config = GatewayConfig(
             host=os.getenv("GATEWAY_HOST", "0.0.0.0"),
             port=int(os.getenv("GATEWAY_PORT", "8001")),
+            cors_origins=cors_origins_str.split(","),
             enable_docs=os.getenv("GATEWAY_ENABLE_DOCS", "true").lower() == "true",
         )
     return _gateway_config

backend/app/gateway/csrf_middleware.py

@@ -4,10 +4,8 @@ Per RFC-001:
 State-changing operations require CSRF protection.
 """
 
-import os
 import secrets
-from collections.abc import Awaitable, Callable
+from collections.abc import Callable
-from urllib.parse import urlsplit
 
 from fastapi import Request, Response
 from starlette.middleware.base import BaseHTTPMiddleware
@@ -21,7 +19,7 @@ CSRF_TOKEN_LENGTH = 64  # bytes
 
 def is_secure_request(request: Request) -> bool:
     """Detect whether the original client request was made over HTTPS."""
-    return _request_scheme(request) == "https"
+    return request.headers.get("x-forwarded-proto", request.url.scheme) == "https"
 
 
 def generate_csrf_token() -> str:
@@ -63,129 +61,15 @@ def is_auth_endpoint(request: Request) -> bool:
     return request.url.path.rstrip("/") in _AUTH_EXEMPT_PATHS
 
 
-def _host_with_optional_port(hostname: str, port: int | None, scheme: str) -> str:
-    """Return normalized host[:port], omitting default ports."""
-    host = hostname.lower()
-    if ":" in host and not host.startswith("["):
-        host = f"[{host}]"
-
-    if port is None or (scheme == "http" and port == 80) or (scheme == "https" and port == 443):
-        return host
-    return f"{host}:{port}"
-
-
-def _normalize_origin(origin: str) -> str | None:
-    """Return a normalized scheme://host[:port] origin, or None for invalid input."""
-    try:
-        parsed = urlsplit(origin.strip())
-        port = parsed.port
-    except ValueError:
-        return None
-
-    scheme = parsed.scheme.lower()
-    if scheme not in {"http", "https"} or not parsed.hostname:
-        return None
-
-    # Browser Origin is only scheme/host/port. Reject URL-shaped or credentialed values.
-    if parsed.username or parsed.password or parsed.path or parsed.query or parsed.fragment:
-        return None
-
-    return f"{scheme}://{_host_with_optional_port(parsed.hostname, port, scheme)}"
-
-
-def _configured_cors_origins() -> set[str]:
-    """Return explicit configured browser origins that may call auth routes."""
-    origins = set()
-    for raw_origin in os.environ.get("GATEWAY_CORS_ORIGINS", "").split(","):
-        origin = raw_origin.strip()
-        if not origin or origin == "*":
-            continue
-        normalized = _normalize_origin(origin)
-        if normalized:
-            origins.add(normalized)
-    return origins
-
-
-def get_configured_cors_origins() -> set[str]:
-    """Return normalized explicit browser origins from GATEWAY_CORS_ORIGINS."""
-    return _configured_cors_origins()
-
-
-def _first_header_value(value: str | None) -> str | None:
-    """Return the first value from a comma-separated proxy header."""
-    if not value:
-        return None
-    first = value.split(",", 1)[0].strip()
-    return first or None
-
-
-def _forwarded_param(request: Request, name: str) -> str | None:
-    """Extract a parameter from the first RFC 7239 Forwarded header entry."""
-    forwarded = _first_header_value(request.headers.get("forwarded"))
-    if not forwarded:
-        return None
-
-    for part in forwarded.split(";"):
-        key, sep, value = part.strip().partition("=")
-        if sep and key.lower() == name:
-            return value.strip().strip('"') or None
-    return None
-
-
-def _request_scheme(request: Request) -> str:
-    """Resolve the original request scheme from trusted proxy headers."""
-    scheme = _forwarded_param(request, "proto") or _first_header_value(request.headers.get("x-forwarded-proto")) or request.url.scheme
-    return scheme.lower()
-
-
-def _request_origin(request: Request) -> str | None:
-    """Build the origin for the URL the browser is targeting."""
-    scheme = _request_scheme(request)
-    host = _forwarded_param(request, "host") or _first_header_value(request.headers.get("x-forwarded-host")) or request.headers.get("host") or request.url.netloc
-
-    forwarded_port = _first_header_value(request.headers.get("x-forwarded-port"))
-    if forwarded_port and ":" not in host.rsplit("]", 1)[-1]:
-        host = f"{host}:{forwarded_port}"
-
-    return _normalize_origin(f"{scheme}://{host}")
-
-
-def is_allowed_auth_origin(request: Request) -> bool:
-    """Allow auth POSTs only from the same origin or explicit configured origins.
-
-    Login/register/initialize are exempt from the double-submit token because
-    first-time browser clients do not have a CSRF token yet. They still create
-    a session cookie, so browser requests with a hostile Origin header must be
-    rejected to prevent login CSRF / session fixation. Requests without Origin
-    are allowed for non-browser clients such as curl and mobile integrations.
-    """
-    origin = request.headers.get("origin")
-    if not origin:
-        return True
-
-    normalized_origin = _normalize_origin(origin)
-    if normalized_origin is None:
-        return False
-
-    request_origin = _request_origin(request)
-    return normalized_origin in _configured_cors_origins() or (request_origin is not None and normalized_origin == request_origin)
-
-
 class CSRFMiddleware(BaseHTTPMiddleware):
     """Middleware that implements CSRF protection using Double Submit Cookie pattern."""
 
     def __init__(self, app: ASGIApp) -> None:
         super().__init__(app)
 
-    async def dispatch(self, request: Request, call_next: Callable[[Request], Awaitable[Response]]) -> Response:
+    async def dispatch(self, request: Request, call_next: Callable) -> Response:
         _is_auth = is_auth_endpoint(request)
 
-        if should_check_csrf(request) and _is_auth and not is_allowed_auth_origin(request):
-            return JSONResponse(
-                status_code=403,
-                content={"detail": "Cross-site auth request denied."},
-            )
-
         if should_check_csrf(request) and not _is_auth:
             cookie_token = request.cookies.get(CSRF_COOKIE_NAME)
             header_token = request.headers.get(CSRF_HEADER_NAME)

backend/app/gateway/deps.py

@@ -3,22 +3,11 @@
 **Getters** (used by routers): raise 503 when a required dependency is
 missing, except ``get_store`` which returns ``None``.
 
-``AppConfig`` is intentionally *not* cached on ``app.state``. Routers and the
-run path resolve it through :func:`deerflow.config.app_config.get_app_config`,
-which performs mtime-based hot reload, so edits to ``config.yaml`` take
-effect on the next request without a process restart. The engines created in
-:func:`langgraph_runtime` (stream bridge, persistence, checkpointer, store,
-run-event store) accept a ``startup_config`` snapshot — they are
-restart-required by design and stay bound to that snapshot to keep the live
-process consistent with itself.
-
 Initialization is handled directly in ``app.py`` via :class:`AsyncExitStack`.
 """
 
 from __future__ import annotations
 
-import asyncio
-import logging
 from collections.abc import AsyncGenerator, Callable
 from contextlib import AsyncExitStack, asynccontextmanager
 from typing import TYPE_CHECKING, TypeVar, cast
@@ -26,144 +15,36 @@ from typing import TYPE_CHECKING, TypeVar, cast
 from fastapi import FastAPI, HTTPException, Request
 from langgraph.types import Checkpointer
 
-from deerflow.config.app_config import AppConfig, get_app_config
+from deerflow.config.app_config import AppConfig
 from deerflow.persistence.feedback import FeedbackRepository
 from deerflow.runtime import RunContext, RunManager, StreamBridge
 from deerflow.runtime.events.store.base import RunEventStore
 from deerflow.runtime.runs.store.base import RunStore
 
-logger = logging.getLogger(__name__)
-
-# Upper bound (seconds) for draining in-flight runs during shutdown, before the
-# AsyncExitStack tears down the checkpointer (and its connection pool). Kept
-# local to avoid an app -> deps -> app import cycle. This is a *separate* budget
-# from ``app.gateway.app._SHUTDOWN_HOOK_TIMEOUT_SECONDS`` (currently also 5.0s,
-# which bounds channel-service stop): the two govern independent teardown steps
-# and may diverge, but both count toward the lifespan shutdown window — revisit
-# them together if their sum must stay within the server's graceful-shutdown
-# timeout.
-_RUN_DRAIN_TIMEOUT_SECONDS = 5.0
-
-
-async def _drain_inflight_runs(run_manager: RunManager) -> None:
-    """Drain in-flight runs before the checkpointer is torn down (issue #3373).
-
-    Shields the (internally-bounded) drain so that even if the lifespan
-    coroutine is itself cancelled mid-shutdown — a second SIGINT or the server's
-    graceful-shutdown timeout, i.e. the same signal storm behind #3373 — the
-    checkpointer pool is not closed while run tasks are still writing
-    checkpoints. On such a cancellation we let the already-running drain finish
-    (it is bounded by ``RunManager.shutdown``'s own timeout) and then propagate
-    the cancellation.
-    """
-    drain = asyncio.create_task(run_manager.shutdown(timeout=_RUN_DRAIN_TIMEOUT_SECONDS))
-    try:
-        await asyncio.shield(drain)
-    except asyncio.CancelledError:
-        # Re-shield so this second wait does not abandon the in-flight drain;
-        # it is bounded, so this cannot hang. Then re-raise to honour shutdown.
-        try:
-            await asyncio.shield(drain)
-        except Exception:
-            logger.exception("In-flight run drain failed after shutdown cancellation")
-        raise
-    except Exception:
-        logger.exception("Failed to drain in-flight runs during shutdown")
-
-
 if TYPE_CHECKING:
     from app.gateway.auth.local_provider import LocalAuthProvider
     from app.gateway.auth.repositories.sqlite import SQLiteUserRepository
     from deerflow.persistence.thread_meta.base import ThreadMetaStore
-    from deerflow.runtime import RunRecord
 
 
 T = TypeVar("T")
 
 
-async def _mark_latest_recovered_threads_error(
+def get_config(request: Request) -> AppConfig:
-    run_manager: RunManager,
+    """Return the app-scoped ``AppConfig`` stored on ``app.state``."""
-    thread_store: ThreadMetaStore,
+    config = getattr(request.app.state, "config", None)
-    recovered_runs: list[RunRecord],
+    if config is None:
-) -> None:
+        raise HTTPException(status_code=503, detail="Configuration not available")
-    """Mark thread status as error only when its newest run was recovered."""
+    return config
-    recovered_by_thread: dict[str, set[str]] = {}
-    for record in recovered_runs:
-        recovered_by_thread.setdefault(record.thread_id, set()).add(record.run_id)
-
-    for thread_id, recovered_run_ids in recovered_by_thread.items():
-        try:
-            latest_runs = await run_manager.list_by_thread(thread_id, user_id=None, limit=1)
-        except Exception:
-            logger.warning("Failed to find latest run for thread %s during run reconciliation", thread_id, exc_info=True)
-            continue
-        if not latest_runs or latest_runs[0].run_id not in recovered_run_ids:
-            continue
-        try:
-            await thread_store.update_status(thread_id, "error", user_id=None)
-        except Exception:
-            logger.warning("Failed to mark thread %s as error during run reconciliation", thread_id, exc_info=True)
-
-
-def get_config() -> AppConfig:
-    """Return the freshest ``AppConfig`` for the current request.
-
-    Routes through :func:`deerflow.config.app_config.get_app_config`, which
-    honours runtime ``ContextVar`` overrides and reloads ``config.yaml`` from
-    disk when its mtime changes. ``AppConfig`` is not cached on ``app.state``
-    at all — the only startup-time snapshot lives as a local
-    ``startup_config`` variable inside ``lifespan()`` and is passed
-    explicitly into :func:`langgraph_runtime` for the engines that are
-    restart-required by design. Routing every request through
-    :func:`get_app_config` closes the bytedance/deer-flow issue #3107 BUG-001
-    split-brain where the worker / lead-agent thread saw a stale startup
-    snapshot.
-
-    Hot-reload boundary: fields backed by startup-time singletons
-    (engines, sandbox provider, IM channels, logging handler) require a
-    process restart to change at runtime. The authoritative list lives in
-    :mod:`deerflow.config.reload_boundary` and is mirrored by the
-    standardised ``"startup-only:"`` prefix on the matching
-    ``Field(description=...)`` in :class:`AppConfig` — IDE hover on those
-    fields will surface the boundary inline. See
-    ``backend/CLAUDE.md`` "Config Hot-Reload Boundary" for the operator
-    summary.
-
-    Any failure to materialise the config (missing file, permission denied,
-    YAML parse error, validation error) is reported as 503 — semantically
-    "the gateway cannot serve requests without a usable configuration" — and
-    logged with the original exception so operators have something to debug.
-    """
-    try:
-        return get_app_config()
-    except Exception as exc:  # noqa: BLE001 - request boundary: log and degrade gracefully
-        logger.exception("Failed to load AppConfig at request time")
-        raise HTTPException(status_code=503, detail="Configuration not available") from exc
 
 
 @asynccontextmanager
-async def langgraph_runtime(app: FastAPI, startup_config: AppConfig) -> AsyncGenerator[None, None]:
+async def langgraph_runtime(app: FastAPI) -> AsyncGenerator[None, None]:
     """Bootstrap and tear down all LangGraph runtime singletons.
 
-    ``startup_config`` is the ``AppConfig`` snapshot taken once during
-    ``lifespan()`` for one-shot infrastructure bootstrap. The engines and
-    stores constructed here (stream bridge, persistence engine, checkpointer,
-    store, run-event store) are restart-required by design — they hold live
-    connections, file handles, or singleton providers — so they bind to this
-    snapshot and survive across `config.yaml` edits. Request-time consumers
-    must still go through :func:`get_config` for any field that should be
-    hot-reloadable. See ``backend/CLAUDE.md`` "Config Hot-Reload Boundary".
-
-    The matching ``run_events_config`` is frozen onto ``app.state`` so
-    :func:`get_run_context` pairs a freshly-loaded ``AppConfig`` with the
-    *startup-time* run-events configuration the underlying ``event_store``
-    was built from — otherwise the runtime could end up combining a live
-    new ``run_events_config`` with an event store still bound to the
-    previous backend.
-
     Usage in ``app.py``::
 
-        async with langgraph_runtime(app, startup_config):
+        async with langgraph_runtime(app):
             yield
     """
     from deerflow.persistence.engine import close_engine, get_session_factory, init_engine_from_config
@@ -172,7 +53,9 @@ async def langgraph_runtime(app: FastAPI, startup_config: AppConfig) -> AsyncGen
     from deerflow.runtime.events.store import make_run_event_store
 
     async with AsyncExitStack() as stack:
-        config = startup_config
+        config = getattr(app.state, "config", None)
+        if config is None:
+            raise RuntimeError("langgraph_runtime() requires app.state.config to be initialized")
 
         app.state.stream_bridge = await stack.enter_async_context(make_stream_bridge(config))
 
@@ -201,38 +84,16 @@ async def langgraph_runtime(app: FastAPI, startup_config: AppConfig) -> AsyncGen
 
         app.state.thread_store = make_thread_store(sf, app.state.store)
 
-        # Run event store. The store and the matching ``run_events_config`` are
+        # Run event store (has its own factory with config-driven backend selection)
-        # both frozen at startup so ``get_run_context`` does not combine a
-        # freshly-reloaded ``AppConfig.run_events`` with a store still bound to
-        # the previous backend.
         run_events_config = getattr(config, "run_events", None)
-        app.state.run_events_config = run_events_config
         app.state.run_event_store = make_run_event_store(run_events_config)
 
         # RunManager with store backing for persistence
         app.state.run_manager = RunManager(store=app.state.run_store)
-        if getattr(config.database, "backend", None) == "sqlite":
-            from deerflow.utils.time import now_iso
-
-            # Startup-only recovery: clean shutdowns return no active rows and
-            # the thread-status update below becomes a no-op.
-            recovered_runs = await app.state.run_manager.reconcile_orphaned_inflight_runs(
-                error="Gateway restarted before this run reached a durable final state.",
-                before=now_iso(),
-            )
-            await _mark_latest_recovered_threads_error(app.state.run_manager, app.state.thread_store, recovered_runs)
 
         try:
             yield
         finally:
-            # Drain in-flight run tasks BEFORE the AsyncExitStack tears down the
-            # checkpointer (and its connection pool). A run still mid-graph would
-            # otherwise leak into asyncio.run() shutdown, where langgraph's
-            # _checkpointer_put_after_previous aput races the closed pool and
-            # raises PoolClosed (issue #3373).
-            run_manager = getattr(app.state, "run_manager", None)
-            if run_manager is not None:
-                await _drain_inflight_runs(run_manager)
             await close_engine()
 
 
@@ -278,20 +139,16 @@ def get_thread_store(request: Request) -> ThreadMetaStore:
 def get_run_context(request: Request) -> RunContext:
     """Build a :class:`RunContext` from ``app.state`` singletons.
 
-    Returns a *base* context with infrastructure dependencies. The
+    Returns a *base* context with infrastructure dependencies.
-    ``app_config`` field is resolved live so per-run fields (e.g.
-    ``models[*].max_tokens``) follow ``config.yaml`` edits; the
-    ``event_store`` / ``run_events_config`` pair stays frozen to the snapshot
-    captured in :func:`langgraph_runtime` so callers never see a store bound
-    to one backend paired with a config pointing at another.
     """
+    config = get_config(request)
     return RunContext(
         checkpointer=get_checkpointer(request),
         store=get_store(request),
         event_store=get_run_event_store(request),
-        run_events_config=getattr(request.app.state, "run_events_config", None),
+        run_events_config=getattr(config, "run_events", None),
         thread_store=get_thread_store(request),
-        app_config=get_config(),
+        app_config=config,
     )
 
 

backend/app/gateway/internal_auth.py

@@ -1,38 +1,26 @@
-"""Authentication for trusted Gateway internal callers."""
+"""Process-local authentication for Gateway internal callers."""
 
 from __future__ import annotations
 
-import os
 import secrets
 from types import SimpleNamespace
 
 from deerflow.runtime.user_context import DEFAULT_USER_ID
 
 INTERNAL_AUTH_HEADER_NAME = "X-DeerFlow-Internal-Token"
-INTERNAL_AUTH_ENV_VAR = "DEER_FLOW_INTERNAL_AUTH_TOKEN"
+_INTERNAL_AUTH_TOKEN = secrets.token_urlsafe(32)
-INTERNAL_SYSTEM_ROLE = "internal"
-
-
-def _load_internal_auth_token() -> str:
-    token = os.environ.get(INTERNAL_AUTH_ENV_VAR)
-    if token:
-        return token
-    return secrets.token_urlsafe(32)
-
-
-_INTERNAL_AUTH_TOKEN = _load_internal_auth_token()
 
 
 def create_internal_auth_headers() -> dict[str, str]:
-    """Return headers that authenticate trusted Gateway internal calls."""
+    """Return headers that authenticate same-process Gateway internal calls."""
     return {INTERNAL_AUTH_HEADER_NAME: _INTERNAL_AUTH_TOKEN}
 
 
 def is_valid_internal_auth_token(token: str | None) -> bool:
-    """Return True when *token* matches this Gateway worker's internal token."""
+    """Return True when *token* matches the process-local internal token."""
     return bool(token) and secrets.compare_digest(token, _INTERNAL_AUTH_TOKEN)
 
 
 def get_internal_user():
     """Return the synthetic user used for trusted internal channel calls."""
-    return SimpleNamespace(id=DEFAULT_USER_ID, system_role=INTERNAL_SYSTEM_ROLE)
+    return SimpleNamespace(id=DEFAULT_USER_ID, system_role="internal")

backend/app/gateway/langgraph_auth.py

@@ -1,12 +1,8 @@
-"""LangGraph compatibility auth handler — shares JWT logic with Gateway.
+"""LangGraph Server auth handler — shares JWT logic with Gateway.
 
-The default DeerFlow runtime is embedded in the FastAPI Gateway; scripts and
+Loaded by LangGraph Server via langgraph.json ``auth.path``.
-Docker deployments do not load this module.  It is retained for LangGraph
+Reuses the same ``decode_token`` / ``get_auth_config`` as Gateway,
-tooling, Studio, or direct LangGraph Server compatibility through
+so both modes validate tokens with the same secret and rules.
-``langgraph.json``'s ``auth.path``.
-
-When that compatibility path is used, this module reuses the same JWT and CSRF
-rules as Gateway so both modes validate sessions consistently.
 
 Two layers:
   1. @auth.authenticate — validates JWT cookie, extracts user_id,

backend/app/gateway/pagination.py

@@ -1,15 +0,0 @@
-"""Shared pagination helpers for gateway routers."""
-
-from __future__ import annotations
-
-
-def trim_run_message_page(rows: list[dict], *, limit: int, after_seq: int | None) -> tuple[list[dict], bool]:
-    """Trim a ``limit + 1`` run-message page while preserving page boundaries."""
-    has_more = len(rows) > limit
-    if not has_more:
-        return rows, False
-
-    if after_seq is not None:
-        return rows[:limit], True
-
-    return rows[-limit:], True

backend/app/gateway/routers/agents.py

@@ -1,6 +1,5 @@
 """CRUD API for custom agents."""
 
-import asyncio
 import logging
 import re
 import shutil
@@ -12,7 +11,6 @@ from pydantic import BaseModel, Field
 from deerflow.config.agents_api_config import get_agents_api_config
 from deerflow.config.agents_config import AgentConfig, list_custom_agents, load_agent_config, load_agent_soul
 from deerflow.config.paths import get_paths
-from deerflow.runtime.user_context import get_effective_user_id
 
 logger = logging.getLogger(__name__)
 router = APIRouter(prefix="/api", tags=["agents"])
@@ -88,11 +86,11 @@ def _require_agents_api_enabled() -> None:
         )
 
 
-def _agent_config_to_response(agent_cfg: AgentConfig, include_soul: bool = False, *, user_id: str | None = None) -> AgentResponse:
+def _agent_config_to_response(agent_cfg: AgentConfig, include_soul: bool = False) -> AgentResponse:
     """Convert AgentConfig to AgentResponse."""
     soul: str | None = None
     if include_soul:
-        soul = load_agent_soul(agent_cfg.name, user_id=user_id) or ""
+        soul = load_agent_soul(agent_cfg.name) or ""
 
     return AgentResponse(
         name=agent_cfg.name,
@@ -118,10 +116,9 @@ async def list_agents() -> AgentsListResponse:
     """
     _require_agents_api_enabled()
 
-    user_id = get_effective_user_id()
     try:
-        agents = list_custom_agents(user_id=user_id)
+        agents = list_custom_agents()
-        return AgentsListResponse(agents=[_agent_config_to_response(a, include_soul=True, user_id=user_id) for a in agents])
+        return AgentsListResponse(agents=[_agent_config_to_response(a, include_soul=True) for a in agents])
     except Exception as e:
         logger.error(f"Failed to list agents: {e}", exc_info=True)
         raise HTTPException(status_code=500, detail=f"Failed to list agents: {str(e)}")
@@ -147,12 +144,7 @@ async def check_agent_name(name: str) -> dict:
     _require_agents_api_enabled()
     _validate_agent_name(name)
     normalized = _normalize_agent_name(name)
-    user_id = get_effective_user_id()
+    available = not get_paths().agent_dir(normalized).exists()
-    paths = get_paths()
-    # Treat the name as taken if either the per-user path or the legacy shared
-    # path holds an agent — picking a name that collides with an unmigrated
-    # legacy agent would shadow the legacy entry once migration runs.
-    available = not paths.user_agent_dir(user_id, normalized).exists() and not paths.agent_dir(normalized).exists()
     return {"available": available, "name": normalized}
 
 
@@ -177,11 +169,10 @@ async def get_agent(name: str) -> AgentResponse:
     _require_agents_api_enabled()
     _validate_agent_name(name)
     name = _normalize_agent_name(name)
-    user_id = get_effective_user_id()
 
     try:
-        agent_cfg = load_agent_config(name, user_id=user_id)
+        agent_cfg = load_agent_config(name)
-        return _agent_config_to_response(agent_cfg, include_soul=True, user_id=user_id)
+        return _agent_config_to_response(agent_cfg, include_soul=True)
     except FileNotFoundError:
         raise HTTPException(status_code=404, detail=f"Agent '{name}' not found")
     except Exception as e:
@@ -211,63 +202,47 @@ async def create_agent_endpoint(request: AgentCreateRequest) -> AgentResponse:
     _require_agents_api_enabled()
     _validate_agent_name(request.name)
     normalized_name = _normalize_agent_name(request.name)
-    user_id = get_effective_user_id()
-    paths = get_paths()
 
-    def _create_agent() -> AgentResponse | None:
+    agent_dir = get_paths().agent_dir(normalized_name)
-        # Worker thread: base-dir resolution, existence checks, directory/file
-        # creation, read-back, and failure cleanup are all blocking filesystem
-        # IO that must stay off the event loop.
-        agent_dir = paths.user_agent_dir(user_id, normalized_name)
-        legacy_dir = paths.agent_dir(normalized_name)
 
-        if legacy_dir.exists():
+    if agent_dir.exists():
-            return None  # signals 409 to the caller
-
-        try:
-            try:
-                agent_dir.mkdir(parents=True, exist_ok=False)
-            except FileExistsError:
-                return None  # signals 409 to the caller
-            # Write config.yaml
-            config_data: dict = {"name": normalized_name}
-            if request.description:
-                config_data["description"] = request.description
-            if request.model is not None:
-                config_data["model"] = request.model
-            if request.tool_groups is not None:
-                config_data["tool_groups"] = request.tool_groups
-            if request.skills is not None:
-                config_data["skills"] = request.skills
-
-            config_file = agent_dir / "config.yaml"
-            with open(config_file, "w", encoding="utf-8") as f:
-                yaml.dump(config_data, f, default_flow_style=False, allow_unicode=True)
-
-            # Write SOUL.md
-            soul_file = agent_dir / "SOUL.md"
-            soul_file.write_text(request.soul, encoding="utf-8")
-
-            logger.info(f"Created agent '{normalized_name}' at {agent_dir}")
-
-            agent_cfg = load_agent_config(normalized_name, user_id=user_id)
-            return _agent_config_to_response(agent_cfg, include_soul=True, user_id=user_id)
-        except Exception:
-            # Clean up partial state on failure before surfacing the error.
-            if agent_dir.exists():
-                shutil.rmtree(agent_dir)
-            raise
-
-    try:
-        response = await asyncio.to_thread(_create_agent)
-    except Exception as e:
-        logger.error(f"Failed to create agent '{request.name}': {e}", exc_info=True)
-        raise HTTPException(status_code=500, detail=f"Failed to create agent: {str(e)}")
-
-    if response is None:
         raise HTTPException(status_code=409, detail=f"Agent '{normalized_name}' already exists")
 
-    return response
+    try:
+        agent_dir.mkdir(parents=True, exist_ok=True)
+
+        # Write config.yaml
+        config_data: dict = {"name": normalized_name}
+        if request.description:
+            config_data["description"] = request.description
+        if request.model is not None:
+            config_data["model"] = request.model
+        if request.tool_groups is not None:
+            config_data["tool_groups"] = request.tool_groups
+        if request.skills is not None:
+            config_data["skills"] = request.skills
+
+        config_file = agent_dir / "config.yaml"
+        with open(config_file, "w", encoding="utf-8") as f:
+            yaml.dump(config_data, f, default_flow_style=False, allow_unicode=True)
+
+        # Write SOUL.md
+        soul_file = agent_dir / "SOUL.md"
+        soul_file.write_text(request.soul, encoding="utf-8")
+
+        logger.info(f"Created agent '{normalized_name}' at {agent_dir}")
+
+        agent_cfg = load_agent_config(normalized_name)
+        return _agent_config_to_response(agent_cfg, include_soul=True)
+
+    except HTTPException:
+        raise
+    except Exception as e:
+        # Clean up on failure
+        if agent_dir.exists():
+            shutil.rmtree(agent_dir)
+        logger.error(f"Failed to create agent '{request.name}': {e}", exc_info=True)
+        raise HTTPException(status_code=500, detail=f"Failed to create agent: {str(e)}")
 
 
 @router.put(
@@ -292,20 +267,13 @@ async def update_agent(name: str, request: AgentUpdateRequest) -> AgentResponse:
     _require_agents_api_enabled()
     _validate_agent_name(name)
     name = _normalize_agent_name(name)
-    user_id = get_effective_user_id()
 
     try:
-        agent_cfg = load_agent_config(name, user_id=user_id)
+        agent_cfg = load_agent_config(name)
     except FileNotFoundError:
         raise HTTPException(status_code=404, detail=f"Agent '{name}' not found")
 
-    paths = get_paths()
+    agent_dir = get_paths().agent_dir(name)
-    agent_dir = paths.user_agent_dir(user_id, name)
-    if not agent_dir.exists() and paths.agent_dir(name).exists():
-        raise HTTPException(
-            status_code=409,
-            detail=(f"Agent '{name}' only exists in the legacy shared layout and is not scoped to a user. Run scripts/migrate_user_isolation.py to move legacy agents into the per-user layout before updating."),
-        )
 
     try:
         # Update config if any config fields changed
@@ -346,8 +314,8 @@ async def update_agent(name: str, request: AgentUpdateRequest) -> AgentResponse:
 
         logger.info(f"Updated agent '{name}'")
 
-        refreshed_cfg = load_agent_config(name, user_id=user_id)
+        refreshed_cfg = load_agent_config(name)
-        return _agent_config_to_response(refreshed_cfg, include_soul=True, user_id=user_id)
+        return _agent_config_to_response(refreshed_cfg, include_soul=True)
 
     except HTTPException:
         raise
@@ -434,38 +402,20 @@ async def delete_agent(name: str) -> None:
         name: The agent name.
 
     Raises:
-        HTTPException: 404 if no per-user copy exists; 409 if only a legacy
+        HTTPException: 404 if agent not found.
-            shared copy exists (suggesting the migration script).
     """
     _require_agents_api_enabled()
     _validate_agent_name(name)
     name = _normalize_agent_name(name)
-    user_id = get_effective_user_id()
-    paths = get_paths()
 
-    def _remove_agent_dir() -> tuple[str, str]:
+    agent_dir = get_paths().agent_dir(name)
-        # Runs in a worker thread: resolving the base dir, probing the directory
+
-        # (`exists`), and removing it (`rmtree`) are all blocking filesystem IO
+    if not agent_dir.exists():
-        # that must stay off the event loop.
+        raise HTTPException(status_code=404, detail=f"Agent '{name}' not found")
-        agent_dir = paths.user_agent_dir(user_id, name)
-        if not agent_dir.exists():
-            outcome = "legacy" if paths.agent_dir(name).exists() else "missing"
-            return outcome, str(agent_dir)
-        shutil.rmtree(agent_dir)
-        return "deleted", str(agent_dir)
 
     try:
-        outcome, agent_dir = await asyncio.to_thread(_remove_agent_dir)
+        shutil.rmtree(agent_dir)
+        logger.info(f"Deleted agent '{name}' from {agent_dir}")
     except Exception as e:
         logger.error(f"Failed to delete agent '{name}': {e}", exc_info=True)
         raise HTTPException(status_code=500, detail=f"Failed to delete agent: {str(e)}")
-
-    if outcome == "legacy":
-        raise HTTPException(
-            status_code=409,
-            detail=(f"Agent '{name}' only exists in the legacy shared layout and is not scoped to a user. Run scripts/migrate_user_isolation.py to move legacy agents into the per-user layout before deleting."),
-        )
-    if outcome == "missing":
-        raise HTTPException(status_code=404, detail=f"Agent '{name}' not found")
-
-    logger.info(f"Deleted agent '{name}' from {agent_dir}")

backend/app/gateway/routers/artifacts.py

@@ -20,9 +20,6 @@ ACTIVE_CONTENT_MIME_TYPES = {
     "image/svg+xml",
 }
 
-MAX_SKILL_ARCHIVE_MEMBER_BYTES = 16 * 1024 * 1024
-_SKILL_ARCHIVE_READ_CHUNK_SIZE = 64 * 1024
-
 
 def _build_content_disposition(disposition_type: str, filename: str) -> str:
     """Build an RFC 5987 encoded Content-Disposition header value."""
@@ -47,22 +44,6 @@ def is_text_file_by_content(path: Path, sample_size: int = 8192) -> bool:
         return False
 
 
-def _read_skill_archive_member(zip_ref: zipfile.ZipFile, info: zipfile.ZipInfo) -> bytes:
-    """Read a .skill archive member while enforcing an uncompressed size cap."""
-    if info.file_size > MAX_SKILL_ARCHIVE_MEMBER_BYTES:
-        raise HTTPException(status_code=413, detail="Skill archive member is too large to preview")
-
-    chunks: list[bytes] = []
-    total_read = 0
-    with zip_ref.open(info, "r") as src:
-        while chunk := src.read(_SKILL_ARCHIVE_READ_CHUNK_SIZE):
-            total_read += len(chunk)
-            if total_read > MAX_SKILL_ARCHIVE_MEMBER_BYTES:
-                raise HTTPException(status_code=413, detail="Skill archive member is too large to preview")
-            chunks.append(chunk)
-    return b"".join(chunks)
-
-
 def _extract_file_from_skill_archive(zip_path: Path, internal_path: str) -> bytes | None:
     """Extract a file from a .skill ZIP archive.
 
@@ -79,16 +60,16 @@ def _extract_file_from_skill_archive(zip_path: Path, internal_path: str) -> byte
     try:
         with zipfile.ZipFile(zip_path, "r") as zip_ref:
             # List all files in the archive
-            infos_by_name = {info.filename: info for info in zip_ref.infolist()}
+            namelist = zip_ref.namelist()
 
             # Try direct path first
-            if internal_path in infos_by_name:
+            if internal_path in namelist:
-                return _read_skill_archive_member(zip_ref, infos_by_name[internal_path])
+                return zip_ref.read(internal_path)
 
             # Try with any top-level directory prefix (e.g., "skill-name/SKILL.md")
-            for name, info in infos_by_name.items():
+            for name in namelist:
                 if name.endswith("/" + internal_path) or name == internal_path:
-                    return _read_skill_archive_member(zip_ref, info)
+                    return zip_ref.read(name)
 
             # Not found
             return None

backend/app/gateway/routers/auth.py

@@ -1,6 +1,5 @@
 """Authentication endpoints."""
 
-import asyncio
 import logging
 import os
 import time
@@ -306,7 +305,7 @@ async def login_local(
 async def register(request: Request, response: Response, body: RegisterRequest):
     """Register a new user account (always 'user' role).
 
-    The first admin is created explicitly through /initialize. This endpoint creates regular users.
+    Admin is auto-created on first boot. This endpoint creates regular users.
     Auto-login by setting the session cookie.
     """
     try:
@@ -383,15 +382,9 @@ async def get_me(request: Request):
     return UserResponse(id=str(user.id), email=user.email, system_role=user.system_role, needs_setup=user.needs_setup)
 
 
-# Per-IP cache: ip → (timestamp, result_dict).
+_SETUP_STATUS_COOLDOWN: dict[str, float] = {}
-# Returns the cached result within the TTL instead of 429, because
+_SETUP_STATUS_COOLDOWN_SECONDS = 60
-# the answer (whether an admin exists) rarely changes and returning
-# 429 breaks multi-tab / post-restart reconnection storms.
-_SETUP_STATUS_CACHE: dict[str, tuple[float, dict]] = {}
-_SETUP_STATUS_CACHE_TTL_SECONDS = 60
 _MAX_TRACKED_SETUP_STATUS_IPS = 10000
-_SETUP_STATUS_INFLIGHT: dict[str, asyncio.Task[dict]] = {}
-_SETUP_STATUS_INFLIGHT_GUARD = asyncio.Lock()
 
 
 @router.get("/setup-status")
@@ -399,56 +392,29 @@ async def setup_status(request: Request):
     """Check if an admin account exists. Returns needs_setup=True when no admin exists."""
     client_ip = _get_client_ip(request)
     now = time.time()
-
+    last_check = _SETUP_STATUS_COOLDOWN.get(client_ip, 0)
-    # Return cached result when within TTL — avoids 429 on multi-tab reconnection.
+    elapsed = now - last_check
-    cached = _SETUP_STATUS_CACHE.get(client_ip)
+    if elapsed < _SETUP_STATUS_COOLDOWN_SECONDS:
-    if cached is not None:
+        retry_after = max(1, int(_SETUP_STATUS_COOLDOWN_SECONDS - elapsed))
-        cached_time, cached_result = cached
+        raise HTTPException(
-        if now - cached_time < _SETUP_STATUS_CACHE_TTL_SECONDS:
+            status_code=status.HTTP_429_TOO_MANY_REQUESTS,
-            return cached_result
+            detail="Setup status check is rate limited",
-
+            headers={"Retry-After": str(retry_after)},
-    async with _SETUP_STATUS_INFLIGHT_GUARD:
+        )
-        # Recheck cache after waiting for the inflight guard.
+    # Evict stale entries when dict grows too large to bound memory usage.
-        now = time.time()
+    if len(_SETUP_STATUS_COOLDOWN) >= _MAX_TRACKED_SETUP_STATUS_IPS:
-        cached = _SETUP_STATUS_CACHE.get(client_ip)
+        cutoff = now - _SETUP_STATUS_COOLDOWN_SECONDS
-        if cached is not None:
+        stale = [k for k, t in _SETUP_STATUS_COOLDOWN.items() if t < cutoff]
-            cached_time, cached_result = cached
+        for k in stale:
-            if now - cached_time < _SETUP_STATUS_CACHE_TTL_SECONDS:
+            del _SETUP_STATUS_COOLDOWN[k]
-                return cached_result
+        # If still too large after evicting expired entries, remove oldest half.
-
+        if len(_SETUP_STATUS_COOLDOWN) >= _MAX_TRACKED_SETUP_STATUS_IPS:
-        task = _SETUP_STATUS_INFLIGHT.get(client_ip)
+            by_time = sorted(_SETUP_STATUS_COOLDOWN.items(), key=lambda kv: kv[1])
-        if task is None:
+            for k, _ in by_time[: len(by_time) // 2]:
-            # Evict stale entries when dict grows too large to bound memory usage.
+                del _SETUP_STATUS_COOLDOWN[k]
-            if len(_SETUP_STATUS_CACHE) >= _MAX_TRACKED_SETUP_STATUS_IPS:
+    _SETUP_STATUS_COOLDOWN[client_ip] = now
-                cutoff = now - _SETUP_STATUS_CACHE_TTL_SECONDS
+    admin_count = await get_local_provider().count_admin_users()
-                stale = [k for k, (t, _) in _SETUP_STATUS_CACHE.items() if t < cutoff]
+    return {"needs_setup": admin_count == 0}
-                for k in stale:
-                    del _SETUP_STATUS_CACHE[k]
-                if len(_SETUP_STATUS_CACHE) >= _MAX_TRACKED_SETUP_STATUS_IPS:
-                    by_time = sorted(_SETUP_STATUS_CACHE.items(), key=lambda entry: entry[1][0])
-                    for k, _ in by_time[: len(by_time) // 2]:
-                        del _SETUP_STATUS_CACHE[k]
-
-            async def _compute_setup_status() -> dict:
-                admin_count = await get_local_provider().count_admin_users()
-                return {"needs_setup": admin_count == 0}
-
-            task = asyncio.create_task(_compute_setup_status())
-            _SETUP_STATUS_INFLIGHT[client_ip] = task
-
-    try:
-        result = await task
-    finally:
-        async with _SETUP_STATUS_INFLIGHT_GUARD:
-            if _SETUP_STATUS_INFLIGHT.get(client_ip) is task:
-                del _SETUP_STATUS_INFLIGHT[client_ip]
-
-    # Cache only the stable "initialized" result to avoid stale setup redirects.
-    if result["needs_setup"] is False:
-        _SETUP_STATUS_CACHE[client_ip] = (time.time(), result)
-    else:
-        _SETUP_STATUS_CACHE.pop(client_ip, None)
-    return result
 
 
 class InitializeAdminRequest(BaseModel):

backend/app/gateway/routers/mcp.py

@@ -1,10 +1,9 @@
 import json
 import logging
-import os
 from pathlib import Path
 from typing import Literal
 
-from fastapi import APIRouter, HTTPException, Request, status
+from fastapi import APIRouter, HTTPException
 from pydantic import BaseModel, Field
 
 from deerflow.config.extensions_config import ExtensionsConfig, get_extensions_config, reload_extensions_config
@@ -13,11 +12,6 @@ logger = logging.getLogger(__name__)
 router = APIRouter(prefix="/api", tags=["mcp"])
 
 
-_MCP_STDIO_COMMAND_ALLOWLIST_ENV = "DEER_FLOW_MCP_STDIO_COMMAND_ALLOWLIST"
-_DEFAULT_MCP_STDIO_COMMAND_ALLOWLIST = frozenset({"npx", "uvx"})
-_SHELL_METACHARS = frozenset(";|&`$<>\n\r")
-
-
 class McpOAuthConfigResponse(BaseModel):
     """OAuth configuration for an MCP server."""
 
@@ -69,178 +63,13 @@ class McpConfigUpdateRequest(BaseModel):
     )
 
 
-_MASKED_VALUE = "***"
-
-
-async def _require_admin_user(request: Request) -> None:
-    """Require the authenticated caller to be an admin user.
-
-    ``AuthMiddleware`` normally stamps ``request.state.user`` before the
-    request reaches this router. Falling back to the strict dependency keeps
-    this route safe even in tests or alternative ASGI compositions that mount
-    the router without the global middleware.
-    """
-    user = getattr(request.state, "user", None)
-    if user is None:
-        from app.gateway.deps import get_current_user_from_request
-
-        user = await get_current_user_from_request(request)
-
-    if getattr(user, "system_role", None) != "admin":
-        raise HTTPException(
-            status_code=status.HTTP_403_FORBIDDEN,
-            detail="Admin privileges required to manage MCP configuration.",
-        )
-
-
-def _allowed_stdio_commands() -> set[str]:
-    """Return executable names allowed for API-managed stdio MCP servers."""
-    raw = os.environ.get(_MCP_STDIO_COMMAND_ALLOWLIST_ENV)
-    base = set(_DEFAULT_MCP_STDIO_COMMAND_ALLOWLIST)
-    if raw is None:
-        return base
-    extra = {item.strip() for item in raw.split(",") if item.strip()}
-    return base | extra
-
-
-def _stdio_command_name(command: str | None, *, server_name: str) -> str:
-    """Normalize and validate a stdio command field from the API boundary."""
-    if command is None or not command.strip():
-        raise HTTPException(
-            status_code=status.HTTP_400_BAD_REQUEST,
-            detail=f"MCP server '{server_name}' with stdio transport requires a command.",
-        )
-
-    stripped = command.strip()
-    has_path_separator = "/" in stripped or "\\" in stripped
-    if stripped != command or has_path_separator or any(ch.isspace() for ch in stripped) or any(ch in stripped for ch in _SHELL_METACHARS):
-        raise HTTPException(
-            status_code=status.HTTP_400_BAD_REQUEST,
-            detail=(f"MCP server '{server_name}' command must be a single executable name; put parameters in args instead."),
-        )
-
-    return stripped
-
-
-def _validate_mcp_update_request(request: McpConfigUpdateRequest) -> None:
-    """Validate API-submitted MCP config before it is persisted.
-
-    Local config files can still express arbitrary advanced setups, but the
-    HTTP API is an untrusted boundary. Restricting stdio commands here reduces
-    the blast radius of a compromised authenticated browser session.
-    """
-    allowed_commands = _allowed_stdio_commands()
-    for name, server in request.mcp_servers.items():
-        transport_type = (server.type or "stdio").lower()
-        if transport_type != "stdio":
-            continue
-
-        command_name = _stdio_command_name(server.command, server_name=name)
-        if command_name not in allowed_commands:
-            allowed = ", ".join(sorted(allowed_commands)) or "<none>"
-            raise HTTPException(
-                status_code=status.HTTP_400_BAD_REQUEST,
-                detail=(f"MCP server '{name}' uses disallowed stdio command '{command_name}'. Allowed commands: {allowed}. Configure {_MCP_STDIO_COMMAND_ALLOWLIST_ENV} to extend this list."),
-            )
-
-
-def _mask_server_config(server: McpServerConfigResponse) -> McpServerConfigResponse:
-    """Return a copy of server config with sensitive fields masked.
-
-    Masks env values, header values, and removes OAuth secrets so they
-    are not exposed through the GET API endpoint.
-    """
-    masked_env = {k: _MASKED_VALUE for k in server.env}
-    masked_headers = {k: _MASKED_VALUE for k in server.headers}
-    masked_oauth = None
-    if server.oauth is not None:
-        masked_oauth = server.oauth.model_copy(
-            update={
-                "client_secret": None,
-                "refresh_token": None,
-            }
-        )
-    return server.model_copy(
-        update={
-            "env": masked_env,
-            "headers": masked_headers,
-            "oauth": masked_oauth,
-        }
-    )
-
-
-def _merge_preserving_secrets(
-    incoming: McpServerConfigResponse,
-    existing: McpServerConfigResponse,
-) -> McpServerConfigResponse:
-    """Merge incoming config with existing, preserving secrets masked by GET.
-
-    When the frontend toggles ``enabled`` it round-trips the full config:
-    GET (masked) → modify enabled → PUT (masked values sent back).
-    This function ensures masked values (``***``) are replaced with the
-    real secrets from the current on-disk config.
-
-    ``***`` is only accepted for keys that already exist in *existing*.
-    New keys must provide a real value.
-
-    For OAuth secrets, ``None`` means "preserve the existing stored value"
-    so masked GET responses can be safely round-tripped. To explicitly clear
-    a stored secret, clients may send an empty string, which is converted
-    to ``None`` before persisting.
-    """
-    merged_env = {}
-    for k, v in incoming.env.items():
-        if v == _MASKED_VALUE:
-            if k in existing.env:
-                merged_env[k] = existing.env[k]
-            else:
-                raise HTTPException(
-                    status_code=400,
-                    detail=f"Cannot set env key '{k}' to masked value '***'; provide a real value.",
-                )
-        else:
-            merged_env[k] = v
-
-    merged_headers = {}
-    for k, v in incoming.headers.items():
-        if v == _MASKED_VALUE:
-            if k in existing.headers:
-                merged_headers[k] = existing.headers[k]
-            else:
-                raise HTTPException(
-                    status_code=400,
-                    detail=f"Cannot set header '{k}' to masked value '***'; provide a real value.",
-                )
-        else:
-            merged_headers[k] = v
-
-    merged_oauth = incoming.oauth
-    if incoming.oauth is not None and existing.oauth is not None:
-        # None = preserve (masked round-trip), "" = explicitly clear, else = new value
-        merged_client_secret = existing.oauth.client_secret if incoming.oauth.client_secret is None else (None if incoming.oauth.client_secret == "" else incoming.oauth.client_secret)
-        merged_refresh_token = existing.oauth.refresh_token if incoming.oauth.refresh_token is None else (None if incoming.oauth.refresh_token == "" else incoming.oauth.refresh_token)
-        merged_oauth = incoming.oauth.model_copy(
-            update={
-                "client_secret": merged_client_secret,
-                "refresh_token": merged_refresh_token,
-            }
-        )
-    return incoming.model_copy(
-        update={
-            "env": merged_env,
-            "headers": merged_headers,
-            "oauth": merged_oauth,
-        }
-    )
-
-
 @router.get(
     "/mcp/config",
     response_model=McpConfigResponse,
     summary="Get MCP Configuration",
     description="Retrieve the current Model Context Protocol (MCP) server configurations.",
 )
-async def get_mcp_configuration(request: Request) -> McpConfigResponse:
+async def get_mcp_configuration() -> McpConfigResponse:
     """Get the current MCP configuration.
 
     Returns:
@@ -254,19 +83,16 @@ async def get_mcp_configuration(request: Request) -> McpConfigResponse:
                     "enabled": true,
                     "command": "npx",
                     "args": ["-y", "@modelcontextprotocol/server-github"],
-                    "env": {"GITHUB_TOKEN": "***"},
+                    "env": {"GITHUB_TOKEN": "ghp_xxx"},
                     "description": "GitHub MCP server for repository operations"
                 }
             }
         }
         ```
     """
-    await _require_admin_user(request)
-
     config = get_extensions_config()
 
-    servers = {name: _mask_server_config(McpServerConfigResponse(**server.model_dump())) for name, server in config.mcp_servers.items()}
+    return McpConfigResponse(mcp_servers={name: McpServerConfigResponse(**server.model_dump()) for name, server in config.mcp_servers.items()})
-    return McpConfigResponse(mcp_servers=servers)
 
 
 @router.put(
@@ -275,7 +101,7 @@ async def get_mcp_configuration(request: Request) -> McpConfigResponse:
     summary="Update MCP Configuration",
     description="Update Model Context Protocol (MCP) server configurations and save to file.",
 )
-async def update_mcp_configuration(request: Request, body: McpConfigUpdateRequest) -> McpConfigResponse:
+async def update_mcp_configuration(request: McpConfigUpdateRequest) -> McpConfigResponse:
     """Update the MCP configuration.
 
     This will:
@@ -308,9 +134,6 @@ async def update_mcp_configuration(request: Request, body: McpConfigUpdateReques
         ```
     """
     try:
-        await _require_admin_user(request)
-        _validate_mcp_update_request(body)
-
         # Get the current config path (or determine where to save it)
         config_path = ExtensionsConfig.resolve_config_path()
 
@@ -319,39 +142,14 @@ async def update_mcp_configuration(request: Request, body: McpConfigUpdateReques
             config_path = Path.cwd().parent / "extensions_config.json"
             logger.info(f"No existing extensions config found. Creating new config at: {config_path}")
 
-        # Load current config to preserve skills
+        # Load current config to preserve skills configuration
         current_config = get_extensions_config()
 
-        # Load raw (un-resolved) JSON from disk to use as the merge source.
+        # Convert request to dict format for JSON serialization
-        # This preserves $VAR placeholders in env values and top-level keys
+        config_data = {
-        # like mcpInterceptors that would otherwise be lost.
+            "mcpServers": {name: server.model_dump() for name, server in request.mcp_servers.items()},
-        raw_servers: dict[str, dict] = {}
+            "skills": {name: {"enabled": skill.enabled} for name, skill in current_config.skills.items()},
-        raw_other_keys: dict = {}
+        }
-        if config_path is not None and config_path.exists():
-            with open(config_path, encoding="utf-8") as f:
-                raw_data = json.load(f)
-            raw_servers = raw_data.get("mcpServers", {})
-            # Preserve any top-level keys beyond mcpServers/skills
-            for key, value in raw_data.items():
-                if key not in ("mcpServers", "skills"):
-                    raw_other_keys[key] = value
-
-        # Merge incoming server configs with raw on-disk secrets
-        merged_servers: dict[str, McpServerConfigResponse] = {}
-        for name, incoming in body.mcp_servers.items():
-            raw_server = raw_servers.get(name)
-            if raw_server is not None:
-                merged_servers[name] = _merge_preserving_secrets(
-                    incoming,
-                    McpServerConfigResponse(**raw_server),
-                )
-            else:
-                merged_servers[name] = incoming
-
-        # Build config data preserving all top-level keys from the original file
-        config_data = dict(raw_other_keys)
-        config_data["mcpServers"] = {name: server.model_dump() for name, server in merged_servers.items()}
-        config_data["skills"] = {name: {"enabled": skill.enabled} for name, skill in current_config.skills.items()}
 
         # Write the configuration to file
         with open(config_path, "w", encoding="utf-8") as f:
@@ -359,15 +157,13 @@ async def update_mcp_configuration(request: Request, body: McpConfigUpdateReques
 
         logger.info(f"MCP configuration updated and saved to: {config_path}")
 
-        # Reload the Gateway configuration and update the global cache. The
+        # NOTE: No need to reload/reset cache here - LangGraph Server (separate process)
-        # agent runtime lives in Gateway, so this keeps API reads and tool
+        # will detect config file changes via mtime and reinitialize MCP tools automatically
-        # execution aligned after extensions_config.json changes.
+
-        reloaded_config = reload_extensions_config()
+        # Reload the configuration and update the global cache
-        servers = {name: _mask_server_config(McpServerConfigResponse(**server.model_dump())) for name, server in reloaded_config.mcp_servers.items()}
+        reloaded_config = reload_extensions_config()
-        return McpConfigResponse(mcp_servers=servers)
+        return McpConfigResponse(mcp_servers={name: McpServerConfigResponse(**server.model_dump()) for name, server in reloaded_config.mcp_servers.items()})
 
-    except HTTPException:
-        raise
     except Exception as e:
         logger.error(f"Failed to update MCP configuration: {e}", exc_info=True)
         raise HTTPException(status_code=500, detail=f"Failed to update MCP configuration: {str(e)}")

backend/app/gateway/routers/runs.py

@@ -7,6 +7,7 @@ is reused so that conversation history is preserved across calls.
 
 from __future__ import annotations
 
+import asyncio
 import logging
 import uuid
 
@@ -15,9 +16,8 @@ from fastapi.responses import StreamingResponse
 
 from app.gateway.authz import require_permission
 from app.gateway.deps import get_checkpointer, get_feedback_repo, get_run_event_store, get_run_manager, get_run_store, get_stream_bridge
-from app.gateway.pagination import trim_run_message_page
 from app.gateway.routers.thread_runs import RunCreateRequest
-from app.gateway.services import sse_consumer, start_run, wait_for_run_completion
+from app.gateway.services import sse_consumer, start_run
 from deerflow.runtime import serialize_channel_values
 
 logger = logging.getLogger(__name__)
@@ -66,25 +66,24 @@ async def stateless_wait(body: RunCreateRequest, request: Request) -> dict:
     Otherwise a new temporary thread is created.
     """
     thread_id = _resolve_thread_id(body)
-    bridge = get_stream_bridge(request)
-    run_mgr = get_run_manager(request)
     record = await start_run(body, thread_id, request)
 
-    completed = True
     if record.task is not None:
-        completed = await wait_for_run_completion(bridge, record, request, run_mgr)
-
-    if completed:
-        checkpointer = get_checkpointer(request)
-        config = {"configurable": {"thread_id": thread_id}}
         try:
-            checkpoint_tuple = await checkpointer.aget_tuple(config)
+            await record.task
-            if checkpoint_tuple is not None:
+        except asyncio.CancelledError:
-                checkpoint = getattr(checkpoint_tuple, "checkpoint", {}) or {}
+            pass
-                channel_values = checkpoint.get("channel_values", {})
+
-                return serialize_channel_values(channel_values)
+    checkpointer = get_checkpointer(request)
-        except Exception:
+    config = {"configurable": {"thread_id": thread_id}}
-            logger.exception("Failed to fetch final state for run %s", record.run_id)
+    try:
+        checkpoint_tuple = await checkpointer.aget_tuple(config)
+        if checkpoint_tuple is not None:
+            checkpoint = getattr(checkpoint_tuple, "checkpoint", {}) or {}
+            channel_values = checkpoint.get("channel_values", {})
+            return serialize_channel_values(channel_values)
+    except Exception:
+        logger.exception("Failed to fetch final state for run %s", record.run_id)
 
     return {"status": record.status.value, "error": record.error}
 
@@ -130,7 +129,8 @@ async def run_messages(
         before_seq=before_seq,
         after_seq=after_seq,
     )
-    data, has_more = trim_run_message_page(rows, limit=limit, after_seq=after_seq)
+    has_more = len(rows) > limit
+    data = rows[:limit] if has_more else rows
     return {"data": data, "has_more": has_more}
 
 

backend/app/gateway/routers/suggestions.py

@@ -1,6 +1,5 @@
 import json
 import logging
-import re
 
 from fastapi import APIRouter, Depends, Request
 from langchain_core.messages import HumanMessage, SystemMessage
@@ -31,31 +30,6 @@ class SuggestionsResponse(BaseModel):
     suggestions: list[str] = Field(default_factory=list, description="Suggested follow-up questions")
 
 
-# Matches a complete <think>...</think> block (case-insensitive, spans newlines).
-_THINK_BLOCK_RE = re.compile(r"<think\b[^>]*>.*?</think\s*>", re.IGNORECASE | re.DOTALL)
-# Matches a dangling, unclosed <think> (model truncated at max_tokens mid-thought).
-_OPEN_THINK_RE = re.compile(r"<think\b[^>]*>", re.IGNORECASE)
-
-
-def _strip_think_blocks(text: str) -> str:
-    """Remove reasoning-model ``<think>...</think>`` blocks from the response.
-
-    Reasoning models such as MiniMax-M3 inline their chain-of-thought into the
-    message ``content`` wrapped in ``<think>...</think>`` (``reasoning_split``
-    defaults to false), rather than exposing a separate ``reasoning_content``
-    field. The thinking text frequently contains ``[`` / ``]`` characters, which
-    corrupted the downstream ``find('[')`` / ``rfind(']')`` JSON extraction and
-    produced empty suggestions. We strip the reasoning before parsing so only
-    the actual answer remains.
-    """
-    text = _THINK_BLOCK_RE.sub("", text)
-    # Drop any unclosed <think> (and everything after it) left by truncation.
-    open_match = _OPEN_THINK_RE.search(text)
-    if open_match:
-        text = text[: open_match.start()]
-    return text.strip()
-
-
 def _strip_markdown_code_fence(text: str) -> str:
     stripped = text.strip()
     if not stripped.startswith("```"):
@@ -67,8 +41,7 @@ def _strip_markdown_code_fence(text: str) -> str:
 
 
 def _parse_json_string_list(text: str) -> list[str] | None:
-    candidate = _strip_think_blocks(text)
+    candidate = _strip_markdown_code_fence(text)
-    candidate = _strip_markdown_code_fence(candidate)
     start = candidate.find("[")
     end = candidate.rfind("]")
     if start == -1 or end == -1 or end <= start:

backend/app/gateway/routers/thread_runs.py

@@ -21,9 +21,8 @@ from pydantic import BaseModel, Field
 
 from app.gateway.authz import require_permission
 from app.gateway.deps import get_checkpointer, get_current_user, get_feedback_repo, get_run_event_store, get_run_manager, get_run_store, get_stream_bridge
-from app.gateway.pagination import trim_run_message_page
+from app.gateway.services import sse_consumer, start_run
-from app.gateway.services import sse_consumer, start_run, wait_for_run_completion
+from deerflow.runtime import RunRecord, serialize_channel_values
-from deerflow.runtime import RunRecord, RunStatus, serialize_channel_values
 
 logger = logging.getLogger(__name__)
 router = APIRouter(prefix="/api/threads", tags=["runs"])
@@ -67,35 +66,6 @@ class RunResponse(BaseModel):
     multitask_strategy: str = "reject"
     created_at: str = ""
     updated_at: str = ""
-    total_input_tokens: int = 0
-    total_output_tokens: int = 0
-    total_tokens: int = 0
-    llm_call_count: int = 0
-    lead_agent_tokens: int = 0
-    subagent_tokens: int = 0
-    middleware_tokens: int = 0
-    message_count: int = 0
-
-
-class ThreadTokenUsageModelBreakdown(BaseModel):
-    tokens: int = 0
-    runs: int = 0
-
-
-class ThreadTokenUsageCallerBreakdown(BaseModel):
-    lead_agent: int = 0
-    subagent: int = 0
-    middleware: int = 0
-
-
-class ThreadTokenUsageResponse(BaseModel):
-    thread_id: str
-    total_tokens: int = 0
-    total_input_tokens: int = 0
-    total_output_tokens: int = 0
-    total_runs: int = 0
-    by_model: dict[str, ThreadTokenUsageModelBreakdown] = Field(default_factory=dict)
-    by_caller: ThreadTokenUsageCallerBreakdown = Field(default_factory=ThreadTokenUsageCallerBreakdown)
 
 
 # ---------------------------------------------------------------------------
@@ -103,12 +73,6 @@ class ThreadTokenUsageResponse(BaseModel):
 # ---------------------------------------------------------------------------
 
 
-def _cancel_conflict_detail(run_id: str, record: RunRecord) -> str:
-    if record.status in (RunStatus.pending, RunStatus.running):
-        return f"Run {run_id} is not active on this worker and cannot be cancelled"
-    return f"Run {run_id} is not cancellable (status: {record.status.value})"
-
-
 def _record_to_response(record: RunRecord) -> RunResponse:
     return RunResponse(
         run_id=record.run_id,
@@ -120,14 +84,6 @@ def _record_to_response(record: RunRecord) -> RunResponse:
         multitask_strategy=record.multitask_strategy,
         created_at=record.created_at,
         updated_at=record.updated_at,
-        total_input_tokens=record.total_input_tokens,
-        total_output_tokens=record.total_output_tokens,
-        total_tokens=record.total_tokens,
-        llm_call_count=record.llm_call_count,
-        lead_agent_tokens=record.lead_agent_tokens,
-        subagent_tokens=record.subagent_tokens,
-        middleware_tokens=record.middleware_tokens,
-        message_count=record.message_count,
     )
 
 
@@ -176,25 +132,24 @@ async def stream_run(thread_id: str, body: RunCreateRequest, request: Request) -
 @require_permission("runs", "create", owner_check=True, require_existing=True)
 async def wait_run(thread_id: str, body: RunCreateRequest, request: Request) -> dict:
     """Create a run and block until it completes, returning the final state."""
-    bridge = get_stream_bridge(request)
-    run_mgr = get_run_manager(request)
     record = await start_run(body, thread_id, request)
 
-    completed = True
     if record.task is not None:
-        completed = await wait_for_run_completion(bridge, record, request, run_mgr)
-
-    if completed:
-        checkpointer = get_checkpointer(request)
-        config = {"configurable": {"thread_id": thread_id}}
         try:
-            checkpoint_tuple = await checkpointer.aget_tuple(config)
+            await record.task
-            if checkpoint_tuple is not None:
+        except asyncio.CancelledError:
-                checkpoint = getattr(checkpoint_tuple, "checkpoint", {}) or {}
+            pass
-                channel_values = checkpoint.get("channel_values", {})
+
-                return serialize_channel_values(channel_values)
+    checkpointer = get_checkpointer(request)
-        except Exception:
+    config = {"configurable": {"thread_id": thread_id}}
-            logger.exception("Failed to fetch final state for run %s", record.run_id)
+    try:
+        checkpoint_tuple = await checkpointer.aget_tuple(config)
+        if checkpoint_tuple is not None:
+            checkpoint = getattr(checkpoint_tuple, "checkpoint", {}) or {}
+            channel_values = checkpoint.get("channel_values", {})
+            return serialize_channel_values(channel_values)
+    except Exception:
+        logger.exception("Failed to fetch final state for run %s", record.run_id)
 
     return {"status": record.status.value, "error": record.error}
 
@@ -204,8 +159,7 @@ async def wait_run(thread_id: str, body: RunCreateRequest, request: Request) ->
 async def list_runs(thread_id: str, request: Request) -> list[RunResponse]:
     """List all runs for a thread."""
     run_mgr = get_run_manager(request)
-    user_id = await get_current_user(request)
+    records = await run_mgr.list_by_thread(thread_id)
-    records = await run_mgr.list_by_thread(thread_id, user_id=user_id)
     return [_record_to_response(r) for r in records]
 
 
@@ -214,8 +168,7 @@ async def list_runs(thread_id: str, request: Request) -> list[RunResponse]:
 async def get_run(thread_id: str, run_id: str, request: Request) -> RunResponse:
     """Get details of a specific run."""
     run_mgr = get_run_manager(request)
-    user_id = await get_current_user(request)
+    record = run_mgr.get(run_id)
-    record = await run_mgr.get(run_id, user_id=user_id)
     if record is None or record.thread_id != thread_id:
         raise HTTPException(status_code=404, detail=f"Run {run_id} not found")
     return _record_to_response(record)
@@ -238,13 +191,16 @@ async def cancel_run(
     - wait=false: Return immediately with 202
     """
     run_mgr = get_run_manager(request)
-    record = await run_mgr.get(run_id)
+    record = run_mgr.get(run_id)
     if record is None or record.thread_id != thread_id:
         raise HTTPException(status_code=404, detail=f"Run {run_id} not found")
 
     cancelled = await run_mgr.cancel(run_id, action=action)
     if not cancelled:
-        raise HTTPException(status_code=409, detail=_cancel_conflict_detail(run_id, record))
+        raise HTTPException(
+            status_code=409,
+            detail=f"Run {run_id} is not cancellable (status: {record.status.value})",
+        )
 
     if wait and record.task is not None:
         try:
@@ -260,14 +216,12 @@ async def cancel_run(
 @require_permission("runs", "read", owner_check=True)
 async def join_run(thread_id: str, run_id: str, request: Request) -> StreamingResponse:
     """Join an existing run's SSE stream."""
+    bridge = get_stream_bridge(request)
     run_mgr = get_run_manager(request)
-    record = await run_mgr.get(run_id)
+    record = run_mgr.get(run_id)
     if record is None or record.thread_id != thread_id:
         raise HTTPException(status_code=404, detail=f"Run {run_id} not found")
-    if record.store_only:
-        raise HTTPException(status_code=409, detail=f"Run {run_id} is not active on this worker and cannot be streamed")
 
-    bridge = get_stream_bridge(request)
     return StreamingResponse(
         sse_consumer(bridge, record, request, run_mgr),
         media_type="text/event-stream",
@@ -279,12 +233,7 @@ async def join_run(thread_id: str, run_id: str, request: Request) -> StreamingRe
     )
 
 
-# Register GET and POST as separate routes so each method gets a unique OpenAPI
+@router.api_route("/{thread_id}/runs/{run_id}/stream", methods=["GET", "POST"], response_model=None)
-# operationId. ``api_route(methods=["GET", "POST"])`` shares one route registration
-# across both methods, which makes FastAPI emit the same ``operationId`` twice and
-# warn about a duplicate operation id during OpenAPI generation.
-@router.get("/{thread_id}/runs/{run_id}/stream", response_model=None)
-@router.post("/{thread_id}/runs/{run_id}/stream", response_model=None)
 @require_permission("runs", "read", owner_check=True)
 async def stream_existing_run(
     thread_id: str,
@@ -301,18 +250,14 @@ async def stream_existing_run(
     remaining buffered events so the client observes a clean shutdown.
     """
     run_mgr = get_run_manager(request)
-    record = await run_mgr.get(run_id)
+    record = run_mgr.get(run_id)
     if record is None or record.thread_id != thread_id:
         raise HTTPException(status_code=404, detail=f"Run {run_id} not found")
-    if record.store_only and action is None:
-        raise HTTPException(status_code=409, detail=f"Run {run_id} is not active on this worker and cannot be streamed")
 
     # Cancel if an action was requested (stop-button / interrupt flow)
     if action is not None:
         cancelled = await run_mgr.cancel(run_id, action=action)
-        if not cancelled:
+        if cancelled and wait and record.task is not None:
-            raise HTTPException(status_code=409, detail=_cancel_conflict_detail(run_id, record))
-        if wait and record.task is not None:
             try:
                 await record.task
             except (asyncio.CancelledError, Exception):
@@ -403,7 +348,8 @@ async def list_run_messages(
         before_seq=before_seq,
         after_seq=after_seq,
     )
-    data, has_more = trim_run_message_page(rows, limit=limit, after_seq=after_seq)
+    has_more = len(rows) > limit
+    data = rows[:limit] if has_more else rows
     return {"data": data, "has_more": has_more}
 
 
@@ -422,17 +368,10 @@ async def list_run_events(
     return await event_store.list_events(thread_id, run_id, event_types=types, limit=limit)
 
 
-@router.get("/{thread_id}/token-usage", response_model=ThreadTokenUsageResponse)
+@router.get("/{thread_id}/token-usage")
 @require_permission("threads", "read", owner_check=True)
-async def thread_token_usage(
+async def thread_token_usage(thread_id: str, request: Request) -> dict:
-    thread_id: str,
-    request: Request,
-    include_active: bool = Query(default=False, description="Include running run progress snapshots"),
-) -> ThreadTokenUsageResponse:
     """Thread-level token usage aggregation."""
     run_store = get_run_store(request)
-    if include_active:
+    agg = await run_store.aggregate_tokens_by_thread(thread_id)
-        agg = await run_store.aggregate_tokens_by_thread(thread_id, include_active=True)
+    return {"thread_id": thread_id, **agg}
-    else:
-        agg = await run_store.aggregate_tokens_by_thread(thread_id)
-    return ThreadTokenUsageResponse(thread_id=thread_id, **agg)

backend/app/gateway/routers/threads.py

@@ -13,11 +13,11 @@ matching the LangGraph Platform wire format expected by the
 from __future__ import annotations
 
 import logging
+import time
 import uuid
 from typing import Any
 
 from fastapi import APIRouter, HTTPException, Request
-from langgraph.checkpoint.base import empty_checkpoint, uuid6
 from pydantic import BaseModel, Field, field_validator
 
 from app.gateway.authz import require_permission
@@ -26,7 +26,6 @@ from app.gateway.utils import sanitize_log_param
 from deerflow.config.paths import Paths, get_paths
 from deerflow.runtime import serialize_channel_values
 from deerflow.runtime.user_context import get_effective_user_id
-from deerflow.utils.time import coerce_iso, now_iso
 
 logger = logging.getLogger(__name__)
 router = APIRouter(prefix="/api/threads", tags=["threads"])
@@ -90,28 +89,6 @@ class ThreadSearchRequest(BaseModel):
     offset: int = Field(default=0, ge=0, description="Pagination offset")
     status: str | None = Field(default=None, description="Filter by thread status")
 
-    @field_validator("metadata")
-    @classmethod
-    def _validate_metadata_filters(cls, v: dict[str, Any]) -> dict[str, Any]:
-        """Reject filter entries the SQL backend cannot compile.
-
-        Enforces consistent behaviour across SQL and memory backends.
-        See ``deerflow.persistence.json_compat`` for the shared validators.
-        """
-        if not v:
-            return v
-        from deerflow.persistence.json_compat import validate_metadata_filter_key, validate_metadata_filter_value
-
-        bad_entries: list[str] = []
-        for key, value in v.items():
-            if not validate_metadata_filter_key(key):
-                bad_entries.append(f"{key!r} (unsafe key)")
-            elif not validate_metadata_filter_value(value):
-                bad_entries.append(f"{key!r} (unsupported value type {type(value).__name__})")
-        if bad_entries:
-            raise ValueError(f"Invalid metadata filter entries: {', '.join(bad_entries)}")
-        return v
-
 
 class ThreadStateResponse(BaseModel):
     """Response model for thread state."""
@@ -256,7 +233,7 @@ async def create_thread(body: ThreadCreateRequest, request: Request) -> ThreadRe
     checkpointer = get_checkpointer(request)
     thread_store = get_thread_store(request)
     thread_id = body.thread_id or str(uuid.uuid4())
-    now = now_iso()
+    now = time.time()
     # ``body.metadata`` is already stripped of server-reserved keys by
     # ``ThreadCreateRequest._strip_reserved`` — see the model definition.
 
@@ -266,8 +243,8 @@ async def create_thread(body: ThreadCreateRequest, request: Request) -> ThreadRe
         return ThreadResponse(
             thread_id=thread_id,
             status=existing_record.get("status", "idle"),
-            created_at=coerce_iso(existing_record.get("created_at", "")),
+            created_at=str(existing_record.get("created_at", "")),
-            updated_at=coerce_iso(existing_record.get("updated_at", "")),
+            updated_at=str(existing_record.get("updated_at", "")),
             metadata=existing_record.get("metadata", {}),
         )
 
@@ -285,6 +262,8 @@ async def create_thread(body: ThreadCreateRequest, request: Request) -> ThreadRe
     # Write an empty checkpoint so state endpoints work immediately
     config = {"configurable": {"thread_id": thread_id, "checkpoint_ns": ""}}
     try:
+        from langgraph.checkpoint.base import empty_checkpoint
+
         ckpt_metadata = {
             "step": -1,
             "source": "input",
@@ -302,8 +281,8 @@ async def create_thread(body: ThreadCreateRequest, request: Request) -> ThreadRe
     return ThreadResponse(
         thread_id=thread_id,
         status="idle",
-        created_at=now,
+        created_at=str(now),
-        updated_at=now,
+        updated_at=str(now),
         metadata=body.metadata,
     )
 
@@ -316,27 +295,20 @@ async def search_threads(body: ThreadSearchRequest, request: Request) -> list[Th
     (SQL-backed for sqlite/postgres, Store-backed for memory mode).
     """
     from app.gateway.deps import get_thread_store
-    from deerflow.persistence.thread_meta import InvalidMetadataFilterError
 
     repo = get_thread_store(request)
-    try:
+    rows = await repo.search(
-        rows = await repo.search(
+        metadata=body.metadata or None,
-            metadata=body.metadata or None,
+        status=body.status,
-            status=body.status,
+        limit=body.limit,
-            limit=body.limit,
+        offset=body.offset,
-            offset=body.offset,
+    )
-        )
-    except InvalidMetadataFilterError as exc:
-        raise HTTPException(status_code=400, detail=str(exc)) from exc
     return [
         ThreadResponse(
             thread_id=r["thread_id"],
             status=r.get("status", "idle"),
-            # ``coerce_iso`` heals legacy unix-second values that
+            created_at=r.get("created_at", ""),
-            # ``MemoryThreadMetaStore`` historically wrote with ``time.time()``;
+            updated_at=r.get("updated_at", ""),
-            # SQL-backed rows already arrive as ISO strings and pass through.
-            created_at=coerce_iso(r.get("created_at", "")),
-            updated_at=coerce_iso(r.get("updated_at", "")),
             metadata=r.get("metadata", {}),
             values={"title": r["display_name"]} if r.get("display_name") else {},
             interrupts={},
@@ -368,8 +340,8 @@ async def patch_thread(thread_id: str, body: ThreadPatchRequest, request: Reques
     return ThreadResponse(
         thread_id=thread_id,
         status=record.get("status", "idle"),
-        created_at=coerce_iso(record.get("created_at", "")),
+        created_at=str(record.get("created_at", "")),
-        updated_at=coerce_iso(record.get("updated_at", "")),
+        updated_at=str(record.get("updated_at", "")),
         metadata=record.get("metadata", {}),
     )
 
@@ -409,8 +381,8 @@ async def get_thread(thread_id: str, request: Request) -> ThreadResponse:
         record = {
             "thread_id": thread_id,
             "status": "idle",
-            "created_at": coerce_iso(ckpt_meta.get("created_at", "")),
+            "created_at": ckpt_meta.get("created_at", ""),
-            "updated_at": coerce_iso(ckpt_meta.get("updated_at", ckpt_meta.get("created_at", ""))),
+            "updated_at": ckpt_meta.get("updated_at", ckpt_meta.get("created_at", "")),
             "metadata": {k: v for k, v in ckpt_meta.items() if k not in ("created_at", "updated_at", "step", "source", "writes", "parents")},
         }
 
@@ -424,8 +396,8 @@ async def get_thread(thread_id: str, request: Request) -> ThreadResponse:
     return ThreadResponse(
         thread_id=thread_id,
         status=status,
-        created_at=coerce_iso(record.get("created_at", "")),
+        created_at=str(record.get("created_at", "")),
-        updated_at=coerce_iso(record.get("updated_at", "")),
+        updated_at=str(record.get("updated_at", "")),
         metadata=record.get("metadata", {}),
         values=serialize_channel_values(channel_values),
     )
@@ -476,10 +448,10 @@ async def get_thread_state(thread_id: str, request: Request) -> ThreadStateRespo
         values=values,
         next=next_tasks,
         metadata=metadata,
-        checkpoint={"id": checkpoint_id, "ts": coerce_iso(metadata.get("created_at", ""))},
+        checkpoint={"id": checkpoint_id, "ts": str(metadata.get("created_at", ""))},
         checkpoint_id=checkpoint_id,
         parent_checkpoint_id=parent_checkpoint_id,
-        created_at=coerce_iso(metadata.get("created_at", "")),
+        created_at=str(metadata.get("created_at", "")),
         tasks=tasks,
     )
 
@@ -529,28 +501,16 @@ async def update_thread_state(thread_id: str, body: ThreadStateUpdateRequest, re
         channel_values.update(body.values)
 
     checkpoint["channel_values"] = channel_values
-    metadata["updated_at"] = now_iso()
+    metadata["updated_at"] = time.time()
 
     if body.as_node:
         metadata["source"] = "update"
         metadata["step"] = metadata.get("step", 0) + 1
         metadata["writes"] = {body.as_node: body.values}
 
-    # Assign a new checkpoint ID so aput performs an INSERT rather than an
-    # in-place REPLACE of the existing row.  Use uuid6 (time-ordered) rather
-    # than uuid4 (random) so the new ID is always lexicographically greater
-    # than the previous one — LangGraph's checkpointers determine the "latest"
-    # checkpoint by max(checkpoint_ids) string order, matching the uuid6 epoch.
-    checkpoint["id"] = str(uuid6())
-
     # aput requires checkpoint_ns in the config — use the same config used for the
-    # read (which always includes checkpoint_ns=""). The fresh checkpoint ID is
+    # read (which always includes checkpoint_ns="").  Do NOT include checkpoint_id
-    # assigned above via checkpoint["id"]; keep checkpoint_id out of the config so
+    # so that aput generates a fresh checkpoint ID for the new snapshot.
-    # the write is keyed by the new checkpoint payload rather than the prior read.
-    # All supported savers (InMemorySaver, AsyncSqliteSaver, AsyncPostgresSaver)
-    # persist and echo back checkpoint["id"] verbatim — none mint their own — so
-    # the new_config below carries the uuid6 we assigned here. (Regression-locked
-    # by test_update_thread_state_inserts_new_checkpoint_each_call.)
     write_config: dict[str, Any] = {
         "configurable": {
             "thread_id": thread_id,
@@ -569,7 +529,7 @@ async def update_thread_state(thread_id: str, body: ThreadStateUpdateRequest, re
 
     # Sync title changes through the ThreadMetaStore abstraction so /threads/search
     # reflects them immediately in both sqlite and memory backends.
-    if thread_store and body.values and "title" in body.values:
+    if body.values and "title" in body.values:
         new_title = body.values["title"]
         if new_title:  # Skip empty strings and None
             try:
@@ -582,7 +542,7 @@ async def update_thread_state(thread_id: str, body: ThreadStateUpdateRequest, re
         next=[],
         metadata=metadata,
         checkpoint_id=new_checkpoint_id,
-        created_at=coerce_iso(metadata.get("created_at", "")),
+        created_at=str(metadata.get("created_at", "")),
     )
 
 
@@ -649,7 +609,7 @@ async def get_thread_history(thread_id: str, body: ThreadHistoryRequest, request
                     parent_checkpoint_id=parent_id,
                     metadata=user_meta,
                     values=values,
-                    created_at=coerce_iso(metadata.get("created_at", "")),
+                    created_at=str(metadata.get("created_at", "")),
                     next=next_tasks,
                 )
             )

backend/app/gateway/routers/uploads.py

@@ -5,7 +5,7 @@ import os
 import stat
 
 from fastapi import APIRouter, Depends, File, HTTPException, Request, UploadFile
-from pydantic import BaseModel, Field
+from pydantic import BaseModel
 
 from app.gateway.authz import require_permission
 from app.gateway.deps import get_config
@@ -15,15 +15,12 @@ from deerflow.runtime.user_context import get_effective_user_id
 from deerflow.sandbox.sandbox_provider import SandboxProvider, get_sandbox_provider
 from deerflow.uploads.manager import (
     PathTraversalError,
-    UnsafeUploadPathError,
-    claim_unique_filename,
     delete_file_safe,
     enrich_file_listing,
     ensure_uploads_dir,
     get_uploads_dir,
     list_files_in_dir,
     normalize_filename,
-    open_upload_file_no_symlink,
     upload_artifact_url,
     upload_virtual_path,
 )
@@ -39,37 +36,12 @@ DEFAULT_MAX_FILE_SIZE = 50 * 1024 * 1024
 DEFAULT_MAX_TOTAL_SIZE = 100 * 1024 * 1024
 
 
-class UploadedFileInfo(BaseModel):
-    """Uploaded file metadata exposed by upload and list APIs."""
-
-    filename: str
-    size: int
-    path: str
-    virtual_path: str
-    artifact_url: str
-    extension: str | None = None
-    modified: float | None = None
-    original_filename: str | None = None
-    markdown_file: str | None = None
-    markdown_path: str | None = None
-    markdown_virtual_path: str | None = None
-    markdown_artifact_url: str | None = None
-
-
 class UploadResponse(BaseModel):
     """Response model for file upload."""
 
     success: bool
-    files: list[UploadedFileInfo]
+    files: list[dict[str, str]]
     message: str
-    skipped_files: list[str] = Field(default_factory=list)
-
-
-class UploadListResponse(BaseModel):
-    """Response model for uploaded file listing."""
-
-    files: list[UploadedFileInfo]
-    count: int
 
 
 class UploadLimits(BaseModel):
@@ -93,30 +65,11 @@ def _make_file_sandbox_writable(file_path: os.PathLike[str] | str) -> None:
         logger.warning("Skipping sandbox chmod for symlinked upload path: %s", file_path)
         return
 
-    writable_mode = stat.S_IMODE(file_stat.st_mode) | stat.S_IWUSR | stat.S_IWGRP | stat.S_IWOTH | stat.S_IRGRP | stat.S_IROTH
+    writable_mode = stat.S_IMODE(file_stat.st_mode) | stat.S_IWUSR | stat.S_IWGRP | stat.S_IWOTH
     chmod_kwargs = {"follow_symlinks": False} if os.chmod in os.supports_follow_symlinks else {}
     os.chmod(file_path, writable_mode, **chmod_kwargs)
 
 
-def _make_file_sandbox_readable(file_path: os.PathLike[str] | str) -> None:
-    """Ensure uploaded files are readable by the sandbox process.
-
-    For Docker sandboxes (AIO), the gateway writes files as root with 0o600
-    permissions, then bind-mounts the host directory into the container. The
-    sandbox process inside the container runs as a non-root user and cannot
-    read those files without group/other read bits. This function adds
-    ``S_IRGRP | S_IROTH`` so the sandbox can read the uploaded content.
-    """
-    file_stat = os.lstat(file_path)
-    if stat.S_ISLNK(file_stat.st_mode):
-        logger.warning("Skipping sandbox chmod for symlinked upload path: %s", file_path)
-        return
-
-    readable_mode = stat.S_IMODE(file_stat.st_mode) | stat.S_IRGRP | stat.S_IROTH
-    chmod_kwargs = {"follow_symlinks": False} if os.chmod in os.supports_follow_symlinks else {}
-    os.chmod(file_path, readable_mode, **chmod_kwargs)
-
-
 def _uses_thread_data_mounts(sandbox_provider: SandboxProvider) -> bool:
     return bool(getattr(sandbox_provider, "uses_thread_data_mounts", False))
 
@@ -163,18 +116,17 @@ def _cleanup_uploaded_paths(paths: list[os.PathLike[str] | str]) -> None:
             logger.warning("Failed to clean up upload path after rejected request: %s", path, exc_info=True)
 
 
-async def _write_upload_file_with_limits(
+async def _write_upload_file_streaming(
     file: UploadFile,
+    file_path: os.PathLike[str] | str,
     *,
-    uploads_dir: os.PathLike[str] | str,
     display_filename: str,
     max_single_file_size: int,
     max_total_size: int,
     total_size: int,
-) -> tuple[os.PathLike[str] | str, int, int]:
+) -> tuple[int, int]:
     file_size = 0
-    file_path, fh = open_upload_file_no_symlink(uploads_dir, display_filename)
+    with open(file_path, "wb") as output:
-    try:
         while chunk := await file.read(UPLOAD_CHUNK_SIZE):
             file_size += len(chunk)
             total_size += len(chunk)
@@ -182,17 +134,8 @@ async def _write_upload_file_with_limits(
                 raise HTTPException(status_code=413, detail=f"File too large: {display_filename}")
             if total_size > max_total_size:
                 raise HTTPException(status_code=413, detail="Total upload size too large")
-            fh.write(chunk)
+            output.write(chunk)
-    except Exception:
+    return file_size, total_size
-        fh.close()
-        try:
-            os.unlink(file_path)
-        except FileNotFoundError:
-            pass
-        raise
-    else:
-        fh.close()
-    return file_path, file_size, total_size
 
 
 def _auto_convert_documents_enabled(app_config: AppConfig) -> bool:
@@ -234,12 +177,7 @@ async def upload_files(
     uploaded_files = []
     written_paths = []
     sandbox_sync_targets = []
-    skipped_files = []
     total_size = 0
-    # Track filenames within this request so duplicate form parts do not
-    # silently truncate each other. Existing uploads keep the historical
-    # overwrite behavior for a single replacement upload.
-    seen_filenames: set[str] = set()
 
     sandbox_provider = get_sandbox_provider()
     sync_to_sandbox = not _uses_thread_data_mounts(sandbox_provider)
@@ -256,22 +194,22 @@ async def upload_files(
             continue
 
         try:
-            original_filename = normalize_filename(file.filename)
+            safe_filename = normalize_filename(file.filename)
-            safe_filename = claim_unique_filename(original_filename, seen_filenames)
         except ValueError:
             logger.warning(f"Skipping file with unsafe filename: {file.filename!r}")
             continue
 
         try:
-            file_path, file_size, total_size = await _write_upload_file_with_limits(
+            file_path = uploads_dir / safe_filename
+            written_paths.append(file_path)
+            file_size, total_size = await _write_upload_file_streaming(
                 file,
-                uploads_dir=uploads_dir,
+                file_path,
                 display_filename=safe_filename,
                 max_single_file_size=limits.max_file_size,
                 max_total_size=limits.max_total_size,
                 total_size=total_size,
             )
-            written_paths.append(file_path)
 
             virtual_path = upload_virtual_path(safe_filename)
 
@@ -280,13 +218,11 @@ async def upload_files(
 
             file_info = {
                 "filename": safe_filename,
-                "size": file_size,
+                "size": str(file_size),
                 "path": str(sandbox_uploads / safe_filename),
                 "virtual_path": virtual_path,
                 "artifact_url": upload_artifact_url(thread_id, safe_filename),
             }
-            if safe_filename != original_filename:
-                file_info["original_filename"] = original_filename
 
             logger.info(f"Saved file: {safe_filename} ({file_size} bytes) to {file_info['path']}")
 
@@ -310,39 +246,20 @@ async def upload_files(
         except HTTPException as e:
             _cleanup_uploaded_paths(written_paths)
             raise e
-        except UnsafeUploadPathError as e:
-            logger.warning("Skipping upload with unsafe destination %s: %s", file.filename, e)
-            skipped_files.append(safe_filename)
-            continue
         except Exception as e:
             logger.error(f"Failed to upload {file.filename}: {e}")
             _cleanup_uploaded_paths(written_paths)
             raise HTTPException(status_code=500, detail=f"Failed to upload {file.filename}: {str(e)}")
 
-    # Uploaded files are created with 0o600 permissions (owner read/write only).
-    # In Docker sandbox deployments the gateway writes as root but the sandbox
-    # process runs as a non-root user (typically UID 1000).  Without group/other
-    # read bits the sandbox cannot access the files — whether the uploads
-    # directory is bind-mounted into the container or synced via
-    # sandbox.update_file.  Always add group/other read bits so every sandbox
-    # configuration can read the uploaded content.
-    for file_path in written_paths:
-        _make_file_sandbox_readable(file_path)
-
     if sync_to_sandbox:
         for file_path, virtual_path in sandbox_sync_targets:
             _make_file_sandbox_writable(file_path)
             sandbox.update_file(virtual_path, file_path.read_bytes())
 
-    message = f"Successfully uploaded {len(uploaded_files)} file(s)"
-    if skipped_files:
-        message += f"; skipped {len(skipped_files)} unsafe file(s)"
-
     return UploadResponse(
-        success=not skipped_files,
+        success=True,
         files=uploaded_files,
-        message=message,
+        message=f"Successfully uploaded {len(uploaded_files)} file(s)",
-        skipped_files=skipped_files,
     )
 
 
@@ -357,9 +274,9 @@ async def get_upload_limits(
     return _get_upload_limits(config)
 
 
-@router.get("/list", response_model=UploadListResponse)
+@router.get("/list", response_model=dict)
 @require_permission("threads", "read", owner_check=True)
-async def list_uploaded_files(thread_id: str, request: Request) -> UploadListResponse:
+async def list_uploaded_files(thread_id: str, request: Request) -> dict:
     """List all files in a thread's uploads directory."""
     try:
         uploads_dir = get_uploads_dir(thread_id)
@@ -373,7 +290,7 @@ async def list_uploaded_files(thread_id: str, request: Request) -> UploadListRes
     for f in result["files"]:
         f["path"] = str(sandbox_uploads / f["filename"])
 
-    return UploadListResponse(**result)
+    return result
 
 
 @router.delete("/{filename}")

backend/app/gateway/services.py

@@ -15,13 +15,10 @@ from collections.abc import Mapping
 from typing import Any
 
 from fastapi import HTTPException, Request
-from langchain_core.messages import BaseMessage
+from langchain_core.messages import HumanMessage
-from langchain_core.messages.utils import convert_to_messages
 
 from app.gateway.deps import get_run_context, get_run_manager, get_stream_bridge
-from app.gateway.internal_auth import INTERNAL_SYSTEM_ROLE
 from app.gateway.utils import sanitize_log_param
-from deerflow.config.app_config import get_app_config
 from deerflow.runtime import (
     END_SENTINEL,
     HEARTBEAT_SENTINEL,
@@ -34,7 +31,6 @@ from deerflow.runtime import (
     UnsupportedStrategyError,
     run_agent,
 )
-from deerflow.runtime.runs.naming import resolve_root_run_name
 
 logger = logging.getLogger(__name__)
 
@@ -78,35 +74,21 @@ def normalize_stream_modes(raw: list[str] | str | None) -> list[str]:
 
 
 def normalize_input(raw_input: dict[str, Any] | None) -> dict[str, Any]:
-    """Convert LangGraph Platform input format to LangChain state dict.
+    """Convert LangGraph Platform input format to LangChain state dict."""
-
-    Delegates dict→message coercion to ``langchain_core.messages.utils.convert_to_messages``
-    so that ``additional_kwargs`` (e.g. uploaded-file metadata — gh #3132), ``id``,
-    ``name``, and non-human roles (ai/system/tool) survive unchanged.  An earlier
-    hand-rolled version only forwarded ``content`` and collapsed every role to
-    ``HumanMessage``, which silently stripped frontend-supplied attachments.
-
-    Malformed message dicts (missing ``role``/``type``/``content``, unsupported
-    role, etc.) raise ``HTTPException(400)`` with the offending index, instead
-    of bubbling up as a 500.  The gateway is a system boundary, so per-entry
-    validation errors are the right shape for clients to retry against.
-    """
     if raw_input is None:
         return {}
     messages = raw_input.get("messages")
     if messages and isinstance(messages, list):
-        converted: list[Any] = []
+        converted = []
-        for index, msg in enumerate(messages):
+        for msg in messages:
-            if isinstance(msg, BaseMessage):
+            if isinstance(msg, dict):
-                converted.append(msg)
+                role = msg.get("role", msg.get("type", "user"))
-            elif isinstance(msg, dict):
+                content = msg.get("content", "")
-                try:
+                if role in ("user", "human"):
-                    converted.extend(convert_to_messages([msg]))
+                    converted.append(HumanMessage(content=content))
-                except (ValueError, TypeError, NotImplementedError) as exc:
+                else:
-                    raise HTTPException(
+                    # TODO: handle other message types (system, ai, tool)
-                        status_code=400,
+                    converted.append(HumanMessage(content=content))
-                        detail=f"Invalid message at input.messages[{index}]: {exc}",
-                    ) from exc
             else:
                 converted.append(msg)
         return {**raw_input, "messages": converted}
@@ -141,14 +123,7 @@ def merge_run_context_overrides(config: dict[str, Any], context: Mapping[str, An
     """Merge whitelisted keys from ``body.context`` into both ``config['configurable']``
     and ``config['context']`` so they are visible to legacy configurable readers and
     to LangGraph ``ToolRuntime.context`` consumers (e.g. the ``setup_agent`` tool —
-    see issue #2677).
+    see issue #2677)."""
-
-    ``user_id`` is intentionally propagated into ``config['context']`` in addition to
-    the whitelisted keys, so non-web callers (e.g. IM channels) that supply identity in
-    ``body.context`` keep it on ``ToolRuntime.context``. It is merged with
-    ``setdefault`` so a server-authenticated id stamped by
-    :func:`inject_authenticated_user_context` always wins over the client-supplied one.
-    """
     if not context:
         return
     configurable = config.setdefault("configurable", {})
@@ -159,29 +134,6 @@ def merge_run_context_overrides(config: dict[str, Any], context: Mapping[str, An
                 configurable.setdefault(key, context[key])
             if isinstance(runtime_context, dict):
                 runtime_context.setdefault(key, context[key])
-    if "user_id" in context and isinstance(runtime_context, dict):
-        runtime_context.setdefault("user_id", context["user_id"])
-
-
-def inject_authenticated_user_context(config: dict[str, Any], request: Request) -> None:
-    """Stamp the authenticated user into the run context for background tools.
-
-    Tool execution may happen after the request handler has returned, so tools
-    that persist user-scoped files should not rely only on ambient ContextVars.
-    The value comes from server-side auth state, never from client context.
-    """
-
-    user = getattr(request.state, "user", None)
-    user_id = getattr(user, "id", None)
-    if user_id is None:
-        return
-
-    if getattr(user, "system_role", None) == INTERNAL_SYSTEM_ROLE:
-        return
-
-    runtime_context = config.setdefault("context", {})
-    if isinstance(runtime_context, dict):
-        runtime_context["user_id"] = str(user_id)
 
 
 def resolve_agent_factory(assistant_id: str | None):
@@ -264,7 +216,6 @@ def build_run_config(
             target = config.setdefault("configurable", {})
         if target is not None and "agent_name" not in target:
             target["agent_name"] = normalized
-        config.setdefault("run_name", resolve_root_run_name(config, normalized))
     if metadata:
         config.setdefault("metadata", {}).update(metadata)
     return config
@@ -298,23 +249,6 @@ async def start_run(
 
     disconnect = DisconnectMode.cancel if body.on_disconnect == "cancel" else DisconnectMode.continue_
 
-    body_context = getattr(body, "context", None) or {}
-    model_name = body_context.get("model_name")
-
-    # Coerce non-string model_name values to str before truncation.
-    if model_name is not None and not isinstance(model_name, str):
-        model_name = str(model_name)
-
-    # Validate model against the allowlist when a model_name is provided.
-    if model_name:
-        app_config = get_app_config()
-        resolved = app_config.get_model_config(model_name)
-        if resolved is None:
-            raise HTTPException(
-                status_code=400,
-                detail=f"Model {model_name!r} is not in the configured model allowlist",
-            )
-
     try:
         record = await run_mgr.create_or_reject(
             thread_id,
@@ -323,7 +257,6 @@ async def start_run(
             metadata=body.metadata or {},
             kwargs={"input": body.input, "config": body.config},
             multitask_strategy=body.multitask_strategy,
-            model_name=model_name,
         )
     except ConflictError as exc:
         raise HTTPException(status_code=409, detail=str(exc)) from exc
@@ -355,7 +288,6 @@ async def start_run(
     # that carries agent configuration (model_name, thinking_enabled, etc.).
     # Only agent-relevant keys are forwarded; unknown keys (e.g. thread_id) are ignored.
     merge_run_context_overrides(config, getattr(body, "context", None))
-    inject_authenticated_user_context(config, request)
 
     stream_modes = normalize_stream_modes(body.stream_mode)
 
@@ -415,51 +347,3 @@ async def sse_consumer(
         if record.status in (RunStatus.pending, RunStatus.running):
             if record.on_disconnect == DisconnectMode.cancel:
                 await run_mgr.cancel(record.run_id)
-
-
-async def wait_for_run_completion(
-    bridge: StreamBridge,
-    record: RunRecord,
-    request: Request,
-    run_mgr: RunManager,
-) -> bool:
-    """Block until the run publishes ``END_SENTINEL``, honouring on_disconnect.
-
-    The non-streaming ``/wait`` endpoints used to ``await record.task``
-    directly with no disconnect handling.  When the client (or an
-    intermediate HTTP proxy) timed out during a long tool call such as
-    ``pip install``, the handler would swallow ``CancelledError`` and
-    serialize whatever checkpoint happened to exist — masking a half-finished
-    run as a normal completion (issue #3265).
-
-    This helper consumes the same bridge that ``sse_consumer`` does so the
-    wait path shares its disconnect semantics: each wake-up polls
-    ``request.is_disconnected()``; on a real disconnect it cancels the
-    background run when ``record.on_disconnect`` is ``cancel``.  The bridge's
-    heartbeat sentinels guarantee at least one wake-up per
-    ``heartbeat_interval`` even when the agent emits no events for a while.
-
-    Returns:
-        ``True`` when ``END_SENTINEL`` was observed (run reached a terminal
-        state), ``False`` when the loop exited because the client
-        disconnected.  Callers must skip checkpoint serialization on
-        ``False`` so a partial checkpoint is not returned as a normal
-        response.
-    """
-    completed = False
-    try:
-        async for entry in bridge.subscribe(record.run_id):
-            # END_SENTINEL means the run reached a terminal state; honour it
-            # even if the client just disconnected so the caller still serializes
-            # the real final checkpoint.
-            if entry is END_SENTINEL:
-                completed = True
-                return True
-            if await request.is_disconnected():
-                break
-            # Heartbeats and regular events: keep waiting for END_SENTINEL.
-        return completed
-    finally:
-        if not completed and record.status in (RunStatus.pending, RunStatus.running):
-            if record.on_disconnect == DisconnectMode.cancel:
-                await run_mgr.cancel(record.run_id)

backend/debug.py

@@ -79,9 +79,7 @@ async def main():
     from langgraph.runtime import Runtime
 
     from deerflow.agents import make_lead_agent
-    from deerflow.config.paths import get_paths
     from deerflow.mcp import initialize_mcp_tools
-    from deerflow.runtime.user_context import get_effective_user_id
 
     # Initialize MCP tools at startup
     try:
@@ -115,8 +113,6 @@ async def main():
         print("Tip: `uv sync --group dev` to enable arrow-key & history support")
     print("=" * 50)
 
-    seen_artifacts: set[str] = set()
-
     while True:
         try:
             if session:
@@ -138,22 +134,6 @@ async def main():
                 last_message = result["messages"][-1]
                 print(f"\nAgent: {last_message.content}")
 
-            # Show files presented to the user this turn (new artifacts only)
-            artifacts = result.get("artifacts") or []
-            new_artifacts = [p for p in artifacts if p not in seen_artifacts]
-            if new_artifacts:
-                thread_id = config["configurable"]["thread_id"]
-                user_id = get_effective_user_id()
-                paths = get_paths()
-                print("\n[Presented files]")
-                for virtual in new_artifacts:
-                    try:
-                        physical = paths.resolve_virtual_path(thread_id, virtual, user_id=user_id)
-                        print(f"  - {virtual}\n    → {physical}")
-                    except ValueError as exc:
-                        print(f"  - {virtual}    (failed to resolve physical path: {exc})")
-                seen_artifacts.update(new_artifacts)
-
         except (KeyboardInterrupt, EOFError):
             print("\nGoodbye!")
             break

backend/docs/API.md

@@ -6,16 +6,16 @@ This document provides a complete reference for the DeerFlow backend APIs.
 
 DeerFlow backend exposes two sets of APIs:
 
-1. **LangGraph-compatible API** - Agent interactions, threads, and streaming (`/api/langgraph/*`)
+1. **LangGraph API** - Agent interactions, threads, and streaming (`/api/langgraph/*`)
 2. **Gateway API** - Models, MCP, skills, uploads, and artifacts (`/api/*`)
 
 All APIs are accessed through the Nginx reverse proxy at port 2026.
 
-## LangGraph-compatible API
+## LangGraph API
 
 Base URL: `/api/langgraph`
 
-The public LangGraph-compatible API follows LangGraph SDK conventions. In the unified nginx deployment, Gateway owns `/api/langgraph/*` and translates those paths to its native `/api/*` run, thread, and streaming routers.
+The LangGraph API is provided by the LangGraph server and follows the LangGraph SDK conventions.
 
 ### Threads
 
@@ -104,11 +104,17 @@ Content-Type: application/json
 **Recursion Limit:**
 
 `config.recursion_limit` caps the number of graph steps LangGraph will execute
-in a single run. The unified Gateway path defaults to `100` in
+in a single run. The `/api/langgraph/*` endpoints go straight to the LangGraph
-`build_run_config` (see `backend/app/gateway/services.py`), which is a safer
+server and therefore inherit LangGraph's native default of **25**, which is
-starting point for plan-mode or subagent-heavy runs. Clients can still set
+too low for plan-mode or subagent-heavy runs — the agent typically errors out
-`recursion_limit` explicitly in the request body; increase it if you run deeply
+with `GraphRecursionError` after the first round of subagent results comes
-nested subagent graphs.
+back, before the lead agent can synthesize the final answer.
+
+DeerFlow's own Gateway and IM-channel paths mitigate this by defaulting to
+`100` in `build_run_config` (see `backend/app/gateway/services.py`), but
+clients calling the LangGraph API directly must set `recursion_limit`
+explicitly in the request body. `100` matches the Gateway default and is a
+safe starting point; increase it if you run deeply nested subagent graphs.
 
 **Configurable Options:**
 - `model_name` (string): Override the default model
@@ -228,13 +234,10 @@ Get current MCP server configurations.
 GET /api/mcp/config
 ```
 
-Requires an authenticated admin session. Sensitive env/header/OAuth secret
-values are masked in the response.
-
 **Response:**
 ```json
 {
-  "mcp_servers": {
+  "mcpServers": {
     "github": {
       "enabled": true,
       "type": "stdio",
@@ -244,6 +247,13 @@ values are masked in the response.
         "GITHUB_TOKEN": "***"
       },
       "description": "GitHub operations"
+    },
+    "filesystem": {
+      "enabled": false,
+      "type": "stdio",
+      "command": "npx",
+      "args": ["-y", "@modelcontextprotocol/server-filesystem"],
+      "description": "File system access"
     }
   }
 }
@@ -258,15 +268,10 @@ PUT /api/mcp/config
 Content-Type: application/json
 ```
 
-Requires an authenticated admin session. API-managed `stdio` MCP servers may
-only use allowed executable names for `command` (default: `npx`, `uvx`). Set
-`DEER_FLOW_MCP_STDIO_COMMAND_ALLOWLIST` to a comma-separated list when a
-deployment needs additional trusted launchers.
-
 **Request Body:**
 ```json
 {
-  "mcp_servers": {
+  "mcpServers": {
     "github": {
       "enabled": true,
       "type": "stdio",
@@ -284,18 +289,8 @@ deployment needs additional trusted launchers.
 **Response:**
 ```json
 {
-  "mcp_servers": {
+  "success": true,
-    "github": {
+  "message": "MCP configuration updated"
-      "enabled": true,
-      "type": "stdio",
-      "command": "npx",
-      "args": ["-y", "@modelcontextprotocol/server-github"],
-      "env": {
-        "GITHUB_TOKEN": "***"
-      },
-      "description": "GitHub operations"
-    }
-  }
 }
 ```
 
@@ -546,28 +541,14 @@ All APIs return errors in a consistent format:
 
 ## Authentication
 
-DeerFlow enforces authentication for all non-public HTTP routes. Public routes are limited to health/docs metadata and these public auth endpoints:
+Currently, DeerFlow does not implement authentication. All APIs are accessible without credentials.
 
-- `POST /api/v1/auth/initialize` creates the first admin account when no admin exists.
+Note: This is about DeerFlow API authentication. MCP outbound connections can still use OAuth for configured HTTP/SSE MCP servers.
-- `POST /api/v1/auth/login/local` logs in with email/password and sets an HttpOnly `access_token` cookie.
-- `POST /api/v1/auth/register` creates a regular `user` account and sets the session cookie.
-- `POST /api/v1/auth/logout` clears the session cookie.
-- `GET /api/v1/auth/setup-status` reports whether the first admin still needs to be created.
 
-The authenticated auth endpoints are:
+For production deployments, it is recommended to:
-
+1. Use Nginx for basic auth or OAuth integration
-- `GET /api/v1/auth/me` returns the current user.
+2. Deploy behind a VPN or private network
-- `POST /api/v1/auth/change-password` changes password, optionally changes email during setup, increments `token_version`, and reissues the cookie.
+3. Implement custom authentication middleware
-
-Protected state-changing requests also require the CSRF double-submit token: send the `csrf_token` cookie value as the `X-CSRF-Token` header. Login/register/initialize/logout are bootstrap auth endpoints: they are exempt from the double-submit token but still reject hostile browser `Origin` headers.
-
-User isolation is enforced from the authenticated user context:
-
-- Thread metadata is scoped by `threads_meta.user_id`; search/read/write/delete APIs only expose the current user's threads.
-- Thread files live under `{base_dir}/users/{user_id}/threads/{thread_id}/user-data/` and are exposed inside the sandbox as `/mnt/user-data/`.
-- Memory and custom agents are stored under `{base_dir}/users/{user_id}/...`.
-
-Note: MCP outbound connections can still use OAuth for configured HTTP/SSE MCP servers; that is separate from DeerFlow API authentication.
 
 ---
 
@@ -586,13 +567,12 @@ location /api/ {
 
 ---
 
-## Streaming Support
+## WebSocket Support
 
-Gateway's LangGraph-compatible API streams run events with Server-Sent Events (SSE):
+The LangGraph server supports WebSocket connections for real-time streaming. Connect to:
 
-```http
+```
-POST /api/langgraph/threads/{thread_id}/runs/stream
+ws://localhost:2026/api/langgraph/threads/{thread_id}/runs/stream
-Accept: text/event-stream
 ```
 
 ---
@@ -628,21 +608,13 @@ const response = await fetch('/api/models');
 const data = await response.json();
 console.log(data.models);
 
-// Create a run and stream SSE events
+// Using EventSource for streaming
-const streamResponse = await fetch(`/api/langgraph/threads/${threadId}/runs/stream`, {
+const eventSource = new EventSource(
-  method: "POST",
+  `/api/langgraph/threads/${threadId}/runs/stream`
-  headers: {
+);
-    "Content-Type": "application/json",
+eventSource.onmessage = (event) => {
-    Accept: "text/event-stream",
+  console.log(JSON.parse(event.data));
-  },
+};
-  body: JSON.stringify({
-    input: { messages: [{ role: "user", content: "Hello" }] },
-    stream_mode: ["values", "messages-tuple", "custom"],
-  }),
-});
-
-const reader = streamResponse.body?.getReader();
-// Decode and parse SSE frames from reader in your client code.
 ```
 
 ### cURL Examples
@@ -677,7 +649,7 @@ curl -X POST http://localhost:2026/api/langgraph/threads/abc123/runs \
   }'
 ```
 
-> The unified Gateway path defaults `config.recursion_limit` to 100 for
+> The `/api/langgraph/*` endpoints bypass DeerFlow's Gateway and inherit
-> plan-mode and subagent-heavy runs. Clients may still set
+> LangGraph's native `recursion_limit` default of 25, which is too low for
-> `config.recursion_limit` explicitly — see the [Create Run](#create-run)
+> plan-mode or subagent runs. Set `config.recursion_limit` explicitly — see
-> section for details.
+> the [Create Run](#create-run) section for details.

backend/docs/ARCHITECTURE.md

@@ -14,28 +14,30 @@ This document provides a comprehensive overview of the DeerFlow backend architec
 │                          Nginx (Port 2026)                               │
 │                    Unified Reverse Proxy Entry Point                      │
 │  ┌────────────────────────────────────────────────────────────────────┐  │
-│  │  /api/langgraph/*  →  Gateway LangGraph-compatible runtime (8001)  │  │
+│  │  /api/langgraph/*  →  LangGraph Server (2024)                      │  │
-│  │  /api/*            →  Gateway REST APIs (8001)                     │  │
+│  │  /api/*            →  Gateway API (8001)                           │  │
 │  │  /*                →  Frontend (3000)                               │  │
 │  └────────────────────────────────────────────────────────────────────┘  │
 └─────────────────────────────────┬────────────────────────────────────────┘
                                   │
-          ┌───────────────────────┴───────────────────────┐
+          ┌───────────────────────┼───────────────────────┐
-          │                                               │
+          │                       │                       │
-          ▼                                               ▼
+          ▼                       ▼                       ▼
-┌─────────────────────────────────────────────┐ ┌─────────────────────┐
+┌─────────────────────┐ ┌─────────────────────┐ ┌─────────────────────┐
-│              Gateway API                    │ │     Frontend        │
+│   LangGraph Server  │ │    Gateway API      │ │     Frontend        │
-│              (Port 8001)                    │ │    (Port 3000)      │
+│     (Port 2024)     │ │    (Port 8001)      │ │    (Port 3000)      │
-│                                             │ │                     │
+│                     │ │                     │ │                     │
-│  - LangGraph-compatible runs/threads API    │ │  - Next.js App      │
+│  - Agent Runtime    │ │  - Models API       │ │  - Next.js App      │
-│  - Embedded Agent Runtime                   │ │  - React UI         │
+│  - Thread Mgmt      │ │  - MCP Config       │ │  - React UI         │
-│  - SSE Streaming                            │ │  - Chat Interface   │
+│  - SSE Streaming    │ │  - Skills Mgmt      │ │  - Chat Interface   │
-│  - Checkpointing                            │ │                     │
+│  - Checkpointing    │ │  - File Uploads     │ │                     │
-│  - Models, MCP, Skills, Uploads, Artifacts  │ │                     │
+│                     │ │  - Thread Cleanup   │ │                     │
-│  - Thread Cleanup                           │ │                     │
+│                     │ │  - Artifacts        │ │                     │
-└─────────────────────────────────────────────┘ └─────────────────────┘
+└─────────────────────┘ └─────────────────────┘ └─────────────────────┘
-          │
+          │                       │
-          ▼
+          │     ┌─────────────────┘
+          │     │
+          ▼     ▼
 ┌──────────────────────────────────────────────────────────────────────────┐
 │                         Shared Configuration                              │
 │  ┌─────────────────────────┐  ┌────────────────────────────────────────┐ │
@@ -50,9 +52,9 @@ This document provides a comprehensive overview of the DeerFlow backend architec
 
 ## Component Details
 
-### Gateway Embedded Agent Runtime
+### LangGraph Server
 
-The agent runtime is embedded in the FastAPI Gateway and built on LangGraph for robust multi-agent workflow orchestration. Nginx rewrites `/api/langgraph/*` to Gateway's native `/api/*` routes, so the public API remains compatible with LangGraph SDK clients without running a separate LangGraph server.
+The LangGraph server is the core agent runtime, built on LangGraph for robust multi-agent workflow orchestration.
 
 **Entry Point**: `packages/harness/deerflow/agents/lead_agent/agent.py:make_lead_agent`
 
@@ -63,8 +65,7 @@ The agent runtime is embedded in the FastAPI Gateway and built on LangGraph for
 - Tool execution orchestration
 - SSE streaming for real-time responses
 
-**Graph registry**: `langgraph.json` remains available for tooling, Studio, or direct LangGraph Server compatibility.
+**Configuration**: `langgraph.json`
-It is not the default service entrypoint; scripts and Docker deployments run the Gateway embedded runtime.
 
 ```json
 {
@@ -77,13 +78,12 @@ It is not the default service entrypoint; scripts and Docker deployments run the
 
 ### Gateway API
 
-FastAPI application providing REST endpoints plus the public LangGraph-compatible `/api/langgraph/*` runtime routes.
+FastAPI application providing REST endpoints for non-agent operations.
 
 **Entry Point**: `app/gateway/app.py`
 
 **Routers**:
 - `models.py` - `/api/models` - Model listing and details
-- `thread_runs.py` / `runs.py` - `/api/threads/{id}/runs`, `/api/runs/*` - LangGraph-compatible runs and streaming
 - `mcp.py` - `/api/mcp` - MCP server configuration
 - `skills.py` - `/api/skills` - Skills management
 - `uploads.py` - `/api/threads/{id}/uploads` - File upload
@@ -91,7 +91,7 @@ FastAPI application providing REST endpoints plus the public LangGraph-compatibl
 - `artifacts.py` - `/api/threads/{id}/artifacts` - Artifact serving
 - `suggestions.py` - `/api/threads/{id}/suggestions` - Follow-up suggestion generation
 
-The web conversation delete flow first deletes Gateway-managed thread state through the LangGraph-compatible route, then the Gateway `threads.py` router removes DeerFlow-managed filesystem data via `Paths.delete_thread_dir()`.
+The web conversation delete flow is now split across both backend surfaces: LangGraph handles `DELETE /api/langgraph/threads/{thread_id}` for thread state, then the Gateway `threads.py` router removes DeerFlow-managed filesystem data via `Paths.delete_thread_dir()`.
 
 ### Agent Architecture
 
@@ -353,10 +353,10 @@ SKILL.md Format:
    POST /api/langgraph/threads/{thread_id}/runs
    {"input": {"messages": [{"role": "user", "content": "Hello"}]}}
 
-2. Nginx → Gateway API (8001)
+2. Nginx → LangGraph Server (2024)
-   `/api/langgraph/*` is rewritten to Gateway's LangGraph-compatible `/api/*` routes
+   Proxied to LangGraph server
 
-3. Gateway embedded runtime
+3. LangGraph Server
    a. Load/create thread state
    b. Execute middleware chain:
       - ThreadDataMiddleware: Set up paths
@@ -412,7 +412,7 @@ SKILL.md Format:
 ### Thread Cleanup Flow
 
 ```
-1. Client deletes conversation via the LangGraph-compatible Gateway route
+1. Client deletes conversation via LangGraph
    DELETE /api/langgraph/threads/{thread_id}
 
 2. Web UI follows up with Gateway cleanup

backend/docs/AUTH_DESIGN.md

@@ -1,331 +0,0 @@
-# 用户认证与隔离设计
-
-本文档描述 DeerFlow 当前内置认证模块的设计，而不是历史 RFC。它覆盖浏览器登录、API 认证、CSRF、用户隔离、首次初始化、密码重置、内部调用和升级迁移。
-
-## 设计目标
-
-认证模块的核心目标是把 DeerFlow 从“本地单用户工具”提升为“可多用户部署的 agent runtime”，并让用户身份贯穿 HTTP API、LangGraph-compatible runtime、文件系统、memory、自定义 agent 和反馈数据。
-
-设计约束：
-
-- 默认强制认证：除健康检查、文档和 auth bootstrap 端点外，HTTP 路由都必须有有效 session。
-- 服务端持有所有权：客户端 metadata 不能声明 `user_id` 或 `owner_id`。
-- 隔离默认开启：repository（仓储）、文件路径、memory、agent 配置默认按当前用户解析。
-- 旧数据可升级：无认证版本留下的 thread 可以在 admin 存在后迁移到 admin。
-- 密码不进日志：首次初始化由操作者设置密码；`reset_admin` 只写 0600 凭据文件。
-
-非目标：
-
-- 当前 OAuth 端点只是占位，尚未实现第三方登录。
-- 当前用户角色只有 `admin` 和 `user`，尚未实现细粒度 RBAC。
-- 当前登录限速是进程内字典，多 worker 下不是全局精确限速。
-
-## 核心模型
-
-```mermaid
-graph TB
-  classDef actor fill:#D8CFC4,stroke:#6E6259,color:#2F2A26;
-  classDef api fill:#C9D7D2,stroke:#5D706A,color:#21302C;
-  classDef state fill:#D7D3E8,stroke:#6B6680,color:#29263A;
-  classDef data fill:#E5D2C4,stroke:#806A5B,color:#30251E;
-
-  Browser["Browser — access_token cookie and csrf_token cookie"]:::actor
-  AuthMiddleware["AuthMiddleware — strict session gate"]:::api
-  CSRFMiddleware["CSRFMiddleware — double-submit token and Origin check"]:::api
-  AuthRoutes["Auth routes — initialize login register logout me change-password"]:::api
-  UserContext["Current user ContextVar — request-scoped identity"]:::state
-  Repositories["Repositories — AUTO resolves user_id from context"]:::state
-  Files["Filesystem — users/{user_id}/threads/{thread_id}/user-data"]:::data
-  Memory["Memory and agents — users/{user_id}/memory.json and agents"]:::data
-
-  Browser --> AuthMiddleware
-  Browser --> CSRFMiddleware
-  AuthMiddleware --> AuthRoutes
-  AuthMiddleware --> UserContext
-  UserContext --> Repositories
-  UserContext --> Files
-  UserContext --> Memory
-```
-
-### 用户表
-
-用户记录定义在 `app.gateway.auth.models.User`，持久化到 `users` 表。关键字段：
-
-| 字段 | 语义 |
-|---|---|
-| `id` | 用户主键，JWT `sub` 使用该值 |
-| `email` | 唯一登录名 |
-| `password_hash` | bcrypt hash，OAuth 用户可为空 |
-| `system_role` | `admin` 或 `user` |
-| `needs_setup` | reset 后要求用户完成邮箱 / 密码设置 |
-| `token_version` | 改密码或 reset 时递增，用于废弃旧 JWT |
-
-### 运行时身份
-
-认证成功后，`AuthMiddleware` 把用户同时写入：
-
-- `request.state.user`
-- `request.state.auth`
-- `deerflow.runtime.user_context` 的 `ContextVar`
-
-`ContextVar` 是这里的核心边界。上层 Gateway 负责写入身份，下层 persistence / file path 只读取结构化的当前用户，不反向依赖 `app.gateway.auth` 具体类型。
-
-可以把 repository 调用的用户参数理解成一个三态 ADT：
-
-```scala
-enum UserScope:
-  case AutoFromContext
-  case Explicit(userId: String)
-  case BypassForMigration
-```
-
-对应 Python 实现是 `AUTO | str | None`：
-
-- `AUTO`：从 `ContextVar` 解析当前用户；没有上下文则抛错。
-- `str`：显式指定用户，主要用于测试或管理脚本。
-- `None`：跳过用户过滤，只允许迁移脚本或 admin CLI 使用。
-
-## 登录与初始化流程
-
-### 首次初始化
-
-首次启动时，如果没有 admin，服务不会自动创建账号，只记录日志提示访问 `/setup`。
-
-流程：
-
-1. 用户访问 `/setup`。
-2. 前端调用 `GET /api/v1/auth/setup-status`。
-3. 如果返回 `{"needs_setup": true}`，前端展示创建 admin 表单。
-4. 表单提交 `POST /api/v1/auth/initialize`。
-5. 服务端确认当前没有 admin，创建 `system_role="admin"`、`needs_setup=false` 的用户。
-6. 服务端设置 `access_token` HttpOnly cookie，用户进入 workspace。
-
-`/api/v1/auth/initialize` 只在没有 admin 时可用。并发初始化由数据库唯一约束兜底，失败方返回 409。
-
-### 普通登录
-
-`POST /api/v1/auth/login/local` 使用 `OAuth2PasswordRequestForm`：
-
-- `username` 是邮箱。
-- `password` 是密码。
-- 成功后签发 JWT，放入 `access_token` HttpOnly cookie。
-- 响应体只返回 `expires_in` 和 `needs_setup`，不返回 token。
-
-登录失败会按客户端 IP 计数。IP 解析只在 TCP peer 属于 `AUTH_TRUSTED_PROXIES` 时信任 `X-Real-IP`，不使用 `X-Forwarded-For`。
-
-### 注册
-
-`POST /api/v1/auth/register` 创建普通 `user`，并自动登录。
-
-当前实现允许在没有 admin 时注册普通用户，但 `setup-status` 仍会返回 `needs_setup=true`，因为 admin 仍不存在。这是当前产品策略边界：如果后续要求“必须先初始化 admin 才能注册普通用户”，需要在 `/register` 增加 admin-exists gate。
-
-### 改密码与 reset setup
-
-`POST /api/v1/auth/change-password` 需要当前密码和新密码：
-
-- 校验当前密码。
-- 更新 bcrypt hash。
-- `token_version += 1`，使旧 JWT 立即失效。
-- 重新签发 cookie。
-- 如果 `needs_setup=true` 且传了 `new_email`，则更新邮箱并清除 `needs_setup`。
-
-`python -m app.gateway.auth.reset_admin` 会：
-
-- 找到 admin 或指定邮箱用户。
-- 生成随机密码。
-- 更新密码 hash。
-- `token_version += 1`。
-- 设置 `needs_setup=true`。
-- 写入 `.deer-flow/admin_initial_credentials.txt`，权限 `0600`。
-
-命令行只输出凭据文件路径，不输出明文密码。
-
-## HTTP 认证边界
-
-`AuthMiddleware` 是 fail-closed（默认拒绝）的全局认证门。
-
-公开路径：
-
-- `/health`
-- `/docs`
-- `/redoc`
-- `/openapi.json`
-- `/api/v1/auth/login/local`
-- `/api/v1/auth/register`
-- `/api/v1/auth/logout`
-- `/api/v1/auth/setup-status`
-- `/api/v1/auth/initialize`
-
-其余路径都要求有效 `access_token` cookie。存在 cookie 但 JWT 无效、过期、用户不存在或 `token_version` 不匹配时，直接返回 401，而不是让请求穿透到业务路由。
-
-路由级别的 owner check 由 `require_permission(..., owner_check=True)` 完成：
-
-- 读类请求允许旧的未追踪 legacy thread 兼容读取。
-- 写 / 删除类请求使用 `require_existing=True`，要求 thread row 存在且属于当前用户，避免删除后缺 row 导致其他用户误通过。
-
-## CSRF 设计
-
-DeerFlow 使用 Double Submit Cookie：
-
-- 服务端设置 `csrf_token` cookie。
-- 前端 state-changing 请求发送同值 `X-CSRF-Token` header。
-- 服务端用 `secrets.compare_digest` 比较 cookie/header。
-
-需要 CSRF 的方法：
-
-- `POST`
-- `PUT`
-- `DELETE`
-- `PATCH`
-
-auth bootstrap 端点（login/register/initialize/logout）不要求 double-submit token，因为首次调用时浏览器还没有 token；但这些端点会校验 browser `Origin`，拒绝 hostile Origin，避免 login CSRF / session fixation。
-
-## 用户隔离
-
-### Thread metadata
-
-Thread metadata 存在 `threads_meta`，关键隔离字段是 `user_id`。
-
-创建 thread 时：
-
-- 客户端传入的 `metadata.user_id` 和 `metadata.owner_id` 会被剥离。
-- `ThreadMetaRepository.create(..., user_id=AUTO)` 从 `ContextVar` 解析真实用户。
-- `/api/threads/search` 默认只返回当前用户的 thread。
-
-读取 / 修改 / 删除时：
-
-- `get()` 默认按当前用户过滤。
-- `check_access()` 用于路由 owner check。
-- 对其他用户的 thread 返回 404，避免泄露资源存在性。
-
-### 文件系统
-
-当前线程文件布局：
-
-```text
-{base_dir}/users/{user_id}/threads/{thread_id}/user-data/
-├── workspace/
-├── uploads/
-└── outputs/
-```
-
-agent 在 sandbox 内看到统一虚拟路径：
-
-```text
-/mnt/user-data/workspace
-/mnt/user-data/uploads
-/mnt/user-data/outputs
-```
-
-`ThreadDataMiddleware` 使用 `get_effective_user_id()` 解析当前用户并生成线程路径。没有认证上下文时会落到 `default` 用户桶，主要用于内部调用、嵌入式 client 或无 HTTP 的本地执行路径。
-
-### Memory
-
-默认 memory 存储：
-
-```text
-{base_dir}/users/{user_id}/memory.json
-{base_dir}/users/{user_id}/agents/{agent_name}/memory.json
-```
-
-有用户上下文时，空或相对 `memory.storage_path` 都使用上述 per-user 默认路径；只有绝对 `memory.storage_path` 会视为显式 opt-out（退出） per-user isolation，所有用户共享该路径。无用户上下文的 legacy 路径仍会把相对 `storage_path` 解析到 `Paths.base_dir` 下。
-
-### 自定义 agent
-
-用户自定义 agent 写入：
-
-```text
-{base_dir}/users/{user_id}/agents/{agent_name}/
-├── config.yaml
-├── SOUL.md
-└── memory.json
-```
-
-旧布局 `{base_dir}/agents/{agent_name}/` 只作为只读兼容回退。更新或删除旧共享 agent 会要求先运行迁移脚本。
-
-## 内部调用与 IM 渠道
-
-IM channel worker 不是浏览器用户，不持有浏览器 cookie。它们通过 Gateway 内部认证：
-
-- 请求带 `X-DeerFlow-Internal-Token`。
-- 同时带匹配的 CSRF cookie/header。
-- 服务端识别为内部用户，`id="default"`、`system_role="internal"`。
-
-这意味着 channel 产生的数据默认进入 `default` 用户桶。这个选择适合“平台级 bot 身份”，但不是“每个 IM 用户单独隔离”。如果后续要做到外部 IM 用户隔离，需要把外部 platform user 映射到 DeerFlow user，并让 channel manager 设置对应的 scoped identity。
-
-## LangGraph-compatible 认证
-
-Gateway 内嵌 runtime 路径由 `AuthMiddleware` 和 `CSRFMiddleware` 保护。
-
-仓库仍保留 `app.gateway.langgraph_auth`，用于 LangGraph Server 直连模式：
-
-- `@auth.authenticate` 校验 JWT cookie、CSRF、用户存在性和 `token_version`。
-- `@auth.on` 在写入 metadata 时注入 `user_id`，并在读路径返回 `{"user_id": current_user}` 过滤条件。
-
-这保证 Gateway 路由和 LangGraph-compatible 直连模式使用同一 JWT 语义。
-
-## 升级与迁移
-
-从无认证版本升级时，可能存在没有 `user_id` 的历史 thread。
-
-当前策略：
-
-1. 首次启动如果没有 admin，只提示访问 `/setup`，不迁移。
-2. 操作者创建 admin。
-3. 后续启动时，`_ensure_admin_user()` 找到 admin，并把 LangGraph store 中缺少 `metadata.user_id` 的 thread 迁移到 admin。
-
-文件系统旧布局迁移由脚本处理：
-
-```bash
-cd backend
-PYTHONPATH=. python scripts/migrate_user_isolation.py --dry-run
-PYTHONPATH=. python scripts/migrate_user_isolation.py --user-id <target-user-id>
-```
-
-迁移脚本覆盖 legacy `memory.json`、`threads/` 和 `agents/` 到 per-user layout。
-
-## 安全不变量
-
-必须长期保持的不变量：
-
-- JWT 只在 HttpOnly cookie 中传输，不出现在响应 JSON。
-- 任何非 public HTTP 路由都不能只靠“cookie 存在”放行，必须严格验证 JWT。
-- `token_version` 不匹配必须拒绝，保证改密码 / reset 后旧 session 失效。
-- 客户端 metadata 中的 `user_id` / `owner_id` 必须剥离。
-- repository 默认 `AUTO` 必须从当前用户上下文解析，不能静默退化成全局查询。
-- 只有迁移脚本和 admin CLI 可以显式传 `user_id=None` 绕过隔离。
-- 本地文件路径必须通过 `Paths` 和 sandbox path validation 解析，不能拼接未校验的用户输入。
-- 捕获认证、迁移、后台任务异常必须记录日志；不能空 catch。
-
-## 已知边界
-
-| 边界 | 当前行为 | 后续方向 |
-|---|---|---|
-| 无 admin 时注册普通用户 | 允许注册普通 `user` | 如产品要求先初始化 admin，给 `/register` 加 gate |
-| 登录限速 | 进程内 dict，单 worker 精确，多 worker 近似 | Redis / DB-backed rate limiter |
-| OAuth | 端点占位，未实现 | 接入 provider 并统一 `token_version` / role 语义 |
-| IM 用户隔离 | channel 使用 `default` 内部用户 | 建立外部用户到 DeerFlow user 的映射 |
-| 绝对 memory path | 显式共享 memory | UI / docs 明确提示 opt-out 风险 |
-
-## 相关文件
-
-| 文件 | 职责 |
-|---|---|
-| `app/gateway/auth_middleware.py` | 全局认证门、JWT 严格验证、写入 user context |
-| `app/gateway/csrf_middleware.py` | CSRF double-submit 和 auth Origin 校验 |
-| `app/gateway/routers/auth.py` | initialize/login/register/logout/me/change-password |
-| `app/gateway/auth/jwt.py` | JWT 创建与解析 |
-| `app/gateway/auth/reset_admin.py` | 密码 reset CLI |
-| `app/gateway/auth/credential_file.py` | 0600 凭据文件写入 |
-| `app/gateway/authz.py` | 路由权限与 owner check |
-| `deerflow/runtime/user_context.py` | 当前用户 ContextVar 与 `AUTO` sentinel |
-| `deerflow/persistence/thread_meta/` | thread metadata owner filter |
-| `deerflow/config/paths.py` | per-user filesystem layout |
-| `deerflow/agents/middlewares/thread_data_middleware.py` | run 时解析用户线程目录 |
-| `deerflow/agents/memory/storage.py` | per-user memory storage |
-| `deerflow/config/agents_config.py` | per-user custom agents |
-| `app/channels/manager.py` | IM channel 内部认证调用 |
-| `scripts/migrate_user_isolation.py` | legacy 数据迁移到 per-user layout |
-| `.deer-flow/data/deerflow.db` | 统一 SQLite 数据库，包含 users / threads_meta / runs / feedback 等表 |
-| `.deer-flow/users/{user_id}/agents/{agent_name}/` | 用户自定义 agent 配置、SOUL 和 agent memory |
-| `.deer-flow/admin_initial_credentials.txt` | `reset_admin` 生成的新凭据文件（0600，读完应删除） |

backend/docs/AUTH_TEST_DOCKER_GAP.md

@@ -24,12 +24,12 @@ All other test plan sections were executed against either:
 
 | Case | Title | What it covers | Why not run |
 |---|---|---|---|
-| TC-DOCKER-01 | `deerflow.db` volume persistence | Verify the `DEER_FLOW_HOME` bind mount survives container restart | needs `docker compose up` |
+| TC-DOCKER-01 | `users.db` volume persistence | Verify the `DEER_FLOW_HOME` bind mount survives container restart | needs `docker compose up` |
 | TC-DOCKER-02 | Session persistence across container restart | `AUTH_JWT_SECRET` env var keeps cookies valid after `docker compose down && up` | needs `docker compose down/up` |
 | TC-DOCKER-03 | Per-worker rate limiter divergence | Confirms in-process `_login_attempts` dict doesn't share state across `gunicorn` workers (4 by default in the compose file); known limitation, documented | needs multi-worker container |
-| TC-DOCKER-04 | IM channels use internal Gateway auth | Verify Feishu/Slack/Telegram dispatchers attach the process-local internal auth header plus CSRF cookie/header when calling Gateway-compatible LangGraph APIs | needs `docker logs` |
+| TC-DOCKER-04 | IM channels skip AuthMiddleware | Verify Feishu/Slack/Telegram dispatchers run in-container against `http://langgraph:2024` without going through nginx | needs `docker logs` |
-| TC-DOCKER-05 | Reset credentials surfacing | `reset_admin` writes a 0600 credential file in `DEER_FLOW_HOME` instead of logging plaintext. The file-based behavior is validated by non-Docker reset tests, so the only Docker-specific gap is verifying the volume mount carries the file out to the host | needs container + host volume |
+| TC-DOCKER-05 | Admin credentials surfacing | **Updated post-simplify** — was "log scrape", now "0600 credential file in `DEER_FLOW_HOME`". The file-based behavior is already validated by TC-1.1 + TC-UPG-13 on sg_dev (non-Docker), so the only Docker-specific gap is verifying the volume mount carries the file out to the host | needs container + host volume |
-| TC-DOCKER-06 | Docker deploy uses Gateway embedded runtime | `./scripts/deploy.sh` produces a Gateway + frontend + nginx topology (no `langgraph` container); same auth flow as local `make dev` | needs `docker compose up` |
+| TC-DOCKER-06 | Gateway-mode Docker deploy | `./scripts/deploy.sh --gateway` produces a 3-container topology (no `langgraph` container); same auth flow as standard mode | needs `docker compose --profile gateway` |
 
 ## Coverage already provided by non-Docker tests
 
@@ -41,9 +41,9 @@ the test cases that ran on sg_dev or local:
 | TC-DOCKER-01 (volume persistence) | TC-REENT-01 on sg_dev (admin row survives gateway restart) — same SQLite file, just no container layer between |
 | TC-DOCKER-02 (session persistence) | TC-API-02/03/06 (cookie roundtrip), plus TC-REENT-04 (multi-cookie) — JWT verification is process-state-free, container restart is equivalent to `pkill uvicorn && uv run uvicorn` |
 | TC-DOCKER-03 (per-worker rate limit) | TC-GW-04 + TC-REENT-09 (single-worker rate limit + 5min expiry). The cross-worker divergence is an architectural property of the in-memory dict; no auth code path differs |
-| TC-DOCKER-04 (IM channels use internal auth) | Code-level: `app/channels/manager.py` creates the `langgraph_sdk` client with `create_internal_auth_headers()` plus CSRF cookie/header, so channel workers do not rely on browser cookies |
+| TC-DOCKER-04 (IM channels skip auth) | Code-level only: `app/channels/manager.py` uses `langgraph_sdk` directly with no cookie handling. The langgraph_auth handler is bypassed by going through SDK, not HTTP |
-| TC-DOCKER-05 (credential surfacing) | `reset_admin` writes `.deer-flow/admin_initial_credentials.txt` with mode 0600 and logs only the path — the only Docker-unique step is whether the bind mount projects this path onto the host, which is a `docker compose` config check, not a runtime behavior change |
+| TC-DOCKER-05 (credential surfacing) | TC-1.1 on sg_dev (file at `~/deer-flow/backend/.deer-flow/admin_initial_credentials.txt`, mode 0600, password 22 chars) — the only Docker-unique step is whether the bind mount projects this path onto the host, which is a `docker compose` config check, not a runtime behavior change |
-| TC-DOCKER-06 (Gateway embedded runtime container) | Section 七 7.2 covered by TC-GW-01..05 + Section 二 (Gateway auth flow on sg_dev) — same Gateway code, container is just a packaging change |
+| TC-DOCKER-06 (gateway-mode container) | Section 七 7.2 covered by TC-GW-01..05 + Section 二 (gateway-mode auth flow on sg_dev) — same Gateway code, container is just a packaging change |
 
 ## Reproduction steps when Docker becomes available
 
@@ -72,6 +72,6 @@ Then run TC-DOCKER-01..06 from the test plan as written.
   about *container packaging* details (bind mounts, multi-worker, log
   collection), not about whether the auth code paths work.
 - **TC-DOCKER-05 was updated in place** in `AUTH_TEST_PLAN.md` to reflect
-  the current reset flow (`reset_admin` → 0600 credentials file, no log leak).
+  the post-simplify reality (credentials file → 0600 file, no log leak).
   The old "grep 'Password:' in docker logs" expectation would have failed
   silently and given a false sense of coverage.

backend/docs/AUTH_TEST_PLAN.md

@@ -4,12 +4,10 @@
 
 | 模式 | 启动命令 | Auth 层 | 端口 |
 |------|---------|---------|------|
-| 标准模式 | `make dev` | Gateway AuthMiddleware（全量） | 2026 (nginx) |
+| 标准模式 | `make dev` | Gateway AuthMiddleware + LangGraph auth | 2026 (nginx) |
+| Gateway 模式 | `make dev-pro` | Gateway AuthMiddleware（全量） | 2026 (nginx) |
 | 直连 Gateway | `cd backend && make gateway` | Gateway AuthMiddleware | 8001 |
-| 直连 LangGraph 兼容性 | 手动运行 LangGraph 工具链时使用 | LangGraph auth | 2024 |
+| 直连 LangGraph | `cd backend && make dev` | LangGraph auth | 2024 |
-
-`make dev`、Docker dev 和生产部署默认都运行 Gateway embedded runtime。
-`app.gateway.langgraph_auth` 仅用于保留的直连 LangGraph 工具链 / Studio 兼容性测试，不是标准服务启动路径。
 
 每种模式下都需执行以下测试。
 
@@ -21,18 +19,19 @@
 
 ```bash
 # 清除已有数据
-rm -f backend/.deer-flow/data/deerflow.db
+rm -f backend/.deer-flow/users.db
 
-# 启动标准模式（Gateway embedded runtime）
+# 选择模式启动
-make dev
+make dev          # 标准模式
+# 或
+make dev-pro      # Gateway 模式
 ```
 
 **验证点：**
-- [ ] 控制台不输出 admin 邮箱或明文密码
+- [ ] 控制台输出 admin 邮箱和随机密码
-- [ ] 控制台提示 `First boot detected — no admin account exists.`
+- [ ] 密码格式为 `secrets.token_urlsafe(16)` 的 22 字符字符串
-- [ ] 控制台提示访问 `/setup` 完成 admin 创建
+- [ ] 邮箱为 `admin@deerflow.dev`
-- [ ] `GET /api/v1/auth/setup-status` 返回 `{"needs_setup": true}`
+- [ ] 提示 `Change it after login: Settings -> Account`
-- [ ] 前端访问 `/login` 会跳转 `/setup`
 
 ### 1.2 非首次启动
 
@@ -43,8 +42,7 @@ make dev
 
 **验证点：**
 - [ ] 控制台不输出密码
-- [ ] `GET /api/v1/auth/setup-status` 返回 `{"needs_setup": false}`
+- [ ] 如果 admin 仍 `needs_setup=True`，控制台有 warning 提示
-- [ ] 已登录用户如果 `needs_setup=True`，访问 workspace 会被引导到 `/setup` 完成改邮箱 / 改密码流程
 
 ### 1.3 环境变量配置
 
@@ -57,7 +55,7 @@ make dev
 
 ## 二、接口流程测试
 
-> 以下用 `BASE=http://localhost:2026` 为例。标准模式经 nginx 暴露此地址。
+> 以下用 `BASE=http://localhost:2026` 为例。标准模式和 Gateway 模式都用此地址。
 > 直连测试替换为对应端口。
 >
 > **CSRF token 提取**：多处用到从 cookie jar 提取 CSRF token，统一使用：
@@ -78,22 +76,19 @@ make dev
 curl -s $BASE/api/v1/auth/setup-status | jq .
 ```
 
-**预期：**
+**预期：** 返回 `{"needs_setup": false}`（admin 在启动时已自动创建，`count_users() > 0`）。仅在启动完成前的极短窗口内可能返回 `true`。
-- 干净数据库且尚未初始化 admin：返回 `{"needs_setup": true}`
-- 已存在 admin：返回 `{"needs_setup": false}`
 
-#### TC-API-02: 首次初始化 Admin
+#### TC-API-02: Admin 首次登录
 
 ```bash
-curl -s -X POST $BASE/api/v1/auth/initialize \
+curl -s -X POST $BASE/api/v1/auth/login/local \
-  -H "Content-Type: application/json" \
+  -d "username=admin@deerflow.dev&password=<控制台密码>" \
-  -d '{"email":"admin@example.com","password":"AdminPass1!"}' \
   -c cookies.txt | jq .
 ```
 
 **预期：**
-- 状态码 201
+- 状态码 200
-- Body: `{"id": "...", "email": "admin@example.com", "system_role": "admin", "needs_setup": false}`
+- Body: `{"expires_in": 604800, "needs_setup": true}`
 - `cookies.txt` 包含 `access_token`（HttpOnly）和 `csrf_token`（非 HttpOnly）
 
 #### TC-API-03: 获取当前用户
@@ -102,9 +97,9 @@ curl -s -X POST $BASE/api/v1/auth/initialize \
 curl -s $BASE/api/v1/auth/me -b cookies.txt | jq .
 ```
 
-**预期：** `{"id": "...", "email": "admin@example.com", "system_role": "admin", "needs_setup": false}`
+**预期：** `{"id": "...", "email": "admin@deerflow.dev", "system_role": "admin", "needs_setup": true}`
 
-#### TC-API-04: 改密码流程
+#### TC-API-04: Setup 流程（改邮箱 + 改密码）
 
 ```bash
 CSRF=$(grep csrf_token cookies.txt | awk '{print $NF}')
@@ -112,36 +107,13 @@ curl -s -X POST $BASE/api/v1/auth/change-password \
   -b cookies.txt \
   -H "Content-Type: application/json" \
   -H "X-CSRF-Token: $CSRF" \
-  -d '{"current_password":"AdminPass1!","new_password":"NewPass123!"}' | jq .
+  -d '{"current_password":"<控制台密码>","new_password":"NewPass123!","new_email":"admin@example.com"}' | jq .
 ```
 
 **预期：**
 - 状态码 200
 - `{"message": "Password changed successfully"}`
-- 再调 `/auth/me` 仍为 `admin@example.com`，`needs_setup` 仍为 `false`
+- 再调 `/auth/me` 邮箱变为 `admin@example.com`，`needs_setup` 变为 `false`
-
-#### TC-API-04a: reset_admin 后的 Setup 流程（改邮箱 + 改密码）
-
-```bash
-cd backend
-python -m app.gateway.auth.reset_admin --email admin@example.com
-# 从 .deer-flow/admin_initial_credentials.txt 读取 reset 后密码
-
-curl -s -X POST $BASE/api/v1/auth/login/local \
-  -d "username=admin@example.com&password=<凭据文件密码>" \
-  -c cookies.txt | jq .
-
-CSRF=$(grep csrf_token cookies.txt | awk '{print $NF}')
-curl -s -X POST $BASE/api/v1/auth/change-password \
-  -b cookies.txt \
-  -H "Content-Type: application/json" \
-  -H "X-CSRF-Token: $CSRF" \
-  -d '{"current_password":"<凭据文件密码>","new_password":"AdminPass2!","new_email":"admin2@example.com"}' | jq .
-```
-
-**预期：**
-- 登录返回 `{"expires_in": 604800, "needs_setup": true}`
-- `change-password` 后 `/auth/me` 邮箱变为 `admin2@example.com`，`needs_setup` 变为 `false`
 
 #### TC-API-05: 普通用户注册
 
@@ -211,18 +183,20 @@ curl -s -X POST $BASE/api/threads/search \
 
 **预期：** 返回 0 或仅包含 user2 自己的 thread
 
-### 2.3 LangGraph-compatible Gateway 路由隔离
+### 2.3 标准模式 LangGraph Server 隔离
 
-#### TC-API-10: LangGraph-compatible 端点需要 cookie
+> 仅在标准模式下测试。Gateway 模式不跑 LangGraph Server。
+
+#### TC-API-10: LangGraph 端点需要 cookie
 
 ```bash
-# 不带 cookie 访问 LangGraph-compatible 接口
+# 不带 cookie 访问 LangGraph 接口
 curl -s -w "%{http_code}" $BASE/api/langgraph/threads
 ```
 
 **预期：** 401
 
-#### TC-API-11: LangGraph-compatible 路由带 cookie 可访问
+#### TC-API-11: LangGraph 带 cookie 可访问
 
 ```bash
 curl -s $BASE/api/langgraph/threads -b user1.txt | jq length
@@ -230,10 +204,10 @@ curl -s $BASE/api/langgraph/threads -b user1.txt | jq length
 
 **预期：** 200，返回 user1 的 thread 列表
 
-#### TC-API-12: LangGraph-compatible 路由隔离 — 用户只看到自己的
+#### TC-API-12: LangGraph 隔离 — 用户只看到自己的
 
 ```bash
-# user2 查 threads
+# user2 查 LangGraph threads
 curl -s $BASE/api/langgraph/threads -b user2.txt | jq length
 ```
 
@@ -519,7 +493,7 @@ curl -s -X POST $BASE/api/v1/auth/register \
 
 ```bash
 # 检查数据库
-sqlite3 backend/.deer-flow/data/deerflow.db "SELECT email, password_hash FROM users LIMIT 3;"
+sqlite3 backend/.deer-flow/users.db "SELECT email, password_hash FROM users LIMIT 3;"
 ```
 
 **预期：** `password_hash` 以 `$2b$` 开头（bcrypt 格式）
@@ -532,25 +506,24 @@ sqlite3 backend/.deer-flow/data/deerflow.db "SELECT email, password_hash FROM us
 
 ### 4.1 首次登录流程
 
-#### TC-UI-01: 无 admin 时访问 workspace 跳转 setup
+#### TC-UI-01: 访问首页跳转登录
 
 1. 打开 `http://localhost:2026/workspace`
-2. **预期：** 自动跳转到 `/setup`
+2. **预期：** 自动跳转到 `/login`
 
-#### TC-UI-02: Setup 页面创建 admin
+#### TC-UI-02: Login 页面
 
-1. 输入 admin 邮箱、密码、确认密码
+1. 输入 admin 邮箱和控制台密码
-2. 点击 Create Admin Account
+2. 点击 Login
+3. **预期：** 跳转到 `/setup`（因为 `needs_setup=true`）
+
+#### TC-UI-03: Setup 页面
+
+1. 输入新邮箱、控制台密码（current）、新密码、确认密码
+2. 点击 Complete Setup
 3. **预期：** 跳转到 `/workspace`
 4. 刷新页面不跳回 `/setup`
 
-#### TC-UI-03: 已初始化后 Login 页面
-
-1. 退出登录后访问 `/login`
-2. 输入 admin 邮箱和密码
-3. 点击 Login
-4. **预期：** 跳转到 `/workspace`
-
 #### TC-UI-04: Setup 密码不匹配
 
 1. 新密码和确认密码不一致
@@ -629,7 +602,7 @@ sqlite3 backend/.deer-flow/data/deerflow.db "SELECT email, password_hash FROM us
 #### TC-UI-15: reset_admin 后重新登录
 
 1. 执行 `cd backend && python -m app.gateway.auth.reset_admin`
-2. 从 `.deer-flow/admin_initial_credentials.txt` 读取新密码并登录
+2. 使用新密码登录
 3. **预期：** 跳转到 `/setup` 页面（`needs_setup` 被重置为 true）
 4. 旧 session 已失效
 
@@ -672,28 +645,18 @@ make install
 make dev
 ```
 
-#### TC-UPG-01: 首次启动等待 admin 初始化
+#### TC-UPG-01: 首次启动创建 admin
 
 **预期：**
-- [ ] 控制台不输出 admin 邮箱或随机密码
+- [ ] 控制台输出 admin 邮箱（`admin@deerflow.dev`）和随机密码
-- [ ] 访问 `/setup` 可创建第一个 admin
 - [ ] 无报错，正常启动
 
 #### TC-UPG-02: 旧 Thread 迁移到 admin
 
 ```bash
-# 创建第一个 admin
-curl -s -X POST http://localhost:2026/api/v1/auth/initialize \
-  -H "Content-Type: application/json" \
-  -d '{"email":"admin@example.com","password":"AdminPass1!"}' \
-  -c cookies.txt
-
-# 重启一次：启动迁移只在已有 admin 的启动路径执行
-make stop && make dev
-
 # 登录 admin
 curl -s -X POST http://localhost:2026/api/v1/auth/login/local \
-  -d "username=admin@example.com&password=AdminPass1!" \
+  -d "username=admin@deerflow.dev&password=<控制台密码>" \
   -c cookies.txt
 
 # 查看 thread 列表
@@ -707,8 +670,8 @@ curl -s -X POST http://localhost:2026/api/threads/search \
 
 **预期：**
 - [ ] 返回的 thread 数量 ≥ 旧版创建的数量
-- [ ] 控制台日志有 `Migrated N orphan LangGraph thread(s) to admin`
+- [ ] 控制台日志有 `Migrated N orphaned thread(s) to admin`
-- [ ] 旧 thread 只对 admin 可见
+- [ ] 每个 thread 的 `metadata.owner_id` 都已被设为 admin 的 ID
 
 #### TC-UPG-03: 旧 Thread 内容完整
 
@@ -720,7 +683,7 @@ curl -s http://localhost:2026/api/threads/<old-thread-id> \
 
 **预期：**
 - [ ] `metadata.title` 保留原值（如 `old-thread-1`）
-- [ ] 响应不回显服务端保留的 `user_id` / `owner_id`
+- [ ] `metadata.owner_id` 已填充
 
 #### TC-UPG-04: 新用户看不到旧 Thread
 
@@ -743,19 +706,18 @@ curl -s -X POST http://localhost:2026/api/threads/search \
 
 ### 5.3 数据库 Schema 兼容
 
-#### TC-UPG-05: 无 deerflow.db 时创建 schema 但不创建默认用户
+#### TC-UPG-05: 无 users.db 时自动创建
 
 ```bash
-ls -la backend/.deer-flow/data/deerflow.db
+ls -la backend/.deer-flow/users.db
-sqlite3 backend/.deer-flow/data/deerflow.db "SELECT COUNT(*) FROM users;"
 ```
 
-**预期：** 文件存在，`sqlite3` 可查到 `users` 表含 `needs_setup`、`token_version` 列；未调用 `/initialize` 前用户数为 0
+**预期：** 文件存在，`sqlite3` 可查到 `users` 表含 `needs_setup`、`token_version` 列
 
-#### TC-UPG-06: deerflow.db WAL 模式
+#### TC-UPG-06: users.db WAL 模式
 
 ```bash
-sqlite3 backend/.deer-flow/data/deerflow.db "PRAGMA journal_mode;"
+sqlite3 backend/.deer-flow/users.db "PRAGMA journal_mode;"
 ```
 
 **预期：** 返回 `wal`
@@ -806,9 +768,9 @@ make dev
 ```
 
 **预期：**
-- [ ] 服务正常启动（忽略 `deerflow.db`，无 auth 相关代码不报错）
+- [ ] 服务正常启动（忽略 `users.db`，无 auth 相关代码不报错）
 - [ ] 旧对话数据仍然可访问
-- [ ] `deerflow.db` 文件残留但不影响运行
+- [ ] `users.db` 文件残留但不影响运行
 
 #### TC-UPG-12: 再次升级到 auth 分支
 
@@ -819,47 +781,51 @@ make dev
 ```
 
 **预期：**
-- [ ] 识别已有 `deerflow.db`，不重新创建 admin
+- [ ] 识别已有 `users.db`，不重新创建 admin
-- [ ] 旧的 admin 账号仍可登录（如果回退期间未删 `deerflow.db`）
+- [ ] 旧的 admin 账号仍可登录（如果回退期间未删 `users.db`）
 
-### 5.7 Admin 初始化与 reset_admin
+### 5.7 休眠 Admin（初始密码未使用/未更改）
 
-> 首次启动不生成默认 admin，也不在日志输出密码。忘记密码时走 `reset_admin`，新密码写入 0600 凭据文件。
+> 首次启动生成 admin + 随机密码，但运维未登录、未改密码。
+> 密码只在首次启动的控制台闪过一次，后续启动不再显示。
 
-#### TC-UPG-13: 未初始化 admin 时重启不创建默认账号
+#### TC-UPG-13: 重启后自动重置密码并打印
 
 ```bash
-rm -f backend/.deer-flow/data/deerflow.db
+# 首次启动，记录密码
+rm -f backend/.deer-flow/users.db
 make dev
+# 控制台输出密码 P0，不登录
 make stop
 
+# 隔了几天，再次启动
 make dev
-curl -s $BASE/api/v1/auth/setup-status | jq .
+# 控制台输出新密码 P1
 ```
 
 **预期：**
-- [ ] 控制台不输出密码
+- [ ] 控制台输出 `Admin account setup incomplete — password reset`
-- [ ] `setup-status` 仍为 `{"needs_setup": true}`
+- [ ] 输出新密码 P1（P0 已失效）
-- [ ] 访问 `/setup` 仍可创建第一个 admin
+- [ ] 用 P1 可以登录，P0 不可以
+- [ ] 登录后 `needs_setup=true`，跳转 `/setup`
+- [ ] `token_version` 递增（旧 session 如有也失效）
 
-#### TC-UPG-14: 密码丢失 — reset_admin 写入凭据文件
+#### TC-UPG-14: 密码丢失 — 无需 CLI，重启即可
 
 ```bash
-python -m app.gateway.auth.reset_admin --email admin@example.com
+# 忘记了控制台密码 → 直接重启服务
-ls -la backend/.deer-flow/admin_initial_credentials.txt
+make stop && make dev
-cat backend/.deer-flow/admin_initial_credentials.txt
+# 控制台自动输出新密码
 ```
 
 **预期：**
-- [ ] 命令行只输出凭据文件路径，不输出明文密码
+- [ ] 无需 `reset_admin`，重启服务即可拿到新密码
-- [ ] 凭据文件权限为 `0600`
+- [ ] `reset_admin` CLI 仍然可用作手动备选方案
-- [ ] 凭据文件包含 email + password 行
-- [ ] 该用户下次登录返回 `needs_setup=true`
 
-#### TC-UPG-15: 未初始化 admin 期间普通用户注册策略边界
+#### TC-UPG-15: 休眠 admin 期间普通用户注册
 
 ```bash
-# admin 尚不存在，普通用户尝试注册
+# admin 存在但从未登录，普通用户先注册
 curl -s -X POST $BASE/api/v1/auth/register \
   -H "Content-Type: application/json" \
   -d '{"email":"earlybird@example.com","password":"EarlyPass1!"}' \
@@ -867,11 +833,11 @@ curl -s -X POST $BASE/api/v1/auth/register \
 ```
 
 **预期：**
-- [ ] 当前代码允许注册普通用户并自动登录（201，角色为 `user`）
+- [ ] 注册成功（201），角色为 `user`
-- [ ] 但 `setup-status` 仍为 `{"needs_setup": true}`，因为 admin 仍不存在
+- [ ] 无法提权为 admin
-- [ ] 这是一个产品策略边界：若要求“必须先有 admin”，需要在 `/register` 增加 admin-exists gate
+- [ ] 普通用户的数据与 admin 隔离
 
-#### TC-UPG-16: 普通用户数据与后续 admin 隔离
+#### TC-UPG-16: 休眠 admin 不影响后续操作
 
 ```bash
 # 普通用户正常创建 thread、发消息
@@ -883,13 +849,14 @@ curl -s -X POST $BASE/api/threads \
   -d '{"metadata":{}}' | jq .thread_id
 ```
 
-**预期：** 普通用户正常创建 thread；后续 admin 创建后，搜索不到该普通用户 thread
+**预期：** 正常创建，不受休眠 admin 影响
 
-#### TC-UPG-17: reset_admin 后完成 Setup
+#### TC-UPG-17: 休眠 admin 最终完成 Setup
 
 ```bash
+# 运维终于登录
 curl -s -X POST $BASE/api/v1/auth/login/local \
-  -d "username=admin@example.com&password=<凭据文件密码>" \
+  -d "username=admin@deerflow.dev&password=<P0或P1>" \
   -c admin.txt | jq .needs_setup
 # 预期: true
 
@@ -899,7 +866,7 @@ curl -s -X POST $BASE/api/v1/auth/change-password \
   -b admin.txt \
   -H "Content-Type: application/json" \
   -H "X-CSRF-Token: $CSRF" \
-  -d '{"current_password":"<凭据文件密码>","new_password":"AdminFinal1!","new_email":"admin@real.com"}' \
+  -d '{"current_password":"<密码>","new_password":"AdminFinal1!","new_email":"admin@real.com"}' \
   -c admin.txt
 
 # 验证
@@ -909,7 +876,7 @@ curl -s $BASE/api/v1/auth/me -b admin.txt | jq '{email, needs_setup}'
 **预期：**
 - [ ] `email` 变为 `admin@real.com`
 - [ ] `needs_setup` 变为 `false`
-- [ ] 后续登录使用新密码
+- [ ] 后续重启控制台不再有 warning
 
 #### TC-UPG-18: 长期未用后 JWT 密钥轮换
 
@@ -923,8 +890,8 @@ make stop && make dev
 
 **预期：**
 - [ ] 服务正常启动
-- [ ] 账号密码仍可登录（密码存在 DB，与 JWT 密钥无关）
+- [ ] 旧密码仍可登录（密码存在 DB，与 JWT 密钥无关）
-- [ ] 旧的 JWT token 失效（密钥变了签名不匹配）
+- [ ] 旧的 JWT token 失效（密钥变了签名不匹配）— 但因为从未登录过也没有旧 token
 
 ---
 
@@ -943,7 +910,7 @@ for i in 1 2 3; do
 done
 
 # 检查 admin 数量
-sqlite3 backend/.deer-flow/data/deerflow.db \
+sqlite3 backend/.deer-flow/users.db \
   "SELECT COUNT(*) FROM users WHERE system_role='admin';"
 ```
 
@@ -1088,7 +1055,7 @@ curl -s -X POST $BASE/api/v1/auth/register \
 wait
 
 # 检查用户数
-sqlite3 backend/.deer-flow/data/deerflow.db \
+sqlite3 backend/.deer-flow/users.db \
   "SELECT COUNT(*) FROM users WHERE email='race@example.com';"
 ```
 
@@ -1198,16 +1165,13 @@ curl -s -w "%{http_code}" -X DELETE "$BASE/api/threads/$TID" \
 ```bash
 cd backend
 python -m app.gateway.auth.reset_admin
-cp .deer-flow/admin_initial_credentials.txt /tmp/deerflow-reset-p1.txt
+# 记录密码 P1
-P1=$(awk -F': ' '/^password:/ {print $2}' /tmp/deerflow-reset-p1.txt)
 
 python -m app.gateway.auth.reset_admin
-cp .deer-flow/admin_initial_credentials.txt /tmp/deerflow-reset-p2.txt
+# 记录密码 P2
-P2=$(awk -F': ' '/^password:/ {print $2}' /tmp/deerflow-reset-p2.txt)
 ```
 
 **预期：**
-- [ ] `.deer-flow/admin_initial_credentials.txt` 每次都会被重写，文件权限为 `0600`
 - [ ] P1 ≠ P2（每次生成新随机密码）
 - [ ] P1 不可用，只有 P2 有效
 - [ ] `token_version` 递增了 2
@@ -1232,11 +1196,21 @@ P2=$(awk -F': ' '/^password:/ {print $2}' /tmp/deerflow-reset-p2.txt)
 ## 七、模式差异测试
 
 > 以下用 `GW=http://localhost:8001` 表示直连 Gateway，`BASE=http://localhost:2026` 表示经 nginx。
-> 标准启动命令：`make dev`（或 `./scripts/serve.sh --dev`）。
+> Gateway 模式启动命令：`make dev-pro`（或 `./scripts/serve.sh --dev --gateway`）。
 
-### 7.1 标准启动模式
+### 7.1 标准模式独有
 
-#### TC-MODE-01: Gateway AuthMiddleware 的 token_version 检查
+> 启动命令：`make dev`（或 `./scripts/serve.sh --dev`）
+
+#### TC-MODE-01: LangGraph Server 独立运行，需 cookie
+
+```bash
+# 无 cookie 访问 LangGraph
+curl -s -w "%{http_code}" -o /dev/null $BASE/api/langgraph/threads/search
+# 预期: 403（LangGraph auth handler 拒绝）
+```
+
+#### TC-MODE-02: LangGraph auth 的 token_version 检查
 
 ```bash
 # 登录拿 cookie
@@ -1249,9 +1223,9 @@ curl -s -X POST $BASE/api/v1/auth/change-password \
   -b cookies.txt -H "Content-Type: application/json" -H "X-CSRF-Token: $CSRF" \
   -d '{"current_password":"正确密码","new_password":"NewPass1!"}' -c new_cookies.txt
 
-# 用旧 cookie 访问 LangGraph-compatible 路由
+# 用旧 cookie 访问 LangGraph
 curl -s -w "%{http_code}" $BASE/api/langgraph/threads/search -b cookies.txt
-# 预期: 401（token_version 不匹配）
+# 预期: 403（token_version 不匹配）
 
 # 用新 cookie 访问
 CSRF2=$(grep csrf_token new_cookies.txt | awk '{print $NF}')
@@ -1260,7 +1234,7 @@ curl -s -w "%{http_code}" -X POST $BASE/api/langgraph/threads/search \
 # 预期: 200
 ```
 
-#### TC-MODE-02: Gateway owner filter 隔离
+#### TC-MODE-03: LangGraph auth 的 owner filter 隔离
 
 ```bash
 # user1 创建 thread
@@ -1285,9 +1259,18 @@ print('OK: user2 sees', len(threads), 'threads, none belong to user1')
 "
 ```
 
-#### TC-MODE-03: 所有请求经 AuthMiddleware
+### 7.2 Gateway 模式独有
+
+> 启动命令：`make dev-pro`（或 `./scripts/serve.sh --dev --gateway`）
+> 无 LangGraph Server 进程，agent runtime 嵌入 Gateway。
+
+#### TC-MODE-04: 所有请求经 AuthMiddleware
 
 ```bash
+# 确认 LangGraph Server 未运行
+curl -s -w "%{http_code}" -o /dev/null http://localhost:2024/ok
+# 预期: 000（连接被拒）
+
 # Gateway API 受保护
 curl -s -w "%{http_code}" -o /dev/null $BASE/api/models
 # 预期: 401
@@ -1298,7 +1281,7 @@ curl -s -w "%{http_code}" -o /dev/null -X POST $BASE/api/langgraph/threads/searc
 # 预期: 401
 ```
 
-#### TC-MODE-04: 标准模式下完整 auth 流程
+#### TC-MODE-05: Gateway 模式下完整 auth 流程
 
 ```bash
 # 登录
@@ -1313,7 +1296,7 @@ curl -s -X POST $BASE/api/langgraph/threads \
   -d '{"metadata":{}}' | python3 -c "import sys,json; print(json.load(sys.stdin)['thread_id'])"
 # 预期: 返回 thread_id
 
-# CSRF 保护（CSRFMiddleware 覆盖所有 Gateway 路由）
+# CSRF 保护（Gateway 模式下 CSRFMiddleware 直接覆盖所有路由）
 curl -s -w "%{http_code}" -o /dev/null -X POST $BASE/api/langgraph/threads \
   -b cookies.txt -H "Content-Type: application/json" -d '{"metadata":{}}'
 # 预期: 403（CSRF token missing）
@@ -1341,8 +1324,7 @@ done
 ```bash
 GW=http://localhost:8001
 
-for path in /health /api/v1/auth/setup-status /api/v1/auth/login/local \
+for path in /health /api/v1/auth/setup-status /api/v1/auth/login/local /api/v1/auth/register; do
-            /api/v1/auth/register /api/v1/auth/initialize /api/v1/auth/logout; do
   echo "$path: $(curl -s -w '%{http_code}' -o /dev/null $GW$path)"
 done
 # 预期: 200 或 405/422（方法不对但不是 401）
@@ -1412,14 +1394,14 @@ done
 
 ### 7.4 Docker 部署
 
-> 启动命令：`./scripts/deploy.sh`
+> 启动命令：`./scripts/deploy.sh`（标准）或 `./scripts/deploy.sh --gateway`（Gateway 模式）
 > Docker Compose 文件：`docker/docker-compose.yaml`
 >
 > 前置条件：
 > - `.env` 中设置 `AUTH_JWT_SECRET`（否则每次容器重启 session 全部失效）
-> - `DEER_FLOW_HOME` 挂载到宿主机目录（持久化 `deerflow.db`）
+> - `DEER_FLOW_HOME` 挂载到宿主机目录（持久化 `users.db`）
 
-#### TC-DOCKER-01: deerflow.db 通过 volume 持久化
+#### TC-DOCKER-01: users.db 通过 volume 持久化
 
 ```bash
 # 启动容器
@@ -1434,13 +1416,13 @@ curl -s -X POST $BASE/api/v1/auth/register \
   -H "Content-Type: application/json" \
   -d '{"email":"docker-test@example.com","password":"DockerTest1!"}' -w "\nHTTP %{http_code}"
 
-# 检查宿主机上的 deerflow.db
+# 检查宿主机上的 users.db
-ls -la ${DEER_FLOW_HOME:-backend/.deer-flow}/data/deerflow.db
+ls -la ${DEER_FLOW_HOME:-backend/.deer-flow}/users.db
-sqlite3 ${DEER_FLOW_HOME:-backend/.deer-flow}/data/deerflow.db \
+sqlite3 ${DEER_FLOW_HOME:-backend/.deer-flow}/users.db \
   "SELECT email FROM users WHERE email='docker-test@example.com';"
 ```
 
-**预期：** deerflow.db 在宿主机 `DEER_FLOW_HOME` 目录中，查询可见刚注册的用户。
+**预期：** users.db 在宿主机 `DEER_FLOW_HOME` 目录中，查询可见刚注册的用户。
 
 #### TC-DOCKER-02: 重启容器后 session 保持
 
@@ -1484,24 +1466,22 @@ done
 
 **已知限制：** In-process rate limiter 不跨 worker 共享。生产环境如需精确限速，需要 Redis 等外部存储。
 
-#### TC-DOCKER-04: IM 渠道使用内部认证
+#### TC-DOCKER-04: IM 渠道不经过 auth
 
 ```bash
-# IM 渠道（Feishu/Slack/Telegram）在 gateway 容器内部通过 LangGraph SDK 调 Gateway
+# IM 渠道（Feishu/Slack/Telegram）在 gateway 容器内部通过 LangGraph SDK 通信
-# 请求携带 process-local internal auth header，并带匹配的 CSRF cookie/header
+# 不走 nginx，不经过 AuthMiddleware
 
 # 验证方式：检查 gateway 日志中 channel manager 的请求不包含 auth 错误
 docker logs deer-flow-gateway 2>&1 | grep -E "ChannelManager|channel" | head -10
 ```
 
-**预期：** 无 auth 相关错误。渠道不依赖浏览器 cookie；服务端通过内部认证头把请求归入 `default` 用户桶。
+**预期：** 无 auth 相关错误。渠道通过 `langgraph-sdk` 直连 LangGraph Server（`http://langgraph:2024`），不走 auth 层。
 
-#### TC-DOCKER-05: reset_admin 密码写入 0600 凭证文件（不再走日志）
+#### TC-DOCKER-05: admin 密码写入 0600 凭证文件（不再走日志）
 
 ```bash
-# 首次启动不会自动生成 admin 密码。先重置已有 admin，凭据文件写在挂载到宿主机的 DEER_FLOW_HOME 下。
+# 凭证文件写在挂载到宿主机的 DEER_FLOW_HOME 下
-docker exec deer-flow-gateway python -m app.gateway.auth.reset_admin --email docker-test@example.com
-
 ls -la ${DEER_FLOW_HOME:-backend/.deer-flow}/admin_initial_credentials.txt
 # 预期文件权限: -rw------- (0600)
 
@@ -1521,26 +1501,25 @@ docker logs deer-flow-gateway 2>&1 | grep -iE "Password: .{15,}" && echo "FAIL:
 - 容器日志输出**路径**（不是密码本身），符合 CodeQL `py/clear-text-logging-sensitive-data` 规则
 - `grep "Password:"` 在日志中**应当无匹配**（旧行为已废弃，simplify pass 移除了日志泄露路径）
 
-#### TC-DOCKER-06: Docker 部署
+#### TC-DOCKER-06: Gateway 模式 Docker 部署
 
 ```bash
-# 标准 Docker 模式：runtime 嵌入 gateway 容器
+# Gateway 模式：无 langgraph 容器
-./scripts/deploy.sh
+./scripts/deploy.sh --gateway
 sleep 15
 
-# 确认 gateway 容器存在
+# 确认 langgraph 容器不存在
-docker ps --filter name=deer-flow-gateway --format '{{.Names}}'
+docker ps --filter name=deer-flow-langgraph --format '{{.Names}}' | wc -l
-# 预期: deer-flow-gateway
+# 预期: 0
 
-# auth 流程正常：未登录受保护接口返回 401
+# auth 流程正常
 curl -s -w "%{http_code}" -o /dev/null $BASE/api/models
 # 预期: 401
 
-curl -s -X POST $BASE/api/v1/auth/initialize \
+curl -s -X POST $BASE/api/v1/auth/login/local \
-  -H "Content-Type: application/json" \
+  -d "username=admin@deerflow.dev&password=<日志密码>" \
-  -d '{"email":"admin@example.com","password":"AdminPass1!"}' \
   -c cookies.txt -w "\nHTTP %{http_code}"
-# 预期: 201
+# 预期: 200
 ```
 
 ### 7.4 补充边界用例
@@ -1608,15 +1587,13 @@ curl -s -D - -X POST $BASE/api/v1/auth/login/local \
 #### TC-EDGE-05: HTTP 无 max_age / HTTPS 有 max_age
 
 ```bash
-GW=http://localhost:8001
-
 # HTTP
-curl -s -D - -X POST $GW/api/v1/auth/login/local \
+curl -s -D - -X POST $BASE/api/v1/auth/login/local \
   -d "username=admin@example.com&password=正确密码" 2>/dev/null \
   | grep "access_token=" | grep -oi "max-age=[0-9]*" || echo "NO max-age (HTTP session cookie)"
 
-# HTTPS：直连 Gateway 才能用 X-Forwarded-Proto 模拟 HTTPS；nginx 会覆盖该 header
+# HTTPS
-curl -s -D - -X POST $GW/api/v1/auth/login/local \
+curl -s -D - -X POST $BASE/api/v1/auth/login/local \
   -H "X-Forwarded-Proto: https" \
   -d "username=admin@example.com&password=正确密码" 2>/dev/null \
   | grep "access_token=" | grep -oi "max-age=[0-9]*"
@@ -1735,10 +1712,10 @@ curl -s -X POST $BASE/api/threads \
   -b cookies.txt \
   -H "Content-Type: application/json" \
   -H "X-CSRF-Token: $CSRF" \
-  -d '{"metadata":{"owner_id":"victim-user-id","user_id":"victim-user-id"}}' | jq .metadata
+  -d '{"metadata":{"owner_id":"victim-user-id"}}' | jq .metadata.owner_id
 ```
 
-**预期：** 返回的 `metadata` 不包含 `owner_id` 或 `user_id`。真实所有权写入 `threads_meta.user_id`，不从客户端 metadata 接收，也不通过 metadata 回显。
+**预期：** 返回的 `metadata.owner_id` 应为当前登录用户的 ID，不是请求中注入的 `victim-user-id`。服务端应覆盖客户端提供的 `user_id`。
 
 #### 7.5.6 HTTP Method 探测
 
@@ -1819,6 +1796,6 @@ cd backend && PYTHONPATH=. uv run pytest \
 # 核心接口冒烟
 curl -s $BASE/health                              # 200
 curl -s $BASE/api/models                          # 401 (无 cookie)
-curl -s $BASE/api/v1/auth/setup-status            # 200
+curl -s -X POST $BASE/api/v1/auth/setup-status    # 200
 curl -s $BASE/api/v1/auth/me -b cookies.txt       # 200 (有 cookie)
 ```

backend/docs/AUTH_UPGRADE.md

@@ -2,16 +2,13 @@
 
 DeerFlow 内置了认证模块。本文档面向从无认证版本升级的用户。
 
-完整设计见 [AUTH_DESIGN.md](AUTH_DESIGN.md)。
-
 ## 核心概念
 
 认证模块采用**始终强制**策略：
 
-- 首次启动时不会自动创建账号；首次访问 `/setup` 时由操作者创建第一个 admin 账号
+- 首次启动时自动创建 admin 账号，随机密码打印到控制台日志
 - 认证从一开始就是强制的，无竞争窗口
-- 已有 admin 后，服务启动时会把历史对话（升级前创建且缺少 `user_id` 的 thread）迁移到 admin 名下
+- 历史对话（升级前创建的 thread）自动迁移到 admin 名下
-- 新数据按用户隔离：thread、workspace/uploads/outputs、memory、自定义 agent 都归属当前用户
 
 ## 升级步骤
 
@@ -28,41 +25,39 @@ cd backend && make install
 make dev
 ```
 
-如果没有 admin 账号，控制台只会提示：
+控制台会输出：
 
 ```
 ============================================================
-  First boot detected — no admin account exists.
+  Admin account created on first boot
-  Visit /setup to complete admin account creation.
+  Email:    admin@deerflow.dev
+  Password: aB3xK9mN_pQ7rT2w
+  Change it after login: Settings → Account
 ============================================================
 ```
 
-首次启动不会在日志里打印随机密码，也不会写入默认 admin。这样避免启动日志泄露凭据，也避免在操作者创建账号前出现可被猜测的默认身份。
+如果未登录就重启了服务，不用担心——只要 setup 未完成，每次启动都会重置密码并重新打印到控制台。
 
-### 3. 创建 admin
+### 3. 登录
 
-访问 `http://localhost:2026/setup`，填写邮箱和密码创建第一个 admin 账号。创建成功后会自动登录并进入 workspace。
+访问 `http://localhost:2026/login`，使用控制台输出的邮箱和密码登录。
 
-如果这是从无认证版本升级，创建 admin 后重启一次服务，让启动迁移把缺少 `user_id` 的历史 thread 归属到 admin。
+### 4. 修改密码
 
-### 4. 登录
+登录后进入 Settings → Account → Change Password。
-
-后续访问 `http://localhost:2026/login`，使用已创建的邮箱和密码登录。
 
 ### 5. 添加用户（可选）
 
-其他用户通过 `/login` 页面注册，自动获得 **user** 角色。每个用户只能看到自己的对话、上传文件、输出文件、memory 和自定义 agent。
+其他用户通过 `/login` 页面注册，自动获得 **user** 角色。每个用户只能看到自己的对话。
 
 ## 安全机制
 
 | 机制 | 说明 |
 |------|------|
 | JWT HttpOnly Cookie | Token 不暴露给 JavaScript，防止 XSS 窃取 |
-| CSRF Double Submit Cookie | 受保护的 POST/PUT/PATCH/DELETE 请求需携带 `X-CSRF-Token`；登录/注册/初始化/登出走 auth 端点 Origin 校验 |
+| CSRF Double Submit Cookie | 所有 POST/PUT/DELETE 请求需携带 `X-CSRF-Token` |
 | bcrypt 密码哈希 | 密码不以明文存储 |
-| Thread owner filter | `threads_meta.user_id` 由服务端认证上下文写入，搜索、读取、更新、删除默认按当前用户过滤 |
+| 多租户隔离 | 用户只能访问自己的 thread |
-| 文件系统隔离 | 线程数据写入 `{base_dir}/users/{user_id}/threads/{thread_id}/user-data/`，sandbox 内统一映射为 `/mnt/user-data/` |
-| Memory / agent 隔离 | 用户 memory 和自定义 agent 写入 `{base_dir}/users/{user_id}/...`；旧共享 agent 只作为只读兼容回退 |
 | HTTPS 自适应 | 检测 `x-forwarded-proto`，自动设置 `Secure` cookie 标志 |
 
 ## 常见操作
@@ -79,27 +74,23 @@ python -m app.gateway.auth.reset_admin
 python -m app.gateway.auth.reset_admin --email user@example.com
 ```
 
-会把新的随机密码写入 `.deer-flow/admin_initial_credentials.txt`，文件权限为 `0600`。命令行只输出文件路径，不输出明文密码。
+会输出新的随机密码。
 
 ### 完全重置
 
-删除统一 SQLite 数据库，重启后重新访问 `/setup` 创建新 admin：
+删除用户数据库，重启后自动创建新 admin：
 
 ```bash
-rm -f backend/.deer-flow/data/deerflow.db
+rm -f backend/.deer-flow/users.db
-# 重启服务后访问 http://localhost:2026/setup
+# 重启服务，控制台输出新密码
 ```
 
 ## 数据存储
 
 | 文件 | 内容 |
 |------|------|
-| `.deer-flow/data/deerflow.db` | 统一 SQLite 数据库（users、threads_meta、runs、feedback 等应用数据） |
+| `.deer-flow/users.db` | SQLite 用户数据库（密码哈希、角色） |
-| `.deer-flow/users/{user_id}/threads/{thread_id}/user-data/` | 用户线程的 workspace、uploads、outputs |
+| `.env` 中的 `AUTH_JWT_SECRET` | JWT 签名密钥（未设置时自动生成临时密钥，重启后 session 失效） |
-| `.deer-flow/users/{user_id}/memory.json` | 用户级 memory |
-| `.deer-flow/users/{user_id}/agents/{agent_name}/` | 用户自定义 agent 配置、SOUL 和 agent memory |
-| `.deer-flow/admin_initial_credentials.txt` | `reset_admin` 生成的新凭据文件（0600，读完应删除） |
-| `.env` 中的 `AUTH_JWT_SECRET` | JWT 签名密钥（未设置时自动生成并持久化到 `.deer-flow/.jwt_secret`，重启后 session 保持） |
 
 ### 生产环境建议
 
@@ -120,21 +111,19 @@ python -c "import secrets; print(secrets.token_urlsafe(32))"
 | `/api/v1/auth/me` | GET | 获取当前用户信息 |
 | `/api/v1/auth/change-password` | POST | 修改密码 |
 | `/api/v1/auth/setup-status` | GET | 检查 admin 是否存在 |
-| `/api/v1/auth/initialize` | POST | 首次初始化第一个 admin（仅无 admin 时可调用） |
 
 ## 兼容性
 
-- **本地开发**（`make dev`）：Gateway embedded runtime 完全兼容；无 admin 时访问 `/setup` 初始化
+- **标准模式**（`make dev`）：完全兼容，admin 自动创建
-- **Gateway embedded runtime**：标准脚本、Docker dev 和生产部署均通过 Gateway 提供认证与 LangGraph-compatible API
+- **Gateway 模式**（`make dev-pro`）：完全兼容
-- **Docker 部署**：完全兼容，`.deer-flow/data/deerflow.db` 需持久化卷挂载
+- **Docker 部署**：完全兼容，`.deer-flow/users.db` 需持久化卷挂载
-- **IM 渠道**（Feishu/Slack/Telegram）：通过 Gateway 内部认证通信，使用 `default` 用户桶
+- **IM 渠道**（Feishu/Slack/Telegram）：通过 LangGraph SDK 通信，不经过认证层
 - **DeerFlowClient**（嵌入式）：不经过 HTTP，不受认证影响
 
 ## 故障排查
 
 | 症状 | 原因 | 解决 |
 |------|------|------|
-| 启动后没看到密码 | 当前实现不在启动日志输出密码 | 首次安装访问 `/setup`；忘记密码用 `reset_admin` |
+| 启动后没看到密码 | admin 已存在（非首次启动） | 用 `reset_admin` 重置，或删 `users.db` |
-| `/login` 自动跳到 `/setup` | 系统还没有 admin | 在 `/setup` 创建第一个 admin |
 | 登录后 POST 返回 403 | CSRF token 缺失 | 确认前端已更新 |
-| 重启后需要重新登录 | `.jwt_secret` 文件被删除且 `.env` 未设置 `AUTH_JWT_SECRET` | 在 `.env` 中设置固定密钥 |
+| 重启后需要重新登录 | `AUTH_JWT_SECRET` 未持久化 | 在 `.env` 中设置固定密钥 |

backend/docs/BLOCKING_IO_DETECTION.md

@@ -1,154 +0,0 @@
-# Blocking IO detection usage and maintenance
-
-This document describes how to use and maintain DeerFlow backend blocking-IO
-detection for async event-loop safety.
-
-The goal is narrow: find and prevent synchronous IO from blocking backend
-async event-loop paths. Static and runtime detection are complementary, but
-they have different jobs.
-
-## Static detector
-
-The static detector is the discovery tool. It scans backend source code and
-reports candidate blocking-IO call sites that may need human review.
-
-Run it from the repository root:
-
-```bash
-make detect-blocking-io
-```
-
-Or from `backend/`:
-
-```bash
-make detect-blocking-io
-```
-
-The report is written to:
-
-```text
-.deer-flow/blocking-io-findings.json
-```
-
-Use this output for review and triage. A static finding is a candidate, not
-proof that production blocks the event loop at runtime. The current static
-rules are intentionally broad; prefer triaging existing output before adding
-new static rules.
-
-Add a static rule only when review finds a recurring high-risk blocking
-pattern that is invisible to the current detector.
-
-## Runtime detector
-
-The runtime detector is the CI regression guard. It uses Blockbuster to fail a
-focused test when code under `app.*` or `deerflow.*` performs blocking IO on
-the asyncio event-loop thread.
-
-Run it from `backend/`:
-
-```bash
-make test-blocking-io
-```
-
-The runtime gate starts from confirmed production bugs and protects those
-paths from regressing. It does not prove that the entire backend is free of
-blocking IO; it only covers the production paths exercised by
-`backend/tests/blocking_io/`.
-
-## Maintenance workflow
-
-Use the static detector to find candidates, then use review to decide which
-async production paths are worth protecting in CI.
-
-The normal workflow is:
-
-1. Run the static detector to find backend blocking-IO candidates.
-2. Use human review to pick high-risk production async paths.
-3. Add or update a focused runtime anchor in `backend/tests/blocking_io/`.
-4. Let CI prevent that path from regressing.
-
-Runtime detection has two maintenance paths.
-
-### Add a runtime rule
-
-Add a runtime rule when Blockbuster's default rules do not cover a generic
-blocking primitive used by production code.
-
-Rules belong in:
-
-```text
-backend/tests/support/detectors/blocking_io_runtime.py
-```
-
-Add them to `_PROJECT_BLOCKING_RULES`, not directly inside individual tests.
-Keeping rules centralized makes it clear which extra primitives DeerFlow
-expects Blockbuster to catch.
-
-Example shape:
-
-```python
-import subprocess
-
-from blockbuster import BlockBusterFunction
-
-_PROJECT_BLOCKING_RULES = (
-    (
-        "subprocess.Popen.__init__",
-        BlockBusterFunction(
-            subprocess.Popen,
-            "__init__",
-            scanned_modules=["app", "deerflow"],
-        ),
-    ),
-)
-```
-
-Do not add a runtime rule just because a business path is not tested. A rule
-only expands what Blockbuster can intercept after code runs.
-
-### Add a runtime anchor
-
-Add a runtime anchor when a high-risk async production path should be protected
-by CI but no existing `backend/tests/blocking_io/` test executes it.
-
-Anchors belong in:
-
-```text
-backend/tests/blocking_io/
-```
-
-A good anchor should:
-
-- Call the real production async entry point.
-- Avoid bypassing the blocking surface with test-only `asyncio.to_thread`
-  wrappers.
-- Use real local filesystem inputs when the bug shape is filesystem IO.
-- Mock only the external dependency boundary, such as a network service or
-  third-party saver class.
-- Fail if a future change moves the blocking operation back onto the event
-  loop.
-
-Avoid testing only the low-level helper unless that helper is the production
-async entry point. The runtime gate is most useful when it protects the caller
-that production actually executes.
-
-## Current runtime coverage
-
-The runtime anchors protect confirmed blocking-IO bug shapes:
-
-- SQLite checkpointer setup, including path resolution and parent-directory
-  creation.
-- Subagent skill metadata loading through `SubagentExecutor._load_skills()`.
-- `JsonlRunEventStore` async API (`put` / `list_*` / `delete_*`): the JSONL
-  run-event backend offloads its synchronous file IO via `asyncio.to_thread`
-  (fix #3084); this anchor drives the real async API under the gate so any
-  blocking IO reintroduced on the loop fails, not only removal of one
-  `to_thread` call.
-- `UploadsMiddleware.before_agent` uploads-directory scan: a sync-only middleware
-  hook runs on the event loop under async graph execution, so the scan is
-  offloaded via `abefore_agent` + `run_in_executor`.
-- Gate health checks: Blockbuster catches unoffloaded calls, opt-out works, and
-  patches are restored after exceptions.
-
-As static detection and review identify more high-risk async paths, add new
-runtime anchors incrementally.

backend/docs/CONFIGURATION.md

@@ -36,7 +36,6 @@ models:
 - OpenAI (`langchain_openai:ChatOpenAI`)
 - Anthropic (`langchain_anthropic:ChatAnthropic`)
 - DeepSeek (`langchain_deepseek:ChatDeepSeek`)
-- Xiaomi MiMo (`deerflow.models.patched_mimo:PatchedChatMiMo`)
 - Claude Code OAuth (`deerflow.models.claude_provider:ClaudeChatModel`)
 - Codex CLI (`deerflow.models.openai_codex_provider:CodexChatModel`)
 - Any LangChain-compatible provider
@@ -95,35 +94,25 @@ models:
         thinking:
           type: enabled
 
-  - name: minimax-m3
+  - name: minimax-m2.5
-    display_name: MiniMax M3
+    display_name: MiniMax M2.5
     use: langchain_openai:ChatOpenAI
-    model: MiniMax-M3
+    model: MiniMax-M2.5
     api_key: $MINIMAX_API_KEY
     base_url: https://api.minimax.io/v1
     max_tokens: 4096
     temperature: 1.0  # MiniMax requires temperature in (0.0, 1.0]
     supports_vision: true
 
-  - name: minimax-m2.7
+  - name: minimax-m2.5-highspeed
-    display_name: MiniMax M2.7
+    display_name: MiniMax M2.5 Highspeed
     use: langchain_openai:ChatOpenAI
-    model: MiniMax-M2.7
+    model: MiniMax-M2.5-highspeed
     api_key: $MINIMAX_API_KEY
     base_url: https://api.minimax.io/v1
     max_tokens: 4096
     temperature: 1.0  # MiniMax requires temperature in (0.0, 1.0]
-    supports_vision: false  # M2.7 is text-only; M3 supports vision
+    supports_vision: true
-
-  - name: minimax-m2.7-highspeed
-    display_name: MiniMax M2.7 Highspeed
-    use: langchain_openai:ChatOpenAI
-    model: MiniMax-M2.7-highspeed
-    api_key: $MINIMAX_API_KEY
-    base_url: https://api.minimax.io/v1
-    max_tokens: 4096
-    temperature: 1.0  # MiniMax requires temperature in (0.0, 1.0]
-    supports_vision: false  # M2.7 is text-only; M3 supports vision
   - name: openrouter-gemini-2.5-flash
     display_name: Gemini 2.5 Flash (OpenRouter)
     use: langchain_openai:ChatOpenAI
@@ -177,37 +166,6 @@ models:
 
 For Gemini accessed **without** thinking (e.g. via OpenRouter where thinking is not activated), the plain `langchain_openai:ChatOpenAI` with `supports_thinking: false` is sufficient and no patch is needed.
 
-**MiMo with thinking via OpenAI-compatible API**:
-
-MiMo returns `reasoning_content` on assistant messages in thinking mode. In multi-turn agent conversations with tool calls, subsequent requests must preserve that historical `reasoning_content` on assistant messages or the MiMo API can return HTTP 400. Standard `langchain_openai:ChatOpenAI` drops this provider-specific field, so use `deerflow.models.patched_mimo:PatchedChatMiMo`:
-
-For pay-as-you-go API keys (`sk-...`), use `https://api.xiaomimimo.com/v1`. For Token Plan keys (`tp-...`), use the regional Token Plan Base URL shown in the MiMo console, such as `https://token-plan-cn.xiaomimimo.com/v1`. MiMo documents these key types as separate and non-interchangeable.
-
-`PatchedChatMiMo` is model-id agnostic. Use it for every MiMo thinking model entry you configure, including model entries referenced by `subagents.*.model` overrides (for example `mimo-v2.5-pro`, `mimo-v2.5`, `mimo-v2-pro`, `mimo-v2-omni`, or `mimo-v2-flash`).
-
-```yaml
-models:
-  - name: mimo-v2.5-pro
-    display_name: MiMo V2.5 Pro
-    use: deerflow.models.patched_mimo:PatchedChatMiMo
-    model: mimo-v2.5-pro
-    api_key: $MIMO_API_KEY
-    base_url: https://api.xiaomimimo.com/v1
-    max_tokens: 8192
-    supports_thinking: true
-    supports_vision: false
-    when_thinking_enabled:
-      extra_body:
-        thinking:
-          type: enabled
-    when_thinking_disabled:
-      extra_body:
-        thinking:
-          type: disabled
-```
-
-`PatchedChatMiMo` preserves MiMo's `choices[].message.reasoning_content`, streaming `delta.reasoning_content`, and request-history assistant `reasoning_content` fields. It does not reuse the DeepSeek provider.
-
 ### Tool Groups
 
 Organize tools into logical groups:
@@ -361,7 +319,6 @@ models:
 - `OPENAI_API_KEY` - OpenAI API key
 - `ANTHROPIC_API_KEY` - Anthropic API key
 - `DEEPSEEK_API_KEY` - DeepSeek API key
-- `MIMO_API_KEY` - Xiaomi MiMo API key
 - `NOVITA_API_KEY` - Novita API key (OpenAI-compatible endpoint)
 - `TAVILY_API_KEY` - Tavily search API key
 - `DEER_FLOW_PROJECT_ROOT` - Project root for relative runtime paths

backend/docs/MCP_SERVER.md

@@ -14,19 +14,6 @@ DeerFlow supports configurable MCP servers and skills to extend its capabilities
 3. Configure each server’s command, arguments, and environment variables as needed.
 4. Restart the application to load and register MCP tools.
 
-## Filesystem MCP Servers
-
-DeerFlow already provides built-in file tools for thread-scoped workspace access.
-Do not add an MCP filesystem server for the same DeerFlow workspace. The
-overlapping file tools use different path semantics, which can make LLM tool
-selection and file access behavior unstable.
-
-DeerFlow does not currently adapt the MCP Roots mode for filesystem servers. In
-particular, it does not publish per-thread MCP roots or map DeerFlow sandbox
-paths such as `/mnt/user-data/...` to paths accepted by
-`@modelcontextprotocol/server-filesystem`. Use DeerFlow's built-in file tools
-for DeerFlow workspace files.
-
 ## OAuth Support (HTTP/SSE MCP Servers)
 
 For `http` and `sse` MCP servers, DeerFlow supports OAuth token acquisition and automatic token refresh.
@@ -101,6 +88,7 @@ MCP servers expose tools that are automatically discovered and integrated into D
 
 MCP servers can provide access to:
 
+- **File systems**
 - **Databases** (e.g., PostgreSQL)
 - **External APIs** (e.g., GitHub, Brave Search)
 - **Browser automation** (e.g., Puppeteer)
@@ -109,4 +97,4 @@ MCP servers can provide access to:
 ## Learn More
 
 For detailed documentation about the Model Context Protocol, visit:  
-https://modelcontextprotocol.io
+https://modelcontextprotocol.io

backend/docs/README.md

@@ -8,7 +8,6 @@ This directory contains detailed documentation for the DeerFlow backend.
 |----------|-------------|
 | [ARCHITECTURE.md](ARCHITECTURE.md) | System architecture overview |
 | [API.md](API.md) | Complete API reference |
-| [AUTH_DESIGN.md](AUTH_DESIGN.md) | User authentication, CSRF, and per-user isolation design |
 | [CONFIGURATION.md](CONFIGURATION.md) | Configuration options |
 | [SETUP.md](SETUP.md) | Quick setup guide |
 
@@ -19,7 +18,6 @@ This directory contains detailed documentation for the DeerFlow backend.
 | [STREAMING.md](STREAMING.md) | Token-level streaming design: Gateway vs DeerFlowClient paths, `stream_mode` semantics, per-id dedup |
 | [FILE_UPLOAD.md](FILE_UPLOAD.md) | File upload functionality |
 | [PATH_EXAMPLES.md](PATH_EXAMPLES.md) | Path types and usage examples |
-| [SANDBOX_MEMORY_PROFILING.md](SANDBOX_MEMORY_PROFILING.md) | Sandbox memory baseline and runtime comparison guide |
 | [summarization.md](summarization.md) | Context summarization feature |
 | [plan_mode_usage.md](plan_mode_usage.md) | Plan mode with TodoList |
 | [AUTO_TITLE_GENERATION.md](AUTO_TITLE_GENERATION.md) | Automatic title generation |
@@ -44,7 +42,6 @@ docs/
 ├── README.md                  # This file
 ├── ARCHITECTURE.md            # System architecture
 ├── API.md                     # API reference
-├── AUTH_DESIGN.md             # User authentication and isolation design
 ├── CONFIGURATION.md           # Configuration guide
 ├── SETUP.md                   # Setup instructions
 ├── FILE_UPLOAD.md             # File upload feature

backend/docs/REPLAY_E2E.md

@@ -1,120 +0,0 @@
-# Record/Replay E2E — front-back contract verification
-
-Deterministic, **key-free** end-to-end checks that a backend change can't
-silently break the frontend (and vice-versa). Two complementary layers, fed by a
-single recording.
-
-## Why
-
-The mock-based frontend e2e hand-writes the backend's JSON/SSE, so a backend
-schema or SSE change passes green ("fake green"). These layers replay a recorded
-**real** run against the **real** backend (and, for Layer 2, the real frontend),
-so contract drift turns the build red instead.
-
-## The two layers
-
-- **Layer 1 — backend golden** (`tests/test_replay_golden.py`): replays a fixture
-  through the real FastAPI gateway with `ReplayChatModel` and asserts the streamed
-  SSE event sequence equals a committed golden. Fast, no browser. Guards protocol
-  *shape*.
-- **Layer 2 — full-stack render** (`frontend/tests/e2e-real-backend/`): real
-  Next.js + real gateway (replay model) + Chromium; asserts the replayed
-  auto-title and a follow-up suggestion render in the browser. Guards semantic
-  *render*. (Complementary to Layer 1 — neither subsumes the other.)
-
-Layer 2 also hosts **cross-stack contract scenarios** — the dangerous class
-where a backend change silently breaks a frontend assumption and *both sides'
-unit tests stay green*. See below.
-
-## Cross-stack scenario: multi-run render order (`multi-run-order.spec.ts`)
-
-Regression guard for issue **#3352** (after context compression, refreshing a
-thread rendered history out of order). Root cause was a front-back desync:
-backend `RunManager.list_by_thread` returns runs **newest-first** (PR #2932),
-while the frontend (`core/threads/hooks.ts`) iterated runs and **prepended** each
-loaded page — inverting chronological order once the checkpoint no longer held
-the older messages. The backend ordering test was green throughout, and the
-frontend regression unit test hardcodes "backend returns newest-first" in a mock,
-so only a *real frontend against a real backend* catches the desync.
-
-This scenario does **not** record a conversation. It uses a **test-only seeder**
-(`tests/seed_runs_router.py`, mounted on the replay gateway only when
-`DEERFLOW_ENABLE_TEST_SEED=1`) to stand up a thread with ≥2 runs and per-run
-message events — and deliberately **no checkpoint**, which is the #3352
-precondition: it forces the frontend's per-run reload path to be the sole source
-of truth so the ordering bug becomes observable. The seeder writes through the
-gateway's own run/event stores using the request's auth context, so the real
-`list_by_thread` → `/runs/{id}/messages` → prepend path runs live. Reverting the
-#3354 frontend fix turns this spec red.
-
-## How replay works
-
-`tests/replay_provider.py::ReplayChatModel` returns recorded assistant turns keyed
-by a **normalized hash of the model caller + conversation**. The conversation is
-human / ai / tool messages — role, text, tool-call name+args; with
-`<system-reminder>`, dates, UUIDs, tmp paths stripped. The caller is the stable
-source of the model call (`lead_agent`, `middleware:title`, `suggest_agent`,
-`subagent:*`, etc.). A miss raises loudly rather than passing silently.
-
-**The system prompt is excluded from the match key.** The lead-agent system
-prompt is a living, frequently-edited implementation detail — its wording changes
-across PRs (e.g. #3195 added a "File Editing Workflow" section). Hashing it would
-make every fixture go stale and red-fail unrelated PRs the moment anyone edits the
-prompt. The conversation flow (user input → tool calls → results → answer) is the
-stable contract that identifies a recorded turn. The caller still stays in the
-key so two different model users with identical conversation text do not compete
-for the same replay bucket. (This mirrors how open-design's mock picker keys on
-the user prompt, not the system internals.) Combined with pinning skills +
-extensions empty and disabling memory/summarization
-(`tests/_replay_fixture.py::build_config_yaml`), a fixture replays the same across
-machines, days, prompt edits, and CI. Replaying needs **no API key**.
-
-A swallowed hash-miss keeps the SSE *event shapes* identical (the gateway wraps it
-into a normal assistant error message), so the Layer-1 golden can't catch a miss
-by shape alone — it inspects `replay_provider.replay_misses()` and fails loud
-instead. Layer-2 already fails on a miss (the recorded turns never render).
-
-## Record a new scenario (needs a real key — dev machine only)
-
-Recording drives the **real frontend** so captured inputs match exactly what the
-browser sends; fixtures contain no API key.
-
-```bash
-# 1. drive the real frontend against a real-model gateway, capturing model calls
-OPENAI_API_KEY=... OPENAI_API_BASE=<openai-compatible-endpoint>/v1 \
-  DEERFLOW_RECORD_OUT=/tmp/rec/turns.jsonl RECORD_MODEL=<model> \
-  bash -c 'cd frontend && pnpm exec playwright test -c playwright.record.config.ts'
-
-# 2. stitch the capture into a fixture
-cd backend && uv run python scripts/build_fixture_from_jsonl.py \
-  --jsonl /tmp/rec/turns.jsonl --meta /tmp/rec/turns.jsonl.meta.json \
-  --out tests/fixtures/replay/<scenario>.<mode>.json --model <model>
-
-# 3. regenerate the committed golden
-DEERFLOW_WRITE_GOLDEN=1 PYTHONPATH=. uv run pytest tests/test_replay_golden.py
-```
-
-## Run (no key)
-
-```bash
-cd backend  && PYTHONPATH=. uv run pytest tests/test_replay_golden.py          # Layer 1
-cd frontend && pnpm exec playwright test -c playwright.real-backend.config.ts  # Layer 2
-```
-
-## CI
-
-`.github/workflows/replay-e2e.yml` runs both layers on changes to **either** side
-of the contract (`frontend/**`, `backend/app/gateway/**`,
-`backend/packages/harness/**`, fixtures). DOM assertions are the gate; the rendered
-screenshot + Playwright HTML report are uploaded as a CI artifact.
-
-## Known limitations
-
-- Visual regression baselines are OS-specific, so they are a **local dev gate
-  only** (gitignored); CI uploads the render as an artifact for human review
-  instead of hard-asserting a cross-OS baseline.
-- Fixtures are coupled to the recording-time prompt; if new
-  environment-dependent content enters the system prompt, extend the
-  normalization in `replay_provider.py` (or pin it in `build_config_yaml`).
-- Re-record a scenario if the agent graph changes how many model calls it makes
-  — the replay raises loudly on a hash miss pointing at the divergence.

backend/docs/SANDBOX_MEMORY_PROFILING.md

@@ -1,81 +0,0 @@
-# Sandbox Memory Profiling
-
-This guide records a repeatable baseline before changing the sandbox runtime.
-Issue #3213 reports per-sandbox memory near 1 GiB in Kubernetes. Before adding
-or recommending a new provider, capture the current AIO sandbox baseline and
-compare candidates with the same DeerFlow workload.
-
-## What to Measure
-
-Measure at least these samples:
-
-1. Empty sandbox after it becomes ready.
-2. After a simple bash command.
-3. After a Python task that imports common packages.
-4. After a Node task when Node-based workloads are expected.
-5. After generating files under `/mnt/user-data/outputs`.
-6. After release and warm reuse.
-7. At the target concurrency level, for example 10, 50, or 100 sandboxes.
-
-`kubectl top` reports Kubernetes/container working set memory. Treat it as a
-capacity signal, not exclusive RSS/PSS. Pod-level memory includes every
-container in the Pod and may include cache charged to the cgroup. If a result
-looks surprising, inspect the sandbox processes and cgroup metrics on the node
-before drawing conclusions.
-
-## Capture a Snapshot
-
-Run this from the repository root:
-
-```bash
-python scripts/sandbox_memory_profile.py \
-  --namespace deer-flow \
-  --selector app=deer-flow-sandbox \
-  --sample empty \
-  --include-processes \
-  --format markdown
-```
-
-Use a descriptive `--sample` value for each phase:
-
-```bash
-python scripts/sandbox_memory_profile.py --sample after-bash --format json
-python scripts/sandbox_memory_profile.py --sample after-python --format json
-python scripts/sandbox_memory_profile.py --sample after-artifact --format json
-```
-
-`--include-processes` runs `kubectl exec ... ps` in each sandbox Pod and adds
-the highest-RSS processes to the report. This helps distinguish Pod-level cgroup
-memory from process RSS. The two numbers will not match exactly because cgroup
-memory can include cache and other kernel-accounted memory.
-
-Save the raw JSON when comparing backends so totals, pod names, images,
-requests, limits, and timestamps can be audited later.
-
-## Candidate Runtime Matrix
-
-For AIO, CubeSandbox, OpenSandbox, gVisor, Kata, or another candidate, compare
-the same workload and record:
-
-| Area | Required Evidence |
-| --- | --- |
-| Capacity | Pod or instance count, total memory, average memory, max memory |
-| Startup | Ready latency at 1, 10, 50, and 100 concurrent sandboxes |
-| Commands | Bash output, timeout behavior, failure shape |
-| Files | `read_file`, `write_file`, binary `update_file`, `list_dir`, `glob`, `grep` |
-| Uploads | Files uploaded by the gateway are visible inside the sandbox |
-| Artifacts | Files written to `/mnt/user-data/outputs` are readable by the backend artifact API |
-| Paths | `/mnt/user-data/workspace`, `/mnt/user-data/uploads`, `/mnt/user-data/outputs`, `/mnt/acp-workspace`, and skills paths keep their expected semantics |
-| Isolation | Different users and threads cannot read each other's data |
-| Cleanup | Release, idle timeout, process restart, and orphan cleanup free resources |
-| Operations | Deployment prerequisites, privileged components, networking, storage, and upgrade path |
-
-## PR Guidance
-
-Do not claim that a new provider fixes high-concurrency memory usage until the
-same DeerFlow workload has been measured on both the current AIO sandbox and the
-candidate backend.
-
-For an experimental provider PR, prefer `Related to #3213` unless the PR also
-includes reproducible DeerFlow workload data that demonstrates the target memory
-reduction and preserves uploads, outputs, artifacts, and isolation behavior.

backend/docs/TODO.md

@@ -26,7 +26,7 @@
   - Replace sync `requests` with `httpx.AsyncClient` in community tools (tavily, jina_ai, firecrawl, infoquest, image_search)
   - [x] Replace sync `model.invoke()` with async `model.ainvoke()` in title_middleware and memory updater
   - Consider `asyncio.to_thread()` wrapper for remaining blocking file I/O
-  - For production: tune Gateway worker/runtime settings for long-running agent workloads
+  - For production: use `langgraph up` (multi-worker) instead of `langgraph dev` (single-worker)
 
 ## Resolved Issues
 

backend/docs/middleware-execution-flow.md

@@ -4,22 +4,22 @@
 
 `create_deerflow_agent` 通过 `RuntimeFeatures` 组装的完整 middleware 链（默认全开时）：
 
-| # | Middleware | `before_agent` | `before_model` | `after_model` | `after_agent` | `wrap_model_call` | `wrap_tool_call` | 主 Agent | Subagent | 来源 |
+| # | Middleware | `before_agent` | `before_model` | `after_model` | `after_agent` | `wrap_tool_call` | 主 Agent | Subagent | 来源 |
-|---|-----------|:-:|:-:|:-:|:-:|:-:|:-:|:-:|:-:|------|
+|---|-----------|:-:|:-:|:-:|:-:|:-:|:-:|:-:|------|
-| 0 | ThreadDataMiddleware | ✓ | | | | | | ✓ | ✓ | `sandbox` |
+| 0 | ThreadDataMiddleware | ✓ | | | | | ✓ | ✓ | `sandbox` |
-| 1 | UploadsMiddleware | ✓ | | | | | | ✓ | ✗ | `sandbox` |
+| 1 | UploadsMiddleware | ✓ | | | | | ✓ | ✗ | `sandbox` |
-| 2 | SandboxMiddleware | ✓ | | | ✓ | | | ✓ | ✓ | `sandbox` |
+| 2 | SandboxMiddleware | ✓ | | | ✓ | | ✓ | ✓ | `sandbox` |
-| 3 | DanglingToolCallMiddleware | | | | | ✓ | | ✓ | ✗ | 始终开启 |
+| 3 | DanglingToolCallMiddleware | | | ✓ | | | ✓ | ✗ | 始终开启 |
-| 4 | GuardrailMiddleware | | | | | | ✓ | ✓ | ✓ | *Phase 2 纳入* |
+| 4 | GuardrailMiddleware | | | | | ✓ | ✓ | ✓ | *Phase 2 纳入* |
-| 5 | ToolErrorHandlingMiddleware | | | | | | ✓ | ✓ | ✓ | 始终开启 |
+| 5 | ToolErrorHandlingMiddleware | | | | | ✓ | ✓ | ✓ | 始终开启 |
-| 6 | SummarizationMiddleware | | ✓ | | | | | ✓ | ✗ | `summarization` |
+| 6 | SummarizationMiddleware | | | ✓ | | | ✓ | ✗ | `summarization` |
-| 7 | TodoMiddleware | | ✓ | ✓ | | ✓ | | ✓ | ✗ | `plan_mode` 参数 |
+| 7 | TodoMiddleware | | | ✓ | | | ✓ | ✗ | `plan_mode` 参数 |
-| 8 | TitleMiddleware | | | ✓ | | | | ✓ | ✗ | `auto_title` |
+| 8 | TitleMiddleware | | | ✓ | | | ✓ | ✗ | `auto_title` |
-| 9 | MemoryMiddleware | | | | ✓ | | | ✓ | ✗ | `memory` |
+| 9 | MemoryMiddleware | | | | ✓ | | ✓ | ✗ | `memory` |
-| 10 | ViewImageMiddleware | | ✓ | | | | | ✓ | ✗ | `vision` |
+| 10 | ViewImageMiddleware | | ✓ | | | | ✓ | ✗ | `vision` |
-| 11 | SubagentLimitMiddleware | | | ✓ | | | | ✓ | ✗ | `subagent` |
+| 11 | SubagentLimitMiddleware | | | ✓ | | | ✓ | ✗ | `subagent` |
-| 12 | LoopDetectionMiddleware | ✓ | | ✓ | ✓ | ✓ | | ✓ | ✗ | 始终开启 |
+| 12 | LoopDetectionMiddleware | | | ✓ | | | ✓ | ✗ | 始终开启 |
-| 13 | ClarificationMiddleware | | | | | | ✓ | ✓ | ✗ | 始终最后 |
+| 13 | ClarificationMiddleware | | | ✓ | | | ✓ | ✗ | 始终最后 |
 
 主 agent **14 个** middleware（`make_lead_agent`），subagent **4 个**（ThreadData、Sandbox、Guardrail、ToolErrorHandling）。`create_deerflow_agent` Phase 1 实现 **13 个**（Guardrail 仅支持自定义实例，无内置默认）。
 
@@ -35,7 +35,7 @@ graph TB
 
     subgraph BA ["<b>before_agent</b> 正序 0→N"]
         direction TB
-        TD["[0] ThreadData<br/>创建线程目录"] --> UL["[1] Uploads<br/>扫描上传文件"] --> SB["[2] Sandbox<br/>获取沙箱"] --> LD_BA["[12] LoopDetection<br/>清理 stale warning"]
+        TD["[0] ThreadData<br/>创建线程目录"] --> UL["[1] Uploads<br/>扫描上传文件"] --> SB["[2] Sandbox<br/>获取沙箱"]
     end
 
     subgraph BM ["<b>before_model</b> 正序 0→N"]
@@ -43,42 +43,34 @@ graph TB
         VI["[10] ViewImage<br/>注入图片 base64"]
     end
 
-    subgraph WM ["<b>wrap_model_call</b>"]
+    SB --> VI
-        direction TB
+    VI --> M["<b>MODEL</b>"]
-        DTC_WM["[3] DanglingToolCall<br/>补悬空 ToolMessage"] --> LD_WM["[12] LoopDetection<br/>注入当前 run warning"]
-    end
-
-    LD_BA --> VI
-    VI --> DTC_WM
-    LD_WM --> M["<b>MODEL</b>"]
 
     subgraph AM ["<b>after_model</b> 反序 N→0"]
         direction TB
-        LD["[12] LoopDetection<br/>检测循环/排队 warning"] --> SL["[11] SubagentLimit<br/>截断多余 task"] --> TI["[8] Title<br/>生成标题"]
+        CL["[13] Clarification<br/>拦截 ask_clarification"] --> LD["[12] LoopDetection<br/>检测循环"] --> SL["[11] SubagentLimit<br/>截断多余 task"] --> TI["[8] Title<br/>生成标题"] --> SM["[6] Summarization<br/>上下文压缩"] --> DTC["[3] DanglingToolCall<br/>补缺失 ToolMessage"]
     end
 
-    M --> LD
+    M --> CL
 
     subgraph AA ["<b>after_agent</b> 反序 N→0"]
         direction TB
-        LD_CLEAN["[12] LoopDetection<br/>清理 pending warning"] --> MEM["[9] Memory<br/>入队记忆"] --> SBR["[2] Sandbox<br/>释放沙箱"]
+        SBR["[2] Sandbox<br/>释放沙箱"] --> MEM["[9] Memory<br/>入队记忆"]
     end
 
-    TI --> LD_CLEAN
+    DTC --> SBR
-    SBR --> END(["response"])
+    MEM --> END(["response"])
 
     classDef beforeNode fill:#a0a8b5,stroke:#636b7a,color:#2d3239
     classDef modelNode fill:#b5a8a0,stroke:#7a6b63,color:#2d3239
-    classDef wrapModelNode fill:#a8a0b5,stroke:#6b637a,color:#2d3239
     classDef afterModelNode fill:#b5a0a8,stroke:#7a636b,color:#2d3239
     classDef afterAgentNode fill:#a0b5a8,stroke:#637a6b,color:#2d3239
     classDef terminalNode fill:#a8b5a0,stroke:#6b7a63,color:#2d3239
 
-    class TD,UL,SB,LD_BA,VI beforeNode
+    class TD,UL,SB,VI beforeNode
-    class DTC_WM,LD_WM wrapModelNode
     class M modelNode
-    class LD,SL,TI afterModelNode
+    class CL,LD,SL,TI,SM,DTC afterModelNode
-    class LD_CLEAN,SBR,MEM afterAgentNode
+    class SBR,MEM afterAgentNode
     class START,END terminalNode
 ```
 
@@ -90,12 +82,13 @@ sequenceDiagram
     participant TD as ThreadDataMiddleware
     participant UL as UploadsMiddleware
     participant SB as SandboxMiddleware
-    participant LD as LoopDetectionMiddleware
     participant VI as ViewImageMiddleware
-    participant DTC as DanglingToolCallMiddleware
     participant M as MODEL
+    participant CL as ClarificationMiddleware
     participant SL as SubagentLimitMiddleware
     participant TI as TitleMiddleware
+    participant SM as SummarizationMiddleware
+    participant DTC as DanglingToolCallMiddleware
     participant MEM as MemoryMiddleware
 
     U ->> TD: invoke
@@ -110,26 +103,19 @@ sequenceDiagram
     activate SB
     Note right of SB: before_agent 获取沙箱
 
-    SB ->> LD: before_agent
+    SB ->> VI: before_model
-    activate LD
-    Note right of LD: before_agent 清理同 thread 旧 run 的 pending warning
-    LD ->> VI: before_model
     activate VI
     Note right of VI: before_model 注入图片 base64
 
-    VI ->> DTC: wrap_model_call
+    VI ->> M: messages + tools
-    activate DTC
-    Note right of DTC: wrap_model_call 补悬空 ToolMessage
-    DTC ->> LD: wrap_model_call
-    Note right of LD: wrap_model_call drain 当前 run warning 并追加到末尾
-    LD ->> M: messages + tools
     activate M
-    M -->> LD: AI response
+    M -->> CL: AI response
     deactivate M
 
-    Note right of LD: after_model 检测循环；warning 入队，hard-stop 清 tool_calls
+    activate CL
-    LD -->> SL: after_model
+    Note right of CL: after_model 拦截 ask_clarification
-    deactivate LD
+    CL -->> SL: after_model
+    deactivate CL
 
     activate SL
     Note right of SL: after_model 截断多余 task
@@ -138,18 +124,22 @@ sequenceDiagram
 
     activate TI
     Note right of TI: after_model 生成标题
-    TI -->> DTC: done
+    TI -->> SM: after_model
     deactivate TI
 
+    activate SM
+    Note right of SM: after_model 上下文压缩
+    SM -->> DTC: after_model
+    deactivate SM
+
+    activate DTC
+    Note right of DTC: after_model 补缺失 ToolMessage
+    DTC -->> VI: done
     deactivate DTC
 
     VI -->> SB: done
     deactivate VI
 
-    Note right of LD: after_agent 清理当前 run 未消费 warning
-
-    Note right of MEM: after_agent 入队记忆
-
     Note right of SB: after_agent 释放沙箱
     SB -->> UL: done
     deactivate SB
@@ -157,6 +147,8 @@ sequenceDiagram
     UL -->> TD: done
     deactivate UL
 
+    Note right of MEM: after_agent 入队记忆
+
     TD -->> U: response
     deactivate TD
 ```
@@ -232,12 +224,12 @@ sequenceDiagram
     participant TD as ThreadData
     participant UL as Uploads
     participant SB as Sandbox
-    participant LD as LoopDetection
     participant VI as ViewImage
-    participant DTC as DanglingToolCall
     participant M as MODEL
+    participant CL as Clarification
     participant SL as SubagentLimit
     participant TI as Title
+    participant SM as Summarization
     participant MEM as Memory
 
     U ->> TD: invoke
@@ -246,40 +238,34 @@ sequenceDiagram
     Note right of UL: before_agent 扫描文件
     UL ->> SB: .
     Note right of SB: before_agent 获取沙箱
-    SB ->> LD: .
-    Note right of LD: before_agent 清理 stale pending warning
 
     loop 每轮对话（tool call 循环）
         SB ->> VI: .
         Note right of VI: before_model 注入图片
-        VI ->> DTC: .
+        VI ->> M: messages + tools
-        Note right of DTC: wrap_model_call 补悬空工具结果
+        M -->> CL: AI response
-        DTC ->> LD: .
+        Note right of CL: after_model 拦截 ask_clarification
-        Note right of LD: wrap_model_call 注入当前 run warning
+        CL -->> SL: .
-        LD ->> M: messages + tools
-        M -->> LD: AI response
-        Note right of LD: after_model 检测循环/排队 warning
-        LD -->> SL: .
         Note right of SL: after_model 截断多余 task
         SL -->> TI: .
         Note right of TI: after_model 生成标题
+        TI -->> SM: .
+        Note right of SM: after_model 上下文压缩
     end
 
-    Note right of LD: after_agent 清理当前 run pending warning
-    LD -->> MEM: .
-    Note right of MEM: after_agent 入队记忆
-    MEM -->> SB: .
     Note right of SB: after_agent 释放沙箱
-    SB -->> U: response
+    SB -->> MEM: .
+    Note right of MEM: after_agent 入队记忆
+    MEM -->> U: response
 ```
 
 > [!warning] 不是洋葱
-> 大部分 middleware 只用一个阶段。SandboxMiddleware 使用 `before_agent`/`after_agent` 做资源获取/释放；LoopDetectionMiddleware 也使用这两个钩子，但用途是清理 run-scoped pending warnings，不是资源生命周期对称。`before_agent` / `after_agent` 只跑一次，`before_model` / `after_model` / `wrap_model_call` 每轮循环都跑。
+> 14 个 middleware 中只有 SandboxMiddleware 有 before/after 对称（获取/释放）。其余都是单向的：要么只在 `before_*` 做事，要么只在 `after_*` 做事。`before_agent` / `after_agent` 只跑一次，`before_model` / `after_model` 每轮循环都跑。
 
 硬依赖只有 2 处：
 
 1. **ThreadData 在 Sandbox 之前** — sandbox 需要线程目录
-2. **Clarification 在列表最后** — `wrap_tool_call` 处理 `ask_clarification` 时优先拦截，并通过 `Command(goto=END)` 中断执行
+2. **Clarification 在列表最后** — `after_model` 反序时最先执行，第一个拦截 `ask_clarification`
 
 ### 结论
 
@@ -287,19 +273,19 @@ sequenceDiagram
 |---|---|---|
 | 每个 middleware | before + after 对称 | 大多只用一个钩子 |
 | 激活条 | 嵌套（外长内短） | 不嵌套（串行） |
-| 反序的意义 | 清理与初始化配对 | 影响 `after_model` / `after_agent` 的执行优先级 |
+| 反序的意义 | 清理与初始化配对 | 仅影响 after_model 的执行优先级 |
 | 典型例子 | Auth: 校验 token / 清理上下文 | ThreadData: 只创建目录，没有清理 |
 
 ## 关键设计点
 
 ### ClarificationMiddleware 为什么在列表最后？
 
-位置最后使它在工具调用包装链中优先拦截 `ask_clarification`。如果命中，它返回 `Command(goto=END)`，把格式化后的澄清问题写成 `ToolMessage` 并中断执行。
+位置最后 = `after_model` 最先执行。它需要**第一个**看到 model 输出，检查是否有 `ask_clarification` tool call。如果有，立即中断（`Command(goto=END)`），后续 middleware 的 `after_model` 不再执行。
 
 ### SandboxMiddleware 的对称性
 
 `before_agent`（正序第 3 个）获取沙箱，`after_agent`（反序第 1 个）释放沙箱。外层进入 → 外层退出，天然的洋葱对称。
 
-### LoopDetectionMiddleware 为什么同时用多个钩子？
+### 大部分 middleware 只用一个钩子
 
-`after_model` 只做检测：重复工具调用达到 warning 阈值时，把 warning 放入 `(thread_id, run_id)` 作用域的 pending 队列。真正注入发生在下一次 `wrap_model_call`：此时上一轮 `AIMessage(tool_calls)` 对应的 `ToolMessage` 已经在请求里，warning 追加在末尾，不会破坏 OpenAI/Moonshot 的 tool-call pairing。`before_agent` 清理同一 thread 下旧 run 的残留 warning，`after_agent` 清理当前 run 没被消费的 warning。
+14 个 middleware 中，只有 SandboxMiddleware 同时用了 `before_agent` + `after_agent`（获取/释放）。其余都只在一个阶段执行。洋葱模型的反序特性主要影响 `after_model` 阶段的执行顺序。

backend/docs/plan_mode_usage.md

@@ -127,8 +127,8 @@ complex_agent = create_agent_for_task("high")
 ## How It Works
 
 1. When `make_lead_agent(config)` is called, it extracts `is_plan_mode` from `config.configurable`
-2. The config is passed to `build_middlewares(config)`
+2. The config is passed to `_build_middlewares(config)`
-3. `build_middlewares()` reads `is_plan_mode` and calls `_create_todo_list_middleware(is_plan_mode)`
+3. `_build_middlewares()` reads `is_plan_mode` and calls `_create_todo_list_middleware(is_plan_mode)`
 4. If `is_plan_mode=True`, a `TodoListMiddleware` instance is created and added to the middleware chain
 5. The middleware automatically adds a `write_todos` tool to the agent's toolset
 6. The agent can use this tool to manage tasks during execution
@@ -141,7 +141,7 @@ make_lead_agent(config)
   │
   ├─> Extracts: is_plan_mode = config.configurable.get("is_plan_mode", False)
   │
-  └─> build_middlewares(config)
+  └─> _build_middlewares(config)
         │
         ├─> ThreadDataMiddleware
         ├─> SandboxMiddleware
@@ -156,7 +156,7 @@ make_lead_agent(config)
 ### Agent Module
 - **Location**: `packages/harness/deerflow/agents/lead_agent/agent.py`
 - **Function**: `_create_todo_list_middleware(is_plan_mode: bool)` - Creates TodoListMiddleware if plan mode is enabled
-- **Function**: `build_middlewares(config: RunnableConfig)` - Builds middleware chain based on runtime config
+- **Function**: `_build_middlewares(config: RunnableConfig)` - Builds middleware chain based on runtime config
 - **Function**: `make_lead_agent(config: RunnableConfig)` - Creates agent with appropriate middlewares
 
 ### Runtime Configuration

backend/packages/harness/deerflow/agents/factory.py

@@ -173,7 +173,7 @@ def _assemble_from_features(
       9.   MemoryMiddleware (memory feature)
       10.  ViewImageMiddleware (vision feature)
       11.  SubagentLimitMiddleware (subagent feature)
-      12.  LoopDetectionMiddleware (loop_detection feature)
+      12.  LoopDetectionMiddleware (always)
       13.  ClarificationMiddleware (always last)
 
     Two-phase ordering:
@@ -272,15 +272,10 @@ def _assemble_from_features(
 
         extra_tools.append(task_tool)
 
-    # --- [12] LoopDetection ---
+    # --- [12] LoopDetection (always) ---
-    if feat.loop_detection is not False:
+    from deerflow.agents.middlewares.loop_detection_middleware import LoopDetectionMiddleware
-        if isinstance(feat.loop_detection, AgentMiddleware):
-            chain.append(feat.loop_detection)
-        else:
-            from deerflow.agents.middlewares.loop_detection_middleware import LoopDetectionMiddleware
-            from deerflow.config.loop_detection_config import LoopDetectionConfig
 
-            chain.append(LoopDetectionMiddleware.from_config(LoopDetectionConfig()))
+    chain.append(LoopDetectionMiddleware())
 
     # --- [13] Clarification (always last among built-ins) ---
     chain.append(ClarificationMiddleware())

backend/packages/harness/deerflow/agents/features.py

@@ -31,7 +31,6 @@ class RuntimeFeatures:
     vision: bool | AgentMiddleware = False
     auto_title: bool | AgentMiddleware = False
     guardrail: Literal[False] | AgentMiddleware = False
-    loop_detection: bool | AgentMiddleware = True
 
 
 # ---------------------------------------------------------------------------

backend/packages/harness/deerflow/agents/lead_agent/agent.py

@@ -1,25 +1,3 @@
-"""Lead agent factory.
-
-INVARIANT — tracing callback placement
-======================================
-
-Tracing callbacks (Langfuse, LangSmith) are attached at the **graph
-invocation root** in :func:`_make_lead_agent` (see the
-``build_tracing_callbacks()`` block that appends to ``config["callbacks"]``).
-Every ``create_chat_model(...)`` call inside this module — and inside any
-middleware reachable from this graph (e.g. ``TitleMiddleware``) — MUST pass
-``attach_tracing=False``.
-
-Forgetting that flag emits duplicate spans (one rooted at the graph, one at
-the model) AND prevents the Langfuse handler's ``propagate_attributes``
-path from firing, so ``session_id`` / ``user_id`` never reach the trace.
-The four current sites are: bootstrap agent, default agent, summarization
-middleware, and the async path inside ``TitleMiddleware``. Any new in-graph
-``create_chat_model`` call must add to this list and pass the flag.
-"""
-
-from __future__ import annotations
-
 import logging
 
 from langchain.agents import create_agent
@@ -31,7 +9,6 @@ from deerflow.agents.memory.summarization_hook import memory_flush_hook
 from deerflow.agents.middlewares.clarification_middleware import ClarificationMiddleware
 from deerflow.agents.middlewares.loop_detection_middleware import LoopDetectionMiddleware
 from deerflow.agents.middlewares.memory_middleware import MemoryMiddleware
-from deerflow.agents.middlewares.safety_finish_reason_middleware import SafetyFinishReasonMiddleware
 from deerflow.agents.middlewares.subagent_limit_middleware import SubagentLimitMiddleware
 from deerflow.agents.middlewares.summarization_middleware import BeforeSummarizationHook, DeerFlowSummarizationMiddleware
 from deerflow.agents.middlewares.title_middleware import TitleMiddleware
@@ -43,14 +20,9 @@ from deerflow.agents.thread_state import ThreadState
 from deerflow.config.agents_config import load_agent_config, validate_agent_name
 from deerflow.config.app_config import AppConfig, get_app_config
 from deerflow.models import create_chat_model
-from deerflow.skills.tool_policy import filter_tools_by_skill_allowed_tools
-from deerflow.skills.types import Skill
-from deerflow.tracing import build_tracing_callbacks
 
 logger = logging.getLogger(__name__)
 
-_BOOTSTRAP_SKILL_NAMES = {"bootstrap"}
-
 
 def _get_runtime_config(config: RunnableConfig) -> dict:
     """Merge legacy configurable options with LangGraph runtime context."""
@@ -99,14 +71,10 @@ def _create_summarization_middleware(*, app_config: AppConfig | None = None) ->
     # Bind "middleware:summarize" tag so RunJournal identifies these LLM calls
     # as middleware rather than lead_agent (SummarizationMiddleware is a
     # LangChain built-in, so we tag the model at creation time).
-    # attach_tracing=False because the graph-level RunnableConfig (set in
-    # ``_make_lead_agent``) already carries tracing callbacks; binding them
-    # again at the model level would emit duplicate spans and break
-    # ``session_id`` / ``user_id`` propagation.
     if config.model_name:
-        model = create_chat_model(name=config.model_name, thinking_enabled=False, app_config=resolved_app_config, attach_tracing=False)
+        model = create_chat_model(name=config.model_name, thinking_enabled=False, app_config=resolved_app_config)
     else:
-        model = create_chat_model(thinking_enabled=False, app_config=resolved_app_config, attach_tracing=False)
+        model = create_chat_model(thinking_enabled=False, app_config=resolved_app_config)
     model = model.with_config(tags=["middleware:summarize"])
 
     # Prepare kwargs
@@ -267,31 +235,20 @@ Being proactive with task management demonstrates thoroughness and ensures all r
 # ViewImageMiddleware should be before ClarificationMiddleware to inject image details before LLM
 # ToolErrorHandlingMiddleware should be before ClarificationMiddleware to convert tool exceptions to ToolMessages
 # ClarificationMiddleware should be last to intercept clarification requests after model calls
-def build_middlewares(
+def _build_middlewares(
     config: RunnableConfig,
     model_name: str | None,
     agent_name: str | None = None,
     custom_middlewares: list[AgentMiddleware] | None = None,
     *,
-    available_skills: set[str] | None = None,
     app_config: AppConfig | None = None,
-    deferred_setup=None,
 ):
-    """Build the lead-agent middleware chain based on runtime configuration.
+    """Build middleware chain based on runtime configuration.
-
-    Public entry point for the lead agent's full middleware composition. Used by
-    ``make_lead_agent`` and by the embedded ``DeerFlowClient`` (a lead-agent variant
-    that needs the identical chain). Keep this name stable: it is imported across a
-    module boundary, so renames/signature changes ripple into ``client.py``.
 
     Args:
         config: Runtime configuration containing configurable options like is_plan_mode.
-        model_name: Resolved runtime model name; gates vision-only middleware.
         agent_name: If provided, MemoryMiddleware will use per-agent memory storage.
         custom_middlewares: Optional list of custom middlewares to inject into the chain.
-        app_config: Explicit AppConfig; falls back to ``get_app_config()`` when omitted.
-        deferred_setup: Optional deferred-MCP-tool setup that attaches
-            ``DeferredToolFilterMiddleware`` when ``tool_search`` is enabled.
 
     Returns:
         List of middleware instances.
@@ -299,19 +256,6 @@ def build_middlewares(
     resolved_app_config = app_config or get_app_config()
     middlewares = build_lead_runtime_middlewares(app_config=resolved_app_config, lazy_init=True)
 
-    # Always inject current date (and optionally memory) as <system-reminder> into the
-    # first HumanMessage to keep the system prompt fully static for prefix-cache reuse.
-    from deerflow.agents.middlewares.dynamic_context_middleware import DynamicContextMiddleware
-
-    middlewares.append(DynamicContextMiddleware(agent_name=agent_name, app_config=resolved_app_config))
-
-    # Deterministically load a full SKILL.md when the user starts the turn with
-    # /skill-name. This keeps the base system prompt metadata-only while giving
-    # explicit user activation priority over model-side relevance guessing.
-    from deerflow.agents.middlewares.skill_activation_middleware import SkillActivationMiddleware
-
-    middlewares.append(SkillActivationMiddleware(available_skills=available_skills, app_config=resolved_app_config))
-
     # Add summarization middleware if enabled
     summarization_middleware = _create_summarization_middleware(app_config=resolved_app_config)
     if summarization_middleware is not None:
@@ -340,13 +284,11 @@ def build_middlewares(
     if model_config is not None and model_config.supports_vision:
         middlewares.append(ViewImageMiddleware())
 
-    # Hide deferred tool schemas from model binding until tool_search promotes them.
+    # Add DeferredToolFilterMiddleware to hide deferred tool schemas from model binding
-    # The deferred set + catalog hash come from the build-time setup (assembled
+    if resolved_app_config.tool_search.enabled:
-    # after tool-policy filtering); promotion is read from graph state.
-    if deferred_setup is not None and deferred_setup.deferred_names:
         from deerflow.agents.middlewares.deferred_tool_filter_middleware import DeferredToolFilterMiddleware
 
-        middlewares.append(DeferredToolFilterMiddleware(deferred_setup.deferred_names, deferred_setup.catalog_hash))
+        middlewares.append(DeferredToolFilterMiddleware())
 
     # Add SubagentLimitMiddleware to truncate excess parallel task calls
     subagent_enabled = cfg.get("subagent_enabled", False)
@@ -355,50 +297,17 @@ def build_middlewares(
         middlewares.append(SubagentLimitMiddleware(max_concurrent=max_concurrent_subagents))
 
     # LoopDetectionMiddleware — detect and break repetitive tool call loops
-    loop_detection_config = resolved_app_config.loop_detection
+    middlewares.append(LoopDetectionMiddleware())
-    if loop_detection_config.enabled:
-        middlewares.append(LoopDetectionMiddleware.from_config(loop_detection_config))
 
     # Inject custom middlewares before ClarificationMiddleware
     if custom_middlewares:
         middlewares.extend(custom_middlewares)
 
-    # SafetyFinishReasonMiddleware — suppress tool execution when the provider
-    # safety-terminated the response. Registered after custom middlewares so
-    # that LangChain's reverse-order after_model dispatch runs Safety first;
-    # cleared tool_calls then flow through Loop/Subagent accounting without
-    # firing extra alarms. See safety_finish_reason_middleware.py docstring.
-    safety_config = resolved_app_config.safety_finish_reason
-    if safety_config.enabled:
-        middlewares.append(SafetyFinishReasonMiddleware.from_config(safety_config))
-
     # ClarificationMiddleware should always be last
     middlewares.append(ClarificationMiddleware())
     return middlewares
 
 
-def _available_skill_names(agent_config, is_bootstrap: bool) -> set[str] | None:
-    if is_bootstrap:
-        return set(_BOOTSTRAP_SKILL_NAMES)
-    if agent_config and agent_config.skills is not None:
-        return set(agent_config.skills)
-    return None
-
-
-def _load_enabled_skills_for_tool_policy(available_skills: set[str] | None, *, app_config: AppConfig) -> list[Skill]:
-    try:
-        from deerflow.agents.lead_agent.prompt import get_enabled_skills_for_config
-
-        skills = get_enabled_skills_for_config(app_config)
-    except Exception:
-        logger.exception("Failed to load skills for allowed-tools policy")
-        raise
-
-    if available_skills is None:
-        return skills
-    return [skill for skill in skills if skill.name in available_skills]
-
-
 def make_lead_agent(config: RunnableConfig):
     """LangGraph graph factory; keep the signature compatible with LangGraph Server."""
     runtime_config = _get_runtime_config(config)
@@ -409,8 +318,7 @@ def make_lead_agent(config: RunnableConfig):
 def _make_lead_agent(config: RunnableConfig, *, app_config: AppConfig):
     # Lazy import to avoid circular dependency
     from deerflow.tools import get_available_tools
-    from deerflow.tools.builtins import setup_agent, update_agent
+    from deerflow.tools.builtins import setup_agent
-    from deerflow.tools.builtins.tool_search import assemble_deferred_tools
 
     cfg = _get_runtime_config(config)
     resolved_app_config = app_config
@@ -425,7 +333,6 @@ def _make_lead_agent(config: RunnableConfig, *, app_config: AppConfig):
     agent_name = validate_agent_name(cfg.get("agent_name"))
 
     agent_config = load_agent_config(agent_name) if not is_bootstrap else None
-    available_skills = _available_skill_names(agent_config, is_bootstrap)
     # Custom agent model from agent config (if any), or None to let _resolve_model_name pick the default
     agent_model_name = agent_config.model if agent_config and agent_config.model else None
 
@@ -464,77 +371,41 @@ def _make_lead_agent(config: RunnableConfig, *, app_config: AppConfig):
             "is_plan_mode": is_plan_mode,
             "subagent_enabled": subagent_enabled,
             "tool_groups": agent_config.tool_groups if agent_config else None,
-            "available_skills": sorted(available_skills) if available_skills is not None else None,
+            "available_skills": ["bootstrap"] if is_bootstrap else (agent_config.skills if agent_config and agent_config.skills is not None else None),
         }
     )
 
-    # Inject tracing callbacks at the graph invocation root so a single LangGraph
-    # run produces one trace with all node / LLM / tool calls as child spans,
-    # AND so the Langfuse handler sees ``on_chain_start(parent_run_id=None)`` and
-    # actually propagates ``langfuse_session_id`` / ``langfuse_user_id`` from
-    # ``config["metadata"]`` onto the trace. Without root-level attachment the
-    # model is a nested observation and the handler strips ``langfuse_*`` keys.
-    tracing_callbacks = build_tracing_callbacks()
-    if tracing_callbacks:
-        existing = config.get("callbacks") or []
-        if not isinstance(existing, list):
-            existing = list(existing)
-        config["callbacks"] = [*existing, *tracing_callbacks]
-
-    skills_for_tool_policy = _load_enabled_skills_for_tool_policy(available_skills, app_config=resolved_app_config)
-
     if is_bootstrap:
         # Special bootstrap agent with minimal prompt for initial custom agent creation flow
-        # Keep the bootstrap skill set intentionally narrow so agent creation
-        # remains deterministic before the custom agent's own config exists.
-        raw_tools = get_available_tools(model_name=model_name, subagent_enabled=subagent_enabled, app_config=resolved_app_config) + [setup_agent]
-        filtered = filter_tools_by_skill_allowed_tools(raw_tools, skills_for_tool_policy)
-        final_tools, setup = assemble_deferred_tools(filtered, enabled=resolved_app_config.tool_search.enabled)
         return create_agent(
-            model=create_chat_model(name=model_name, thinking_enabled=thinking_enabled, app_config=resolved_app_config, attach_tracing=False),
+            model=create_chat_model(name=model_name, thinking_enabled=thinking_enabled, app_config=resolved_app_config),
-            tools=final_tools,
+            tools=get_available_tools(model_name=model_name, subagent_enabled=subagent_enabled, app_config=resolved_app_config) + [setup_agent],
-            middleware=build_middlewares(
+            middleware=_build_middlewares(config, model_name=model_name, app_config=resolved_app_config),
-                config,
-                model_name=model_name,
-                available_skills=set(_BOOTSTRAP_SKILL_NAMES),
-                app_config=resolved_app_config,
-                deferred_setup=setup,
-            ),
             system_prompt=apply_prompt_template(
                 subagent_enabled=subagent_enabled,
                 max_concurrent_subagents=max_concurrent_subagents,
-                available_skills=set(_BOOTSTRAP_SKILL_NAMES),
+                available_skills=set(["bootstrap"]),
                 app_config=resolved_app_config,
-                deferred_names=setup.deferred_names,
             ),
             state_schema=ThreadState,
         )
 
-    # Custom agents can update their own SOUL.md / config via update_agent.
-    # The default agent (no agent_name) does not see this tool.
-    extra_tools = [update_agent] if agent_name else []
     # Default lead agent (unchanged behavior)
-    raw_tools = get_available_tools(model_name=model_name, groups=agent_config.tool_groups if agent_config else None, subagent_enabled=subagent_enabled, app_config=resolved_app_config)
-    filtered = filter_tools_by_skill_allowed_tools(raw_tools + extra_tools, skills_for_tool_policy)
-    final_tools, setup = assemble_deferred_tools(filtered, enabled=resolved_app_config.tool_search.enabled)
     return create_agent(
-        model=create_chat_model(name=model_name, thinking_enabled=thinking_enabled, reasoning_effort=reasoning_effort, app_config=resolved_app_config, attach_tracing=False),
+        model=create_chat_model(name=model_name, thinking_enabled=thinking_enabled, reasoning_effort=reasoning_effort, app_config=resolved_app_config),
-        tools=final_tools,
+        tools=get_available_tools(
-        middleware=build_middlewares(
-            config,
             model_name=model_name,
-            agent_name=agent_name,
+            groups=agent_config.tool_groups if agent_config else None,
-            available_skills=available_skills,
+            subagent_enabled=subagent_enabled,
             app_config=resolved_app_config,
-            deferred_setup=setup,
         ),
+        middleware=_build_middlewares(config, model_name=model_name, agent_name=agent_name, app_config=resolved_app_config),
         system_prompt=apply_prompt_template(
             subagent_enabled=subagent_enabled,
             max_concurrent_subagents=max_concurrent_subagents,
             agent_name=agent_name,
-            available_skills=available_skills,
+            available_skills=set(agent_config.skills) if agent_config and agent_config.skills is not None else None,
             app_config=resolved_app_config,
-            deferred_names=setup.deferred_names,
         ),
         state_schema=ThreadState,
     )

backend/packages/harness/deerflow/agents/lead_agent/prompt.py

@@ -3,6 +3,7 @@ from __future__ import annotations
 import asyncio
 import logging
 import threading
+from datetime import datetime
 from functools import lru_cache
 from typing import TYPE_CHECKING
 
@@ -10,7 +11,6 @@ from deerflow.config.agents_config import load_agent_soul
 from deerflow.skills.storage import get_or_new_skill_storage
 from deerflow.skills.types import Skill, SkillCategory
 from deerflow.subagents import get_available_subagent_names
-from deerflow.tools.builtins.tool_search import get_deferred_tools_prompt_section
 
 if TYPE_CHECKING:
     from deerflow.config.app_config import AppConfig
@@ -20,7 +20,6 @@ logger = logging.getLogger(__name__)
 _ENABLED_SKILLS_REFRESH_WAIT_TIMEOUT_SECONDS = 5.0
 _enabled_skills_lock = threading.Lock()
 _enabled_skills_cache: list[Skill] | None = None
-_enabled_skills_by_config_cache: dict[int, tuple[object, list[Skill]]] = {}
 _enabled_skills_refresh_active = False
 _enabled_skills_refresh_version = 0
 _enabled_skills_refresh_event = threading.Event()
@@ -85,7 +84,6 @@ def _invalidate_enabled_skills_cache() -> threading.Event:
     _get_cached_skills_prompt_section.cache_clear()
     with _enabled_skills_lock:
         _enabled_skills_cache = None
-        _enabled_skills_by_config_cache.clear()
         _enabled_skills_refresh_version += 1
         _enabled_skills_refresh_event.clear()
         if _enabled_skills_refresh_active:
@@ -109,15 +107,6 @@ def warm_enabled_skills_cache(timeout_seconds: float = _ENABLED_SKILLS_REFRESH_W
 
 
 def _get_enabled_skills():
-    return get_cached_enabled_skills()
-
-
-def get_cached_enabled_skills() -> list[Skill]:
-    """Return the cached enabled-skills list, kicking off a background refresh on miss.
-
-    Safe to call from request paths: never blocks on disk I/O. Returns an empty
-    list on cache miss; the next call will see the warmed result.
-    """
     with _enabled_skills_lock:
         cached = _enabled_skills_cache
 
@@ -128,29 +117,17 @@ def get_cached_enabled_skills() -> list[Skill]:
     return []
 
 
-def get_enabled_skills_for_config(app_config: AppConfig | None = None) -> list[Skill]:
+def _get_enabled_skills_for_config(app_config: AppConfig | None = None) -> list[Skill]:
     """Return enabled skills using the caller's config source.
 
-    When a concrete ``app_config`` is supplied, cache the loaded skills by that
+    When a concrete ``app_config`` is supplied, bypass the global enabled-skills
-    config object's identity so request-scoped config injection still resolves
+    cache so the skill list and skill paths are resolved from the same config
-    skill paths from the matching config without rescanning storage on every
+    object. This keeps request-scoped config injection consistent even while the
-    agent factory call.
+    release branch still supports global fallback paths.
     """
     if app_config is None:
         return _get_enabled_skills()
-
+    return list(get_or_new_skill_storage(app_config=app_config).load_skills(enabled_only=True))
-    cache_key = id(app_config)
-    with _enabled_skills_lock:
-        cached = _enabled_skills_by_config_cache.get(cache_key)
-        if cached is not None:
-            cached_config, cached_skills = cached
-            if cached_config is app_config:
-                return list(cached_skills)
-
-    skills = list(get_or_new_skill_storage(app_config=app_config).load_skills(enabled_only=True))
-    with _enabled_skills_lock:
-        _enabled_skills_by_config_cache[cache_key] = (app_config, skills)
-    return list(skills)
 
 
 def _skill_mutability_label(category: SkillCategory | str) -> str:
@@ -367,7 +344,8 @@ You are {agent_name}, an open-source super agent.
 </role>
 
 {soul}
-{self_update_section}
+{memory_context}
+
 <thinking_style>
 - Think concisely and strategically about the user's request BEFORE taking action
 - Break down the task: What is clear? What is ambiguous? What is missing?
@@ -543,14 +521,6 @@ combined with a FastAPI gateway for REST API access [citation:FastAPI](https://f
 {subagent_reminder}- Skill First: Always load the relevant skill before starting **complex** tasks.
 - Progressive Loading: Load resources incrementally as referenced in skills
 - Output Files: Final deliverables must be in `/mnt/user-data/outputs`
-- File Editing Workflow: When revising an existing file, prefer
-  `str_replace` over `write_file` — it sends only the diff and avoids
-  re-emitting the whole file (mirrors Claude Code's Edit and Codex's
-  apply_patch). When writing long new content from scratch, split it
-  into sections: the first `write_file` call creates the file, then use
-  `write_file` with append=True to extend it section by section. This
-  keeps each tool call small and avoids mid-stream chunk-gap timeouts
-  on oversized single-shot writes. (See issue #3189.)  
 - Clarity: Be direct and helpful, avoid unnecessary meta-commentary
 - Including Images and Mermaid: Images and Mermaid diagrams are always welcomed in the Markdown format, and you're encouraged to use `![Image Description](image_path)\n\n` or "```mermaid" to display images in response or Markdown files
 - Multi-task: Better utilize parallel tool calling to call multiple tools at one time for better performance
@@ -625,11 +595,6 @@ You have access to skills that provide optimized workflows for specific tasks. E
 4. Load referenced resources only when needed during execution
 5. Follow the skill's instructions precisely
 
-**Explicit Slash Skill Activation:**
-- If the user starts a request with `/<skill-name>`, that skill was explicitly requested for the current turn.
-- Follow the activated skill before choosing a general workflow.
-- The runtime injects the activated skill content for explicit slash activations; do not call `read_file` for that SKILL.md again unless the injected skill references supporting resources you need.
-
 **Skills are located at:** {container_base_path}
 {skill_evolution_section}
 {skills_list}
@@ -639,7 +604,7 @@ You have access to skills that provide optimized workflows for specific tasks. E
 
 def get_skills_prompt_section(available_skills: set[str] | None = None, *, app_config: AppConfig | None = None) -> str:
     """Generate the skills prompt section with available skills list."""
-    skills = get_enabled_skills_for_config(app_config)
+    skills = _get_enabled_skills_for_config(app_config)
 
     if app_config is None:
         try:
@@ -678,25 +643,34 @@ def get_agent_soul(agent_name: str | None) -> str:
     return ""
 
 
-def _build_self_update_section(agent_name: str | None) -> str:
+def get_deferred_tools_prompt_section(*, app_config: AppConfig | None = None) -> str:
-    """Prompt block that teaches the custom agent to persist self-updates via update_agent."""
+    """Generate <available-deferred-tools> block for the system prompt.
-    if not agent_name:
+
+    Lists only deferred tool names so the agent knows what exists
+    and can use tool_search to load them.
+    Returns empty string when tool_search is disabled or no tools are deferred.
+    """
+    from deerflow.tools.builtins.tool_search import get_deferred_registry
+
+    if app_config is None:
+        try:
+            from deerflow.config import get_app_config
+
+            config = get_app_config()
+        except Exception:
+            return ""
+    else:
+        config = app_config
+
+    if not config.tool_search.enabled:
         return ""
-    return f"""<self_update>
-You are running as the custom agent **{agent_name}** with a persisted SOUL.md and config.yaml.
 
-When the user asks you to update your own description, personality, behaviour, skill set, tool groups, or default model,
+    registry = get_deferred_registry()
-you MUST persist the change with the `update_agent` tool. Do NOT use `bash`, `write_file`, or any sandbox tool to edit
+    if not registry:
-SOUL.md or config.yaml — those write into a temporary sandbox/tool workspace and the changes will be lost on the next turn.
+        return ""
 
-Rules:
+    names = "\n".join(e.name for e in registry.entries)
-- Always pass the FULL replacement text for `soul` (no patch semantics). Start from your current SOUL above and apply the user's edits.
+    return f"<available-deferred-tools>\n{names}\n</available-deferred-tools>"
-- Only pass the fields that should change. Omit the others to preserve them.
-- Never pass literal strings like `"null"`, `"none"`, or `"undefined"` for unchanged fields.
-- Pass `skills=[]` to disable all skills, or omit `skills` to keep the existing whitelist.
-- After `update_agent` returns successfully, tell the user the change is persisted and will take effect on the next turn.
-</self_update>
-"""
 
 
 def _build_acp_section(*, app_config: AppConfig | None = None) -> str:
@@ -757,8 +731,10 @@ def apply_prompt_template(
     agent_name: str | None = None,
     available_skills: set[str] | None = None,
     app_config: AppConfig | None = None,
-    deferred_names: frozenset[str] = frozenset(),
 ) -> str:
+    # Get memory context
+    memory_context = _get_memory_context(agent_name, app_config=app_config)
+
     # Include subagent section only if enabled (from runtime parameter)
     n = max_concurrent_subagents
     subagent_section = _build_subagent_section(n, app_config=app_config) if subagent_enabled else ""
@@ -785,25 +761,24 @@ def apply_prompt_template(
     skills_section = get_skills_prompt_section(available_skills, app_config=app_config)
 
     # Get deferred tools section (tool_search)
-    deferred_tools_section = get_deferred_tools_prompt_section(deferred_names=deferred_names)
+    deferred_tools_section = get_deferred_tools_prompt_section(app_config=app_config)
 
     # Build ACP agent section only if ACP agents are configured
     acp_section = _build_acp_section(app_config=app_config)
     custom_mounts_section = _build_custom_mounts_section(app_config=app_config)
     acp_and_mounts_section = "\n".join(section for section in (acp_section, custom_mounts_section) if section)
 
-    # Build and return the fully static system prompt.
+    # Format the prompt with dynamic skills and memory
-    # Memory and current date are injected per-turn via DynamicContextMiddleware
+    prompt = SYSTEM_PROMPT_TEMPLATE.format(
-    # as a <system-reminder> in the first HumanMessage, keeping this prompt
-    # identical across users and sessions for maximum prefix-cache reuse.
-    return SYSTEM_PROMPT_TEMPLATE.format(
         agent_name=agent_name or "DeerFlow 2.0",
         soul=get_agent_soul(agent_name),
-        self_update_section=_build_self_update_section(agent_name),
         skills_section=skills_section,
         deferred_tools_section=deferred_tools_section,
+        memory_context=memory_context,
         subagent_section=subagent_section,
         subagent_reminder=subagent_reminder,
         subagent_thinking=subagent_thinking,
         acp_section=acp_and_mounts_section,
     )
+
+    return prompt + f"\n<current_date>{datetime.now().strftime('%Y-%m-%d, %A')}</current_date>"

backend/packages/harness/deerflow/agents/memory/prompt.py

@@ -1,14 +1,9 @@
 """Prompt templates for memory update and injection."""
 
-from __future__ import annotations
-
-import logging
 import math
 import re
 from typing import Any
 
-logger = logging.getLogger(__name__)
-
 try:
     import tiktoken
 
@@ -165,39 +160,6 @@ Rules:
 Return ONLY valid JSON."""
 
 
-# Module-level tiktoken encoding cache.  Populated lazily on first use;
-# subsequent calls are a dict lookup (no network I/O).  Pre-warming at
-# startup via :func:`warm_tiktoken_cache` avoids blocking a request on the
-# (potentially slow) first ``get_encoding`` call.
-_tiktoken_encoding_cache: dict[str, tiktoken.Encoding] = {}
-
-
-def _get_tiktoken_encoding(encoding_name: str = "cl100k_base") -> tiktoken.Encoding | None:
-    """Return a cached tiktoken encoding, or ``None`` on failure / unavailability.
-
-    On the very first call for a given *encoding_name*, tiktoken may need to
-    download the BPE data from ``openaipublic.blob.core.windows.net``.  In
-    network-restricted environments (e.g. deployments behind the GFW) this
-    download can block for tens of minutes before the OS TCP timeout kicks in.
-    The caller must therefore be prepared for this to block and should run it
-    off the event loop (e.g. via ``asyncio.to_thread``).
-    """
-    if not TIKTOKEN_AVAILABLE:
-        return None
-
-    cached = _tiktoken_encoding_cache.get(encoding_name)
-    if cached is not None:
-        return cached
-
-    try:
-        encoding = tiktoken.get_encoding(encoding_name)
-        _tiktoken_encoding_cache[encoding_name] = encoding
-        return encoding
-    except Exception:
-        logger.warning("Failed to load tiktoken encoding %r; falling back to char-based estimation", encoding_name, exc_info=True)
-        return None
-
-
 def _count_tokens(text: str, encoding_name: str = "cl100k_base") -> int:
     """Count tokens in text using tiktoken.
 
@@ -208,30 +170,18 @@ def _count_tokens(text: str, encoding_name: str = "cl100k_base") -> int:
     Returns:
         The number of tokens in the text.
     """
-    encoding = _get_tiktoken_encoding(encoding_name)
+    if not TIKTOKEN_AVAILABLE:
-    if encoding is None:
         # Fallback to character-based estimation if tiktoken is not available
-        # or the encoding failed to load.
         return len(text) // 4
 
     try:
+        encoding = tiktoken.get_encoding(encoding_name)
         return len(encoding.encode(text))
     except Exception:
         # Fallback to character-based estimation on error
         return len(text) // 4
 
 
-def warm_tiktoken_cache() -> bool:
-    """Pre-warm the tiktoken encoding cache.
-
-    Call at startup (off the event loop) so the first request never blocks
-    on the BPE download.  Returns ``True`` if the encoding was loaded
-    successfully (or was already cached), ``False`` if tiktoken is
-    unavailable or the download failed.
-    """
-    return _get_tiktoken_encoding("cl100k_base") is not None
-
-
 def _coerce_confidence(value: Any, default: float = 0.0) -> float:
     """Coerce a confidence-like value to a bounded float in [0, 1].
 

backend/packages/harness/deerflow/agents/memory/queue.py

@@ -40,15 +40,6 @@ class MemoryUpdateQueue:
         self._timer: threading.Timer | None = None
         self._processing = False
 
-    @staticmethod
-    def _queue_key(
-        thread_id: str,
-        user_id: str | None,
-        agent_name: str | None,
-    ) -> tuple[str, str | None, str | None]:
-        """Return the debounce identity for a memory update target."""
-        return (thread_id, user_id, agent_name)
-
     def add(
         self,
         thread_id: str,
@@ -124,9 +115,8 @@ class MemoryUpdateQueue:
         correction_detected: bool,
         reinforcement_detected: bool,
     ) -> None:
-        queue_key = self._queue_key(thread_id, user_id, agent_name)
         existing_context = next(
-            (context for context in self._queue if self._queue_key(context.thread_id, context.user_id, context.agent_name) == queue_key),
+            (context for context in self._queue if context.thread_id == thread_id),
             None,
         )
         merged_correction_detected = correction_detected or (existing_context.correction_detected if existing_context is not None else False)
@@ -140,7 +130,7 @@ class MemoryUpdateQueue:
             reinforcement_detected=merged_reinforcement_detected,
         )
 
-        self._queue = [context for context in self._queue if self._queue_key(context.thread_id, context.user_id, context.agent_name) != queue_key]
+        self._queue = [c for c in self._queue if c.thread_id != thread_id]
         self._queue.append(context)
 
     def _reset_timer(self) -> None:

backend/packages/harness/deerflow/agents/memory/summarization_hook.py

@@ -6,7 +6,6 @@ from deerflow.agents.memory.message_processing import detect_correction, detect_
 from deerflow.agents.memory.queue import get_memory_queue
 from deerflow.agents.middlewares.summarization_middleware import SummarizationEvent
 from deerflow.config.memory_config import get_memory_config
-from deerflow.runtime.user_context import resolve_runtime_user_id
 
 
 def memory_flush_hook(event: SummarizationEvent) -> None:
@@ -22,13 +21,11 @@ def memory_flush_hook(event: SummarizationEvent) -> None:
 
     correction_detected = detect_correction(filtered_messages)
     reinforcement_detected = not correction_detected and detect_reinforcement(filtered_messages)
-    user_id = resolve_runtime_user_id(event.runtime)
     queue = get_memory_queue()
     queue.add_nowait(
         thread_id=event.thread_id,
         messages=filtered_messages,
         agent_name=event.agent_name,
-        user_id=user_id,
         correction_detected=correction_detected,
         reinforcement_detected=reinforcement_detected,
     )

backend/packages/harness/deerflow/agents/memory/updater.py

@@ -227,110 +227,6 @@ def _extract_text(content: Any) -> str:
     return str(content)
 
 
-_REQUIRED_MEMORY_UPDATE_TOP_LEVEL_KEYS = frozenset({"user", "history", "newFacts", "factsToRemove"})
-
-
-def _normalize_memory_update_fact(fact: Any) -> dict[str, Any] | None:
-    """Normalize a single fact entry from a model-produced memory update."""
-    if not isinstance(fact, dict):
-        return None
-
-    raw_content = fact.get("content")
-    if not isinstance(raw_content, str):
-        return None
-    content = raw_content.strip()
-    if not content:
-        return None
-
-    raw_category = fact.get("category")
-    category = raw_category.strip() if isinstance(raw_category, str) and raw_category.strip() else "context"
-
-    raw_confidence = fact.get("confidence", 0.5)
-    if isinstance(raw_confidence, bool):
-        return None
-    if isinstance(raw_confidence, str):
-        raw_confidence = raw_confidence.strip()
-        if not raw_confidence:
-            return None
-        try:
-            raw_confidence = float(raw_confidence)
-        except ValueError:
-            return None
-    elif isinstance(raw_confidence, (int, float)):
-        raw_confidence = float(raw_confidence)
-    else:
-        return None
-
-    if not math.isfinite(raw_confidence):
-        return None
-
-    normalized_fact = {
-        "content": content,
-        "category": category,
-        "confidence": raw_confidence,
-    }
-    source_error = fact.get("sourceError")
-    if isinstance(source_error, str):
-        normalized_source_error = source_error.strip()
-        if normalized_source_error:
-            normalized_fact["sourceError"] = normalized_source_error
-
-    return normalized_fact
-
-
-def _normalize_memory_update_data(update_data: dict[str, Any]) -> dict[str, Any]:
-    """Coerce parsed memory update data into the shape consumed by _apply_updates."""
-    user = update_data.get("user")
-    history = update_data.get("history")
-    new_facts = update_data.get("newFacts")
-    facts_to_remove = update_data.get("factsToRemove")
-    normalized_facts_to_remove = [fact_id for fact_id in facts_to_remove if isinstance(fact_id, str)] if isinstance(facts_to_remove, list) else []
-    normalized_new_facts = []
-    dropped_new_fact = not isinstance(new_facts, list)
-    if isinstance(new_facts, list):
-        for fact in new_facts:
-            normalized_fact = _normalize_memory_update_fact(fact)
-            if normalized_fact is not None:
-                normalized_new_facts.append(normalized_fact)
-            else:
-                dropped_new_fact = True
-
-    if normalized_facts_to_remove and dropped_new_fact:
-        raise json.JSONDecodeError(
-            "Unsafe partial memory update: factsToRemove with malformed newFacts",
-            json.dumps(update_data, ensure_ascii=False),
-            0,
-        )
-
-    return {
-        "user": user if isinstance(user, dict) else {},
-        "history": history if isinstance(history, dict) else {},
-        "newFacts": normalized_new_facts,
-        "factsToRemove": normalized_facts_to_remove,
-    }
-
-
-def _parse_memory_update_response(response_content: Any) -> dict[str, Any]:
-    """Parse the first valid memory-update JSON object from an LLM response.
-
-    Some providers may wrap JSON in thinking traces, prose, or markdown fences
-    even when prompted to return JSON only. This parser accepts safely
-    extractable JSON objects but does not repair truncated or malformed JSON.
-    """
-    response_text = _extract_text(response_content).strip()
-    decoder = json.JSONDecoder()
-
-    for match in re.finditer(r"\{", response_text):
-        try:
-            parsed, _end = decoder.raw_decode(response_text[match.start() :])
-        except json.JSONDecodeError:
-            continue
-        if isinstance(parsed, dict) and _REQUIRED_MEMORY_UPDATE_TOP_LEVEL_KEYS.issubset(parsed):
-            return _normalize_memory_update_data(parsed)
-
-    raise json.JSONDecodeError("No valid memory update JSON object found", response_text, 0)
-
-
 # Matches sentences that describe a file-upload *event* rather than general
 # file-related work.  Deliberately narrow to avoid removing legitimate facts
 # such as "User works with CSV files" or "prefers PDF export".
@@ -442,7 +338,7 @@ class MemoryUpdater:
             reinforcement_detected=reinforcement_detected,
         )
         prompt = MEMORY_UPDATE_PROMPT.format(
-            current_memory=json.dumps(current_memory, indent=2, ensure_ascii=False),
+            current_memory=json.dumps(current_memory, indent=2),
             conversation=conversation_text,
             correction_hint=correction_hint,
         )
@@ -457,7 +353,13 @@ class MemoryUpdater:
         user_id: str | None = None,
     ) -> bool:
         """Parse the model response, apply updates, and persist memory."""
-        update_data = _parse_memory_update_response(response_content)
+        response_text = _extract_text(response_content).strip()
+
+        if response_text.startswith("```"):
+            lines = response_text.split("\n")
+            response_text = "\n".join(lines[1:-1] if lines[-1] == "```" else lines[1:])
+
+        update_data = json.loads(response_text)
         # Deep-copy before in-place mutation so a subsequent save() failure
         # cannot corrupt the still-cached original object reference.
         updated_memory = self._apply_updates(copy.deepcopy(current_memory), update_data, thread_id)

backend/packages/harness/deerflow/agents/middlewares/dangling_tool_call_middleware.py

@@ -15,7 +15,6 @@ to the end of the message list as before_model + add_messages reducer would do.
 
 import json
 import logging
-from collections import defaultdict, deque
 from collections.abc import Awaitable, Callable
 from typing import override
 
@@ -26,11 +25,6 @@ from langchain_core.messages import ToolMessage
 
 logger = logging.getLogger(__name__)
 
-# Workaround for issue #2894: malformed write_file calls can carry huge Markdown
-# payloads in invalid tool-call args. Keep recovery error details short so the
-# synthetic ToolMessage does not echo large or malformed content back to the model.
-_MAX_RECOVERY_ERROR_DETAIL_LEN = 500
-
 
 class DanglingToolCallMiddleware(AgentMiddleware[AgentState]):
     """Inserts placeholder ToolMessages for dangling tool calls before model invocation.
@@ -42,144 +36,94 @@ class DanglingToolCallMiddleware(AgentMiddleware[AgentState]):
 
     @staticmethod
     def _message_tool_calls(msg) -> list[dict]:
-        """Return normalized tool calls from structured fields or raw provider payloads.
+        """Return normalized tool calls from structured fields or raw provider payloads."""
-
-        LangChain stores malformed provider function calls in ``invalid_tool_calls``.
-        They do not execute, but provider adapters may still serialize enough of
-        the call id/name back into the next request that strict OpenAI-compatible
-        validators expect a matching ToolMessage. Treat them as dangling calls so
-        the next model request stays well-formed and the model sees a recoverable
-        tool error instead of another provider 400.
-        """
-        normalized: list[dict] = []
-
         tool_calls = getattr(msg, "tool_calls", None) or []
-        normalized.extend(list(tool_calls))
+        if tool_calls:
+            return list(tool_calls)
 
         raw_tool_calls = (getattr(msg, "additional_kwargs", None) or {}).get("tool_calls") or []
-        if not tool_calls:
+        normalized: list[dict] = []
-            for raw_tc in raw_tool_calls:
+        for raw_tc in raw_tool_calls:
-                if not isinstance(raw_tc, dict):
+            if not isinstance(raw_tc, dict):
-                    continue
-
-                function = raw_tc.get("function")
-                name = raw_tc.get("name")
-                if not name and isinstance(function, dict):
-                    name = function.get("name")
-
-                args = raw_tc.get("args", {})
-                if not args and isinstance(function, dict):
-                    raw_args = function.get("arguments")
-                    if isinstance(raw_args, str):
-                        try:
-                            parsed_args = json.loads(raw_args)
-                        except (TypeError, ValueError, json.JSONDecodeError):
-                            parsed_args = {}
-                        args = parsed_args if isinstance(parsed_args, dict) else {}
-
-                normalized.append(
-                    {
-                        "id": raw_tc.get("id"),
-                        "name": name or "unknown",
-                        "args": args if isinstance(args, dict) else {},
-                    }
-                )
-
-        for invalid_tc in getattr(msg, "invalid_tool_calls", None) or []:
-            if not isinstance(invalid_tc, dict):
                 continue
+
+            function = raw_tc.get("function")
+            name = raw_tc.get("name")
+            if not name and isinstance(function, dict):
+                name = function.get("name")
+
+            args = raw_tc.get("args", {})
+            if not args and isinstance(function, dict):
+                raw_args = function.get("arguments")
+                if isinstance(raw_args, str):
+                    try:
+                        parsed_args = json.loads(raw_args)
+                    except (TypeError, ValueError, json.JSONDecodeError):
+                        parsed_args = {}
+                    args = parsed_args if isinstance(parsed_args, dict) else {}
+
             normalized.append(
                 {
-                    "id": invalid_tc.get("id"),
+                    "id": raw_tc.get("id"),
-                    "name": invalid_tc.get("name") or "unknown",
+                    "name": name or "unknown",
-                    "args": {},
+                    "args": args if isinstance(args, dict) else {},
-                    "invalid": True,
-                    "error": invalid_tc.get("error"),
                 }
             )
 
         return normalized
 
-    @staticmethod
-    def _synthetic_tool_message_content(tool_call: dict) -> str:
-        if tool_call.get("invalid"):
-            name = tool_call.get("name")
-            error = tool_call.get("error")
-            error_text = error[:_MAX_RECOVERY_ERROR_DETAIL_LEN] if isinstance(error, str) and error else ""
-            # Workaround for issue #2894: malformed write_file calls can carry huge Markdown
-            # payloads in invalid tool-call args. Keep recovery guidance actionable without
-            # echoing large or malformed content back to the model.
-            if name == "write_file":
-                details = f" Parser error: {error_text}" if error_text else ""
-                return (
-                    "[write_file failed before execution: the tool-call arguments were not valid JSON, "
-                    "so no file was written. This often happens when the model tries to write a very "
-                    "large Markdown file in a single tool call, especially when `content` contains "
-                    "unescaped quotes, inline JSON, backslashes, or code fences. Do not retry the same "
-                    "large `write_file` payload for this artifact; provide the report/content directly "
-                    "as normal assistant text in your next response. If a file write is still needed "
-                    f"later, split the file into smaller sections instead of one large payload.{details}]"
-                )
-            if error_text:
-                return f"[Tool call could not be executed because its arguments were invalid: {error_text}]"
-            return "[Tool call could not be executed because its arguments were invalid.]"
-        return "[Tool call was interrupted and did not return a result.]"
-
     def _build_patched_messages(self, messages: list) -> list | None:
-        """Return messages with tool results grouped after their tool-call AIMessage.
+        """Return a new message list with patches inserted at the correct positions.
 
-        This normalizes model-bound causal order before provider serialization while
+        For each AIMessage with dangling tool_calls (no corresponding ToolMessage),
-        preserving already-valid transcripts unchanged.
+        a synthetic ToolMessage is inserted immediately after that AIMessage.
+        Returns None if no patches are needed.
         """
-        tool_messages_by_id: dict[str, deque[ToolMessage]] = defaultdict(deque)
+        # Collect IDs of all existing ToolMessages
+        existing_tool_msg_ids: set[str] = set()
         for msg in messages:
             if isinstance(msg, ToolMessage):
-                tool_messages_by_id[msg.tool_call_id].append(msg)
+                existing_tool_msg_ids.add(msg.tool_call_id)
 
-        tool_call_ids: set[str] = set()
+        # Check if any patching is needed
+        needs_patch = False
         for msg in messages:
             if getattr(msg, "type", None) != "ai":
                 continue
             for tc in self._message_tool_calls(msg):
                 tc_id = tc.get("id")
-                if tc_id:
+                if tc_id and tc_id not in existing_tool_msg_ids:
-                    tool_call_ids.add(tc_id)
+                    needs_patch = True
+                    break
+            if needs_patch:
+                break
 
+        if not needs_patch:
+            return None
+
+        # Build new list with patches inserted right after each dangling AIMessage
         patched: list = []
+        patched_ids: set[str] = set()
         patch_count = 0
         for msg in messages:
-            if isinstance(msg, ToolMessage) and msg.tool_call_id in tool_call_ids:
-                continue
-
             patched.append(msg)
             if getattr(msg, "type", None) != "ai":
                 continue
-
             for tc in self._message_tool_calls(msg):
                 tc_id = tc.get("id")
-                if not tc_id:
+                if tc_id and tc_id not in existing_tool_msg_ids and tc_id not in patched_ids:
-                    continue
-
-                tool_msg_queue = tool_messages_by_id.get(tc_id)
-                existing_tool_msg = tool_msg_queue.popleft() if tool_msg_queue else None
-                if existing_tool_msg is not None:
-                    patched.append(existing_tool_msg)
-                else:
                     patched.append(
                         ToolMessage(
-                            content=self._synthetic_tool_message_content(tc),
+                            content="[Tool call was interrupted and did not return a result.]",
                             tool_call_id=tc_id,
                             name=tc.get("name", "unknown"),
                             status="error",
                         )
                     )
+                    patched_ids.add(tc_id)
                     patch_count += 1
 
-        if patched == messages:
+        logger.warning(f"Injecting {patch_count} placeholder ToolMessage(s) for dangling tool calls")
-            return None
-
-        if patch_count:
-            logger.warning(f"Injecting {patch_count} placeholder ToolMessage(s) for dangling tool calls")
         return patched
 
     @override

backend/packages/harness/deerflow/agents/middlewares/deferred_tool_filter_middleware.py

@@ -1,15 +1,12 @@
 """Middleware to filter deferred tool schemas from model binding.
 
-When tool_search is enabled, MCP tools are still passed to ToolNode for
+When tool_search is enabled, MCP tools are registered in the DeferredToolRegistry
-execution, but their schemas must NOT be sent to the LLM via bind_tools until
+and passed to ToolNode for execution, but their schemas should NOT be sent to the
-the model has discovered them via tool_search. This middleware removes the
+LLM via bind_tools (that's the whole point of deferral — saving context tokens).
-still-deferred tools from request.tools before model binding, and blocks tool
-calls to tools that have not been promoted yet.
 
-The deferred name set and the catalog hash are injected at construction time
+This middleware intercepts wrap_model_call and removes deferred tools from
-(no ContextVar). Promotion state is read from graph state (``state["promoted"]``),
+request.tools so that model.bind_tools only receives active tool schemas.
-scoped by catalog hash so a stale persisted promotion cannot expose a renamed
+The agent discovers deferred tools at runtime via the tool_search tool.
-or drifted tool.
 """
 
 import logging
@@ -27,49 +24,47 @@ logger = logging.getLogger(__name__)
 
 
 class DeferredToolFilterMiddleware(AgentMiddleware[AgentState]):
-    """Hide deferred tool schemas from the bound model until promoted.
+    """Remove deferred tools from request.tools before model binding.
 
     ToolNode still holds all tools (including deferred) for execution routing,
-    but the LLM only sees active tool schemas plus tools that have already been
+    but the LLM only sees active tool schemas — deferred tools are discoverable
-    promoted (recorded in ``state["promoted"]`` under the current catalog hash).
+    via tool_search at runtime.
     """
 
-    def __init__(self, deferred_names: frozenset[str], catalog_hash: str | None):
-        super().__init__()
-        self._deferred = deferred_names
-        self._catalog_hash = catalog_hash
-
-    def _promoted(self, state) -> set[str]:
-        promoted = (state or {}).get("promoted")
-        if promoted and promoted.get("catalog_hash") == self._catalog_hash:
-            return set(promoted.get("names") or [])
-        return set()
-
-    def _hidden(self, state) -> set[str]:
-        return set(self._deferred) - self._promoted(state)
-
     def _filter_tools(self, request: ModelRequest) -> ModelRequest:
-        if not self._deferred:
+        from deerflow.tools.builtins.tool_search import get_deferred_registry
+
+        registry = get_deferred_registry()
+        if not registry:
             return request
-        hide = self._hidden(request.state)
+
-        if not hide:
+        deferred_names = registry.deferred_names
-            return request
+        active_tools = [t for t in request.tools if getattr(t, "name", None) not in deferred_names]
-        active = [t for t in request.tools if getattr(t, "name", None) not in hide]
+
-        if len(active) < len(request.tools):
+        if len(active_tools) < len(request.tools):
-            logger.debug("Filtered %d deferred tool schema(s) from model binding", len(request.tools) - len(active))
+            logger.debug(f"Filtered {len(request.tools) - len(active_tools)} deferred tool schema(s) from model binding")
-        return request.override(tools=active)
+
+        return request.override(tools=active_tools)
 
     def _blocked_tool_message(self, request: ToolCallRequest) -> ToolMessage | None:
-        if not self._deferred:
+        from deerflow.tools.builtins.tool_search import get_deferred_registry
+
+        registry = get_deferred_registry()
+        if not registry:
             return None
-        name = str(request.tool_call.get("name") or "")
+
-        if not name or name not in self._hidden(request.state):
+        tool_name = str(request.tool_call.get("name") or "")
+        if not tool_name:
             return None
+
+        if not registry.contains(tool_name):
+            return None
+
         tool_call_id = str(request.tool_call.get("id") or "missing_tool_call_id")
         return ToolMessage(
-            content=(f"Error: Tool '{name}' is deferred and has not been promoted yet. Call tool_search first to expose and promote this tool's schema, then retry."),
+            content=(f"Error: Tool '{tool_name}' is deferred and has not been promoted yet. Call tool_search first to expose and promote this tool's schema, then retry."),
             tool_call_id=tool_call_id,
-            name=name,
+            name=tool_name,
             status="error",
         )
 

backend/packages/harness/deerflow/agents/middlewares/dynamic_context_middleware.py

@@ -1,232 +0,0 @@
-"""Middleware to inject dynamic context (memory, current date) as a system-reminder.
-
-The system prompt is kept fully static for maximum prefix-cache reuse across users
-and sessions.  The current date is always injected.  Per-user memory is also injected
-when ``memory.injection_enabled`` is True in the app config.  Both are delivered once
-per conversation as a dedicated <system-reminder> HumanMessage inserted before the
-first user message (frozen-snapshot pattern).
-
-When a conversation spans midnight the middleware detects the date change and injects
-a lightweight date-update reminder as a separate HumanMessage before the current turn.
-This correction is persisted so subsequent turns on the new day see a consistent history
-and do not re-inject.
-
-Reminder format:
-
-    <system-reminder>
-    <memory>...</memory>
-
-    <current_date>2026-05-08, Friday</current_date>
-    </system-reminder>
-
-Date-update format:
-
-    <system-reminder>
-    <current_date>2026-05-09, Saturday</current_date>
-    </system-reminder>
-"""
-
-from __future__ import annotations
-
-import asyncio
-import logging
-import re
-import uuid
-from datetime import datetime
-from typing import TYPE_CHECKING, override
-
-from langchain.agents.middleware import AgentMiddleware
-from langchain_core.messages import HumanMessage
-from langgraph.runtime import Runtime
-
-if TYPE_CHECKING:
-    from deerflow.config.app_config import AppConfig
-
-logger = logging.getLogger(__name__)
-
-# Upper bound (seconds) for a single _inject() offload.  If the warm-up at
-# gateway startup failed silently, the first request may still hit a cold
-# tiktoken BPE download that blocks until the OS TCP timeout (~26 min).
-# This cap ensures the request degrades gracefully instead of hanging.
-_INJECT_TIMEOUT_SECONDS = 5.0
-
-_DATE_RE = re.compile(r"<current_date>([^<]+)</current_date>")
-_DYNAMIC_CONTEXT_REMINDER_KEY = "dynamic_context_reminder"
-_SUMMARY_MESSAGE_NAME = "summary"
-
-
-def _extract_date(content: str) -> str | None:
-    """Return the first <current_date> value found in *content*, or None."""
-    m = _DATE_RE.search(content)
-    return m.group(1) if m else None
-
-
-def is_dynamic_context_reminder(message: object) -> bool:
-    """Return whether *message* is a hidden dynamic-context reminder."""
-    return isinstance(message, HumanMessage) and bool(message.additional_kwargs.get(_DYNAMIC_CONTEXT_REMINDER_KEY))
-
-
-def _last_injected_date(messages: list) -> str | None:
-    """Scan messages in reverse and return the most recently injected date.
-
-    Detection uses the ``dynamic_context_reminder`` additional_kwargs flag rather
-    than content substring matching, so user messages containing ``<system-reminder>``
-    are not mistakenly treated as injected reminders.
-    """
-    for msg in reversed(messages):
-        if is_dynamic_context_reminder(msg):
-            content_str = msg.content if isinstance(msg.content, str) else str(msg.content)
-            return _extract_date(content_str)
-    return None
-
-
-def _is_user_injection_target(message: object) -> bool:
-    """Return whether *message* can receive a dynamic-context reminder."""
-    return isinstance(message, HumanMessage) and not is_dynamic_context_reminder(message) and message.name != _SUMMARY_MESSAGE_NAME
-
-
-class DynamicContextMiddleware(AgentMiddleware):
-    """Inject memory and current date into HumanMessages as a <system-reminder>.
-
-    First turn
-    ----------
-    Prepends a full system-reminder (memory + date) to the first HumanMessage and
-    persists it (same message ID).  The first message is then frozen for the whole
-    session — its content never changes again, so the prefix cache can hit on every
-    subsequent turn.
-
-    Midnight crossing
-    -----------------
-    If the conversation spans midnight, the current date differs from the date that
-    was injected earlier.  In that case a lightweight date-update reminder is prepended
-    to the **current** (last) HumanMessage and persisted.  Subsequent turns on the new
-    day see the corrected date in history and skip re-injection.
-    """
-
-    def __init__(self, agent_name: str | None = None, *, app_config: AppConfig | None = None):
-        super().__init__()
-        self._agent_name = agent_name
-        self._app_config = app_config
-
-    def _build_full_reminder(self) -> str:
-        from deerflow.agents.lead_agent.prompt import _get_memory_context
-
-        # Memory injection is gated by injection_enabled; date is always included.
-        injection_enabled = self._app_config.memory.injection_enabled if self._app_config else True
-        memory_context = _get_memory_context(self._agent_name, app_config=self._app_config) if injection_enabled else ""
-        current_date = datetime.now().strftime("%Y-%m-%d, %A")
-
-        lines: list[str] = ["<system-reminder>"]
-        if memory_context:
-            lines.append(memory_context.strip())
-            lines.append("")  # blank line separating memory from date
-        lines.append(f"<current_date>{current_date}</current_date>")
-        lines.append("</system-reminder>")
-
-        return "\n".join(lines)
-
-    def _build_date_update_reminder(self) -> str:
-        current_date = datetime.now().strftime("%Y-%m-%d, %A")
-        return "\n".join(
-            [
-                "<system-reminder>",
-                f"<current_date>{current_date}</current_date>",
-                "</system-reminder>",
-            ]
-        )
-
-    @staticmethod
-    def _make_reminder_and_user_messages(original: HumanMessage, reminder_content: str) -> tuple[HumanMessage, HumanMessage]:
-        """Return (reminder_msg, user_msg) using the ID-swap technique.
-
-        reminder_msg takes the original message's ID so that add_messages replaces it
-        in-place (preserving position).  user_msg carries the original content with a
-        derived ``{id}__user`` ID and is appended immediately after by add_messages.
-
-        If the original message has no ID a stable UUID is generated so the derived
-        ``{id}__user`` ID never collapses to the ambiguous ``None__user`` string.
-        """
-        stable_id = original.id or str(uuid.uuid4())
-        reminder_msg = HumanMessage(
-            content=reminder_content,
-            id=stable_id,
-            additional_kwargs={"hide_from_ui": True, _DYNAMIC_CONTEXT_REMINDER_KEY: True},
-        )
-        user_msg = HumanMessage(
-            content=original.content,
-            id=f"{stable_id}__user",
-            name=original.name,
-            additional_kwargs=original.additional_kwargs,
-        )
-        return reminder_msg, user_msg
-
-    def _inject(self, state) -> dict | None:
-        messages = list(state.get("messages", []))
-        if not messages:
-            return None
-
-        current_date = datetime.now().strftime("%Y-%m-%d, %A")
-        last_date = _last_injected_date(messages)
-        logger.debug(
-            "DynamicContextMiddleware._inject: msg_count=%d last_date=%r current_date=%r",
-            len(messages),
-            last_date,
-            current_date,
-        )
-
-        if last_date is None:
-            # ── First turn: inject full reminder as a separate HumanMessage ─────
-            first_idx = next((i for i, m in enumerate(messages) if _is_user_injection_target(m)), None)
-            if first_idx is None:
-                return None
-            full_reminder = self._build_full_reminder()
-            logger.info(
-                "DynamicContextMiddleware: injecting full reminder (len=%d, has_memory=%s) into first HumanMessage id=%r",
-                len(full_reminder),
-                "<memory>" in full_reminder,
-                messages[first_idx].id,
-            )
-            reminder_msg, user_msg = self._make_reminder_and_user_messages(messages[first_idx], full_reminder)
-            return {"messages": [reminder_msg, user_msg]}
-
-        if last_date == current_date:
-            # ── Same day: nothing to do ──────────────────────────────────────────
-            return None
-
-        # ── Midnight crossed: inject date-update reminder as a separate HumanMessage ──
-        last_human_idx = next((i for i in reversed(range(len(messages))) if _is_user_injection_target(messages[i])), None)
-        if last_human_idx is None:
-            return None
-
-        reminder_msg, user_msg = self._make_reminder_and_user_messages(messages[last_human_idx], self._build_date_update_reminder())
-        logger.info("DynamicContextMiddleware: midnight crossing detected — injected date update before current turn")
-        return {"messages": [reminder_msg, user_msg]}
-
-    @override
-    def before_agent(self, state, runtime: Runtime) -> dict | None:
-        return self._inject(state)
-
-    @override
-    async def abefore_agent(self, state, runtime: Runtime) -> dict | None:
-        # _inject() performs synchronous file I/O (memory JSON loading) and
-        # potentially blocking network calls (tiktoken encoding download on
-        # first use).  Offload to a thread so the event loop is never blocked
-        # — a blocking call here starves all concurrent HTTP handlers (auth,
-        # SSE heartbeats, etc.).  See issue #3402.
-        #
-        # Bounded timeout: if startup warm-up failed silently (e.g. network
-        # blip during deploy), the first request's cold tiktoken download can
-        # block for tens of minutes (OS TCP timeout).  Time-box injection so
-        # the request degrades gracefully (no memory context) rather than
-        # hanging.
-        try:
-            return await asyncio.wait_for(
-                asyncio.to_thread(self._inject, state),
-                timeout=_INJECT_TIMEOUT_SECONDS,
-            )
-        except TimeoutError:
-            logger.warning(
-                "DynamicContextMiddleware: injection timed out (%.1fs); skipping memory/date injection for this turn",
-                _INJECT_TIMEOUT_SECONDS,
-            )
-            return None

backend/packages/harness/deerflow/agents/middlewares/llm_error_handling_middleware.py

@@ -62,41 +62,6 @@ _AUTH_PATTERNS = (
     "未授权",
 )
 
-# Per-exception retry budget overrides.
-#
-# Some transient errors are retriable in principle but expensive to retry at
-# the default budget. StreamChunkTimeoutError in particular fires after the
-# upstream provider has already stalled for `stream_chunk_timeout` seconds
-# (typically 120-240s); a full 3-attempt loop can therefore stack 6-12 minutes
-# of dead air before surfacing the failure to the user. We keep exactly one
-# retry (cheap reconnect that catches genuine transient TCP blips) and then
-# fail fast — the same buffered payload is overwhelmingly likely to fail
-# again at the upstream provider for the same reason.
-#
-# Keys are exception class *names* (not classes) so we don't introduce
-# import-time coupling on optional dependencies like langchain-openai. The
-# value is the absolute max attempt count, NOT additional retries — so a
-# value of 2 means "1 first attempt + 1 retry" (the CR-requested
-# "keep one retry" behavior).
-_RETRY_BUDGET_OVERRIDES: dict[str, int] = {
-    "StreamChunkTimeoutError": 2,
-}
-
-# Exception class names that indicate the upstream stream-chunk watchdog
-# fired because the model stalled mid-flight. These deserve a more specific
-# user-facing message than the generic "temporarily unavailable" copy,
-# because the typical root cause is a long tool-call serialization stalling
-# the upstream stream — and the most actionable advice we can give the user
-# is "ask for a shorter / split output" rather than "wait and retry".
-# Generic connection drops (httpx RemoteProtocolError / ReadError) are
-# intentionally excluded: they routinely fire on transient network blips
-# with normal payloads, where the "split the work" guidance is misleading.
-_STREAM_DROP_EXCEPTIONS: frozenset[str] = frozenset(
-    {
-        "StreamChunkTimeoutError",
-    }
-)
-
 
 class LLMErrorHandlingMiddleware(AgentMiddleware[AgentState]):
     """Retry transient LLM errors and surface graceful assistant messages."""
@@ -118,18 +83,6 @@ class LLMErrorHandlingMiddleware(AgentMiddleware[AgentState]):
         self._circuit_state = "closed"
         self._circuit_probe_in_flight = False
 
-    def _max_attempts_for(self, exc: BaseException) -> int:
-        """Return the effective max attempt count for this exception.
-
-        Falls back to `self.retry_max_attempts` unless the exception class name
-        appears in the per-exception override table.
-        """
-        override = _RETRY_BUDGET_OVERRIDES.get(type(exc).__name__)
-        if override is None:
-            return self.retry_max_attempts
-
-        return min(override, self.retry_max_attempts)
-
     def _check_circuit(self) -> bool:
         """Returns True if circuit is OPEN (fast fail), False otherwise."""
         with self._circuit_lock:
@@ -200,7 +153,6 @@ class LLMErrorHandlingMiddleware(AgentMiddleware[AgentState]):
             "InternalServerError",
             "ReadError",  # httpx.ReadError: connection dropped mid-stream
             "RemoteProtocolError",  # httpx: server closed connection unexpectedly
-            "StreamChunkTimeoutError",  # langchain-openai: chunk gap exceeded stream_chunk_timeout
         }:
             return True, "transient"
         if status_code in _RETRIABLE_STATUS_CODES:
@@ -225,24 +177,6 @@ class LLMErrorHandlingMiddleware(AgentMiddleware[AgentState]):
     def _build_circuit_breaker_message(self) -> str:
         return "The configured LLM provider is currently unavailable due to continuous failures. Circuit breaker is engaged to protect the system. Please wait a moment before trying again."
 
-    def _build_error_fallback_message(
-        self,
-        content: str,
-        *,
-        error_type: str,
-        reason: str,
-        detail: str,
-    ) -> AIMessage:
-        return AIMessage(
-            content=content,
-            additional_kwargs={
-                "deerflow_error_fallback": True,
-                "error_type": error_type,
-                "error_reason": reason,
-                "error_detail": detail,
-            },
-        )
-
     def _build_user_message(self, exc: BaseException, reason: str) -> str:
         detail = _extract_error_detail(exc)
         if reason == "quota":
@@ -250,31 +184,9 @@ class LLMErrorHandlingMiddleware(AgentMiddleware[AgentState]):
         if reason == "auth":
             return "The configured LLM provider rejected the request because authentication or access is invalid. Please check the provider credentials and try again."
         if reason in {"busy", "transient"}:
-            # Stream-drop failures (chunk-gap timeout, peer-closed connection,
-            # raw read error) almost always point at a single oversized
-            # tool-call payload — the model spent so long serializing JSON
-            # arguments that the upstream provider buffered and the stream
-            # gap exceeded `stream_chunk_timeout`. Surfacing this distinct
-            # cause lets the user split or shorten their next request
-            # instead of helplessly retrying the same prompt.
-            if type(exc).__name__ in _STREAM_DROP_EXCEPTIONS:
-                return (
-                    "The model's streaming response was interrupted before it could "
-                    "finish. This usually happens when a single response or tool call "
-                    "is very large — please ask the assistant to split the work into "
-                    "smaller steps, or shorten the requested output, and try again."
-                )
             return "The configured LLM provider is temporarily unavailable after multiple retries. Please wait a moment and continue the conversation."
         return f"LLM request failed: {detail}"
 
-    def _build_user_fallback_message(self, exc: BaseException, reason: str) -> AIMessage:
-        return self._build_error_fallback_message(
-            self._build_user_message(exc, reason),
-            error_type=type(exc).__name__,
-            reason=reason,
-            detail=_extract_error_detail(exc),
-        )
-
     def _emit_retry_event(self, attempt: int, wait_ms: int, reason: str) -> None:
         try:
             from langgraph.config import get_stream_writer
@@ -300,12 +212,7 @@ class LLMErrorHandlingMiddleware(AgentMiddleware[AgentState]):
         handler: Callable[[ModelRequest], ModelResponse],
     ) -> ModelCallResult:
         if self._check_circuit():
-            return self._build_error_fallback_message(
+            return AIMessage(content=self._build_circuit_breaker_message())
-                self._build_circuit_breaker_message(),
-                error_type="CircuitBreakerOpen",
-                reason="circuit_open",
-                detail="LLM circuit breaker is open",
-            )
 
         attempt = 1
         while True:
@@ -321,8 +228,7 @@ class LLMErrorHandlingMiddleware(AgentMiddleware[AgentState]):
                 raise
             except Exception as exc:
                 retriable, reason = self._classify_error(exc)
-                max_attempts = self._max_attempts_for(exc)
+                if retriable and attempt < self.retry_max_attempts:
-                if retriable and attempt < max_attempts:
                     wait_ms = self._build_retry_delay_ms(attempt, exc)
                     logger.warning(
                         "Transient LLM error on attempt %d/%d; retrying in %dms: %s",
@@ -343,7 +249,7 @@ class LLMErrorHandlingMiddleware(AgentMiddleware[AgentState]):
                 )
                 if retriable:
                     self._record_failure()
-                return self._build_user_fallback_message(exc, reason)
+                return AIMessage(content=self._build_user_message(exc, reason))
 
     @override
     async def awrap_model_call(
@@ -352,12 +258,7 @@ class LLMErrorHandlingMiddleware(AgentMiddleware[AgentState]):
         handler: Callable[[ModelRequest], Awaitable[ModelResponse]],
     ) -> ModelCallResult:
         if self._check_circuit():
-            return self._build_error_fallback_message(
+            return AIMessage(content=self._build_circuit_breaker_message())
-                self._build_circuit_breaker_message(),
-                error_type="CircuitBreakerOpen",
-                reason="circuit_open",
-                detail="LLM circuit breaker is open",
-            )
 
         attempt = 1
         while True:
@@ -373,8 +274,7 @@ class LLMErrorHandlingMiddleware(AgentMiddleware[AgentState]):
                 raise
             except Exception as exc:
                 retriable, reason = self._classify_error(exc)
-                max_attempts = self._max_attempts_for(exc)
+                if retriable and attempt < self.retry_max_attempts:
-                if retriable and attempt < max_attempts:
                     wait_ms = self._build_retry_delay_ms(attempt, exc)
                     logger.warning(
                         "Transient LLM error on attempt %d/%d; retrying in %dms: %s",
@@ -395,7 +295,7 @@ class LLMErrorHandlingMiddleware(AgentMiddleware[AgentState]):
                 )
                 if retriable:
                     self._record_failure()
-                return self._build_user_fallback_message(exc, reason)
+                return AIMessage(content=self._build_user_message(exc, reason))
 
 
 def _matches_any(detail: str, patterns: tuple[str, ...]) -> bool:

backend/packages/harness/deerflow/agents/middlewares/loop_detection_middleware.py

@@ -6,58 +6,25 @@ arguments indefinitely until the recursion limit kills the run.
 Detection strategy:
   1. After each model response, hash the tool calls (name + args).
   2. Track recent hashes in a sliding window.
-  3. If the same hash appears >= warn_threshold times, queue a
+  3. If the same hash appears >= warn_threshold times, inject a
-     "you are repeating yourself — wrap up" warning for the current
+     "you are repeating yourself — wrap up" system message (once per hash).
-     thread/run. The warning is **injected at the next model call** (in
-     ``wrap_model_call``) as a ``HumanMessage`` appended to the message
-     list, *after* all ToolMessage responses to the previous
-     AIMessage(tool_calls).
   4. If it appears >= hard_limit times, strip all tool_calls from the
      response so the agent is forced to produce a final text answer.
-
-Why the warning is injected at ``wrap_model_call`` instead of
-``after_model``:
-
-  ``after_model`` fires immediately after the model emits an
-  ``AIMessage`` that may carry ``tool_calls``. The tools node has not
-  run yet, so no matching ``ToolMessage`` exists in the history. Any
-  message we add here lands *between* the assistant's tool_calls and
-  their responses. OpenAI/Moonshot reject the next request with
-  ``"tool_call_ids did not have response messages"`` because their
-  validators require the assistant's tool_calls to be followed
-  immediately by tool messages. Anthropic also disallows mid-stream
-  ``SystemMessage``. By deferring the warning to ``wrap_model_call``,
-  every prior ToolMessage is already present in the request's message
-  list and the warning is appended at the end — pairing intact, no
-  ``AIMessage`` semantics are mutated.
-
-Queued warnings are intentionally transient. If a run ends before the
-next model request drains a queued warning, ``after_agent`` drops it
-instead of carrying it into a later invocation for the same thread. The
-hard-stop path still forces termination when the configured safety limit
-is reached.
 """
 
-from __future__ import annotations
-
 import hashlib
 import json
 import logging
 import threading
 from collections import OrderedDict, defaultdict
-from collections.abc import Awaitable, Callable
 from copy import deepcopy
-from typing import TYPE_CHECKING, override
+from typing import override
 
 from langchain.agents import AgentState
 from langchain.agents.middleware import AgentMiddleware
-from langchain.agents.middleware.types import ModelCallResult, ModelRequest, ModelResponse
 from langchain_core.messages import HumanMessage
 from langgraph.runtime import Runtime
 
-if TYPE_CHECKING:
-    from deerflow.config.loop_detection_config import LoopDetectionConfig
-
 logger = logging.getLogger(__name__)
 
 # Defaults — can be overridden via constructor
@@ -67,7 +34,6 @@ _DEFAULT_WINDOW_SIZE = 20  # track last N tool calls
 _DEFAULT_MAX_TRACKED_THREADS = 100  # LRU eviction limit
 _DEFAULT_TOOL_FREQ_WARN = 30  # warn after 30 calls to the same tool type
 _DEFAULT_TOOL_FREQ_HARD_LIMIT = 50  # force-stop after 50 calls to the same tool type
-_MAX_PENDING_WARNINGS_PER_RUN = 4
 
 
 def _normalize_tool_call_args(raw_args: object) -> tuple[dict, str | None]:
@@ -174,9 +140,6 @@ _TOOL_FREQ_HARD_STOP_MSG = "[FORCED STOP] Tool {tool_name} called {count} times
 class LoopDetectionMiddleware(AgentMiddleware[AgentState]):
     """Detects and breaks repetitive tool call loops.
 
-    Threshold parameters are validated upstream by :class:`LoopDetectionConfig`;
-    construct via :meth:`from_config` to ensure values pass Pydantic validation.
-
     Args:
         warn_threshold: Number of identical tool call sets before injecting
             a warning message. Default: 3.
@@ -192,14 +155,6 @@ class LoopDetectionMiddleware(AgentMiddleware[AgentState]):
             Default: 30.
         tool_freq_hard_limit: Number of calls to the same tool type before
             forcing a stop. Default: 50.
-        tool_freq_overrides: Per-tool overrides for frequency thresholds,
-            keyed by tool name. Each value is a ``(warn, hard_limit)`` tuple
-            that replaces ``tool_freq_warn`` / ``tool_freq_hard_limit`` for
-            that specific tool. Tools not listed here fall back to the global
-            thresholds. Useful for raising limits on intentionally
-            high-frequency tools (e.g. ``bash`` in batch pipelines) without
-            weakening protection on all other tools. Default: ``None``
-            (no overrides).
     """
 
     def __init__(
@@ -210,7 +165,6 @@ class LoopDetectionMiddleware(AgentMiddleware[AgentState]):
         max_tracked_threads: int = _DEFAULT_MAX_TRACKED_THREADS,
         tool_freq_warn: int = _DEFAULT_TOOL_FREQ_WARN,
         tool_freq_hard_limit: int = _DEFAULT_TOOL_FREQ_HARD_LIMIT,
-        tool_freq_overrides: dict[str, tuple[int, int]] | None = None,
     ):
         super().__init__()
         self.warn_threshold = warn_threshold
@@ -219,50 +173,21 @@ class LoopDetectionMiddleware(AgentMiddleware[AgentState]):
         self.max_tracked_threads = max_tracked_threads
         self.tool_freq_warn = tool_freq_warn
         self.tool_freq_hard_limit = tool_freq_hard_limit
-        self._tool_freq_overrides: dict[str, tuple[int, int]] = tool_freq_overrides or {}
         self._lock = threading.Lock()
+        # Per-thread tracking using OrderedDict for LRU eviction
         self._history: OrderedDict[str, list[str]] = OrderedDict()
         self._warned: dict[str, set[str]] = defaultdict(set)
+        # Per-thread, per-tool-type cumulative call counts
         self._tool_freq: dict[str, dict[str, int]] = defaultdict(lambda: defaultdict(int))
         self._tool_freq_warned: dict[str, set[str]] = defaultdict(set)
-        # Per-thread/run queue of warnings to inject at the next model call.
-        # Populated by ``after_model`` (detection) and drained by
-        # ``wrap_model_call`` (injection); see module docstring.
-        self._pending_warnings: dict[tuple[str, str], list[str]] = defaultdict(list)
-        self._pending_warning_touch_order: OrderedDict[tuple[str, str], None] = OrderedDict()
-        self._max_pending_warning_keys = max(1, self.max_tracked_threads * 2)
-
-    @classmethod
-    def from_config(cls, config: LoopDetectionConfig) -> LoopDetectionMiddleware:
-        """Construct from a Pydantic-validated config, trusting its validation."""
-        return cls(
-            warn_threshold=config.warn_threshold,
-            hard_limit=config.hard_limit,
-            window_size=config.window_size,
-            max_tracked_threads=config.max_tracked_threads,
-            tool_freq_warn=config.tool_freq_warn,
-            tool_freq_hard_limit=config.tool_freq_hard_limit,
-            tool_freq_overrides={name: (o.warn, o.hard_limit) for name, o in config.tool_freq_overrides.items()},
-        )
 
     def _get_thread_id(self, runtime: Runtime) -> str:
         """Extract thread_id from runtime context for per-thread tracking."""
         thread_id = runtime.context.get("thread_id") if runtime.context else None
         if thread_id:
-            return str(thread_id)
+            return thread_id
         return "default"
 
-    def _get_run_id(self, runtime: Runtime) -> str:
-        """Extract run_id from runtime context for per-run warning scoping."""
-        run_id = runtime.context.get("run_id") if runtime.context else None
-        if run_id:
-            return str(run_id)
-        return "default"
-
-    def _pending_key(self, runtime: Runtime) -> tuple[str, str]:
-        """Return the pending-warning key for the current thread/run."""
-        return self._get_thread_id(runtime), self._get_run_id(runtime)
-
     def _evict_if_needed(self) -> None:
         """Evict least recently used threads if over the limit.
 
@@ -273,52 +198,8 @@ class LoopDetectionMiddleware(AgentMiddleware[AgentState]):
             self._warned.pop(evicted_id, None)
             self._tool_freq.pop(evicted_id, None)
             self._tool_freq_warned.pop(evicted_id, None)
-            for key in list(self._pending_warnings):
-                if key[0] == evicted_id:
-                    self._drop_pending_warning_key_locked(key)
             logger.debug("Evicted loop tracking for thread %s (LRU)", evicted_id)
 
-    def _drop_pending_warning_key_locked(self, key: tuple[str, str]) -> None:
-        """Drop all pending-warning bookkeeping for one thread/run key.
-
-        Must be called while holding self._lock.
-        """
-        self._pending_warnings.pop(key, None)
-        self._pending_warning_touch_order.pop(key, None)
-
-    def _touch_pending_warning_key_locked(self, key: tuple[str, str]) -> None:
-        """Mark a pending-warning key as recently used.
-
-        Must be called while holding self._lock.
-        """
-        self._pending_warning_touch_order[key] = None
-        self._pending_warning_touch_order.move_to_end(key)
-
-    def _prune_pending_warning_state_locked(self, protected_key: tuple[str, str]) -> None:
-        """Cap pending-warning state across abnormal or concurrent runs.
-
-        Must be called while holding self._lock.
-        """
-        overflow = len(self._pending_warning_touch_order) - self._max_pending_warning_keys
-        if overflow <= 0:
-            return
-
-        candidates = [key for key in self._pending_warning_touch_order if key != protected_key]
-        for key in candidates[:overflow]:
-            self._drop_pending_warning_key_locked(key)
-
-    def _queue_pending_warning(self, runtime: Runtime, warning: str) -> None:
-        """Queue one transient warning for the current thread/run with caps."""
-        pending_key = self._pending_key(runtime)
-        with self._lock:
-            warnings = self._pending_warnings[pending_key]
-            if warning not in warnings:
-                warnings.append(warning)
-            if len(warnings) > _MAX_PENDING_WARNINGS_PER_RUN:
-                del warnings[: len(warnings) - _MAX_PENDING_WARNINGS_PER_RUN]
-            self._touch_pending_warning_key_locked(pending_key)
-            self._prune_pending_warning_state_locked(protected_key=pending_key)
-
     def _track_and_check(self, state: AgentState, runtime: Runtime) -> tuple[str | None, bool]:
         """Track tool calls and check for loops.
 
@@ -359,12 +240,6 @@ class LoopDetectionMiddleware(AgentMiddleware[AgentState]):
             if len(history) > self.window_size:
                 history[:] = history[-self.window_size :]
 
-            warned_hashes = self._warned.get(thread_id)
-            if warned_hashes is not None:
-                warned_hashes.intersection_update(history)
-                if not warned_hashes:
-                    self._warned.pop(thread_id, None)
-
             count = history.count(call_hash)
             tool_names = [tc.get("name", "?") for tc in tool_calls]
 
@@ -405,12 +280,7 @@ class LoopDetectionMiddleware(AgentMiddleware[AgentState]):
                 freq[name] += 1
                 tc_count = freq[name]
 
-                if name in self._tool_freq_overrides:
+                if tc_count >= self.tool_freq_hard_limit:
-                    eff_warn, eff_hard = self._tool_freq_overrides[name]
-                else:
-                    eff_warn, eff_hard = self.tool_freq_warn, self.tool_freq_hard_limit
-
-                if tc_count >= eff_hard:
                     logger.error(
                         "Tool frequency hard limit reached — forcing stop",
                         extra={
@@ -421,7 +291,7 @@ class LoopDetectionMiddleware(AgentMiddleware[AgentState]):
                     )
                     return _TOOL_FREQ_HARD_STOP_MSG.format(tool_name=name, count=tc_count), True
 
-                if tc_count >= eff_warn:
+                if tc_count >= self.tool_freq_warn:
                     warned = self._tool_freq_warned[thread_id]
                     if name not in warned:
                         warned.add(name)
@@ -478,10 +348,7 @@ class LoopDetectionMiddleware(AgentMiddleware[AgentState]):
         warning, hard_stop = self._track_and_check(state, runtime)
 
         if hard_stop:
-            # Strip tool_calls from the last AIMessage to force text output.
+            # Strip tool_calls from the last AIMessage to force text output
-            # Once tool_calls are stripped, the AIMessage no longer requires
-            # matching ToolMessage responses, so mutating it in place here
-            # is safe for OpenAI/Moonshot pairing validators.
             messages = state.get("messages", [])
             last_msg = messages[-1]
             content = self._append_text(last_msg.content, warning or _HARD_STOP_MSG)
@@ -489,48 +356,16 @@ class LoopDetectionMiddleware(AgentMiddleware[AgentState]):
             return {"messages": [stripped_msg]}
 
         if warning:
-            # Defer injection to the next model call. We must NOT alter the
+            # Inject as HumanMessage instead of SystemMessage to avoid
-            # AIMessage(tool_calls=...) here (would put framework words in
+            # Anthropic's "multiple non-consecutive system messages" error.
-            # the model's mouth, polluting downstream consumers like
+            # Anthropic models require system messages only at the start of
-            # MemoryMiddleware), nor insert a separate non-tool message
+            # the conversation; injecting one mid-conversation crashes
-            # (would break OpenAI/Moonshot tool-call pairing because the
+            # langchain_anthropic's _format_messages(). HumanMessage works
-            # tools node has not produced ToolMessage responses yet). The
+            # with all providers. See #1299.
-            # warning is delivered via ``wrap_model_call`` below.
+            return {"messages": [HumanMessage(content=warning, name="loop_warning")]}
-            self._queue_pending_warning(runtime, warning)
-            return None
 
         return None
 
-    def _clear_other_run_pending_warnings(self, runtime: Runtime) -> None:
-        """Drop stale pending warnings for previous runs in this thread."""
-        thread_id, current_run_id = self._pending_key(runtime)
-        with self._lock:
-            for key in list(self._pending_warnings):
-                if key[0] == thread_id and key[1] != current_run_id:
-                    self._drop_pending_warning_key_locked(key)
-
-    def _clear_current_run_pending_warnings(self, runtime: Runtime) -> None:
-        """Drop pending warnings owned by the current thread/run."""
-        pending_key = self._pending_key(runtime)
-        with self._lock:
-            self._drop_pending_warning_key_locked(pending_key)
-
-    @staticmethod
-    def _format_warning_message(warnings: list[str]) -> str:
-        """Merge pending warnings into one prompt message."""
-        deduped = list(dict.fromkeys(warnings))
-        return "\n\n".join(deduped)
-
-    @override
-    def before_agent(self, state: AgentState, runtime: Runtime) -> dict | None:
-        self._clear_other_run_pending_warnings(runtime)
-        return None
-
-    @override
-    async def abefore_agent(self, state: AgentState, runtime: Runtime) -> dict | None:
-        self._clear_other_run_pending_warnings(runtime)
-        return None
-
     @override
     def after_model(self, state: AgentState, runtime: Runtime) -> dict | None:
         return self._apply(state, runtime)
@@ -539,59 +374,6 @@ class LoopDetectionMiddleware(AgentMiddleware[AgentState]):
     async def aafter_model(self, state: AgentState, runtime: Runtime) -> dict | None:
         return self._apply(state, runtime)
 
-    @override
-    def after_agent(self, state: AgentState, runtime: Runtime) -> dict | None:
-        self._clear_current_run_pending_warnings(runtime)
-        return None
-
-    @override
-    async def aafter_agent(self, state: AgentState, runtime: Runtime) -> dict | None:
-        self._clear_current_run_pending_warnings(runtime)
-        return None
-
-    def _drain_pending_warnings(self, runtime: Runtime) -> list[str]:
-        """Pop and return all queued warnings for *runtime*'s thread/run."""
-        pending_key = self._pending_key(runtime)
-        with self._lock:
-            warnings = self._pending_warnings.pop(pending_key, [])
-            self._pending_warning_touch_order.pop(pending_key, None)
-        return warnings
-
-    def _augment_request(self, request: ModelRequest) -> ModelRequest:
-        """Append queued loop warnings (if any) to the outgoing message list.
-
-        The warning is placed *after* every existing message, including the
-        ToolMessage responses to the previous AIMessage(tool_calls). This
-        keeps ``assistant tool_calls -> tool_messages`` pairing intact for
-        OpenAI/Moonshot, avoids the Anthropic mid-stream SystemMessage
-        restriction (we use HumanMessage), and never mutates an existing
-        AIMessage.
-        """
-        warnings = self._drain_pending_warnings(request.runtime)
-        if not warnings:
-            return request
-        new_messages = [
-            *request.messages,
-            HumanMessage(content=self._format_warning_message(warnings), name="loop_warning"),
-        ]
-        return request.override(messages=new_messages)
-
-    @override
-    def wrap_model_call(
-        self,
-        request: ModelRequest,
-        handler: Callable[[ModelRequest], ModelResponse],
-    ) -> ModelCallResult:
-        return handler(self._augment_request(request))
-
-    @override
-    async def awrap_model_call(
-        self,
-        request: ModelRequest,
-        handler: Callable[[ModelRequest], Awaitable[ModelResponse]],
-    ) -> ModelCallResult:
-        return await handler(self._augment_request(request))
-
     def reset(self, thread_id: str | None = None) -> None:
         """Clear tracking state. If thread_id given, clear only that thread."""
         with self._lock:
@@ -600,13 +382,8 @@ class LoopDetectionMiddleware(AgentMiddleware[AgentState]):
                 self._warned.pop(thread_id, None)
                 self._tool_freq.pop(thread_id, None)
                 self._tool_freq_warned.pop(thread_id, None)
-                for key in list(self._pending_warnings):
-                    if key[0] == thread_id:
-                        self._drop_pending_warning_key_locked(key)
             else:
                 self._history.clear()
                 self._warned.clear()
                 self._tool_freq.clear()
                 self._tool_freq_warned.clear()
-                self._pending_warnings.clear()
-                self._pending_warning_touch_order.clear()

backend/packages/harness/deerflow/agents/middlewares/safety_finish_reason_middleware.py

@@ -1,317 +0,0 @@
-"""Suppress tool execution when the provider safety-terminated the response.
-
-Background — see issue bytedance/deer-flow#3028.
-
-Some providers (OpenAI ``finish_reason='content_filter'``, Anthropic
-``stop_reason='refusal'``, Gemini ``finish_reason='SAFETY'`` ...) can stop
-generation mid-stream while still returning partially-formed ``tool_calls``.
-LangChain's tool router treats any AIMessage with a non-empty ``tool_calls``
-field as "go execute these", so half-truncated arguments — e.g. a markdown
-``write_file`` that stops in the middle of a sentence — get dispatched as if
-they were complete. The agent then sees the truncated file, tries to fix it,
-gets filtered again, and loops.
-
-This middleware sits at ``after_model`` and gates that behaviour: when a
-configured ``SafetyTerminationDetector`` fires *and* the AIMessage carries
-tool calls, we strip the tool calls (both structured and raw provider
-payloads), append a user-facing explanation, and stash observability fields
-in ``additional_kwargs.safety_termination`` so logs, traces, and SSE
-consumers can see what happened.
-
-Hook choice: ``after_model`` (not ``wrap_model_call``) because the response
-is a *normal* return — not an exception — and we want to participate in the
-same after-model chain as ``LoopDetectionMiddleware``, with which we share
-the same tool-call-suppression mechanic but a different trigger.
-
-Placement: register *after* ``LoopDetectionMiddleware`` in the middleware
-list. LangChain factory wires ``after_model`` edges in reverse list order
-(``langchain/agents/factory.py:add_edge("model", middleware_w_after_model[-1])``,
-then walks ``range(len-1, 0, -1)``), so the *last* registered middleware is
-the *first* to observe the model output. Registering Safety after Loop
-means Safety sees the raw response first, clears tool calls if it fires,
-and Loop then accounts against the cleaned message.
-"""
-
-from __future__ import annotations
-
-import logging
-from typing import TYPE_CHECKING, override
-
-from langchain.agents import AgentState
-from langchain.agents.middleware import AgentMiddleware
-from langchain_core.messages import AIMessage
-from langgraph.runtime import Runtime
-
-from deerflow.agents.middlewares.safety_termination_detectors import (
-    SafetyTermination,
-    SafetyTerminationDetector,
-    default_detectors,
-)
-from deerflow.agents.middlewares.tool_call_metadata import clone_ai_message_with_tool_calls
-
-if TYPE_CHECKING:
-    from deerflow.config.safety_finish_reason_config import SafetyFinishReasonConfig
-
-logger = logging.getLogger(__name__)
-
-
-_USER_FACING_MESSAGE = (
-    "The model provider stopped this response with a safety-related signal "
-    "({reason_field}={reason_value!r}, detector={detector!r}). Any tool "
-    "calls produced in this turn were suppressed because their arguments "
-    "may be truncated and unsafe to execute. Please rephrase the request "
-    "or ask for a narrower output."
-)
-
-
-class SafetyFinishReasonMiddleware(AgentMiddleware[AgentState]):
-    """Strip tool_calls from AIMessages flagged by a SafetyTerminationDetector."""
-
-    def __init__(self, detectors: list[SafetyTerminationDetector] | None = None) -> None:
-        super().__init__()
-        # Copy so caller mutations after construction don't leak into us.
-        self._detectors: list[SafetyTerminationDetector] = list(detectors) if detectors else default_detectors()
-
-    @classmethod
-    def from_config(cls, config: SafetyFinishReasonConfig) -> SafetyFinishReasonMiddleware:
-        """Construct from validated Pydantic config, honouring the
-        reflection-loaded detector list when provided.
-
-        An explicit empty list is intentionally rejected — it would silently
-        disable detection while leaving the middleware in the chain, which
-        is the worst of both worlds. Use ``enabled: false`` instead.
-        """
-        if config.detectors is None:
-            return cls()
-
-        if not config.detectors:
-            raise ValueError("safety_finish_reason.detectors must be omitted (use built-ins) or contain at least one entry; use enabled=false to disable the middleware entirely.")
-
-        from deerflow.reflection import resolve_variable
-
-        detectors: list[SafetyTerminationDetector] = []
-        for entry in config.detectors:
-            detector_cls = resolve_variable(entry.use)
-            kwargs = dict(entry.config) if entry.config else {}
-            detector = detector_cls(**kwargs)
-            if not isinstance(detector, SafetyTerminationDetector):
-                raise TypeError(f"{entry.use} did not produce a SafetyTerminationDetector (got {type(detector).__name__}); ensure it has a `name` attribute and a `detect(message)` method")
-            detectors.append(detector)
-        return cls(detectors=detectors)
-
-    # ----- detection -------------------------------------------------------
-
-    def _detect(self, message: AIMessage) -> SafetyTermination | None:
-        for detector in self._detectors:
-            try:
-                hit = detector.detect(message)
-            except Exception:  # noqa: BLE001 - never let a buggy detector break the agent run
-                logger.exception("SafetyTerminationDetector %r raised; treating as no-match", getattr(detector, "name", type(detector).__name__))
-                continue
-            if hit is not None:
-                return hit
-        return None
-
-    # ----- message rewriting ----------------------------------------------
-
-    @staticmethod
-    def _append_user_message(content: object, text: str) -> str | list:
-        """Append a plain-text explanation to AIMessage content.
-
-        Mirrors ``LoopDetectionMiddleware._append_text`` so list-content
-        responses (Anthropic thinking blocks, vLLM reasoning splits) keep
-        their structure instead of being string-coerced into a TypeError.
-        """
-        if content is None or content == "":
-            return text
-        if isinstance(content, list):
-            return [*content, {"type": "text", "text": f"\n\n{text}"}]
-        if isinstance(content, str):
-            return content + f"\n\n{text}"
-        return str(content) + f"\n\n{text}"
-
-    def _build_suppressed_message(
-        self,
-        message: AIMessage,
-        termination: SafetyTermination,
-    ) -> AIMessage:
-        suppressed_names = [tc.get("name") or "unknown" for tc in (message.tool_calls or [])]
-        explanation = _USER_FACING_MESSAGE.format(
-            reason_field=termination.reason_field,
-            reason_value=termination.reason_value,
-            detector=termination.detector,
-        )
-        new_content = self._append_user_message(message.content, explanation)
-
-        # clone_ai_message_with_tool_calls handles structured tool_calls,
-        # raw additional_kwargs.tool_calls, and function_call in one shot.
-        # It only rewrites finish_reason when the old value was "tool_calls",
-        # which is not our case — content_filter / refusal / SAFETY stay put
-        # so downstream SSE / converters keep seeing the real provider reason.
-        cleared = clone_ai_message_with_tool_calls(message, [], content=new_content)
-
-        # Re-clone additional_kwargs so we don't accidentally mutate the
-        # dict returned by clone_ai_message_with_tool_calls (which already
-        # made a shallow copy, but downstream model_copy still references
-        # it). Then stamp the observability record.
-        kwargs = dict(getattr(cleared, "additional_kwargs", None) or {})
-        kwargs["safety_termination"] = {
-            "detector": termination.detector,
-            "reason_field": termination.reason_field,
-            "reason_value": termination.reason_value,
-            "suppressed_tool_call_count": len(suppressed_names),
-            "suppressed_tool_call_names": suppressed_names,
-            "extras": dict(termination.extras) if termination.extras else {},
-        }
-        return cleared.model_copy(update={"additional_kwargs": kwargs})
-
-    # ----- observability ---------------------------------------------------
-
-    def _emit_event(
-        self,
-        termination: SafetyTermination,
-        suppressed_names: list[str],
-        runtime: Runtime,
-    ) -> None:
-        """Notify SSE consumers (e.g. the web UI) that a tool turn was
-        suppressed so they can reconcile any "tool starting..." placeholders
-        already streamed to the user. Failures are logged at debug and
-        ignored — this is a best-effort signal."""
-        try:
-            from langgraph.config import get_stream_writer
-
-            writer = get_stream_writer()
-        except Exception:  # noqa: BLE001
-            logger.debug("get_stream_writer unavailable; skipping safety_termination event", exc_info=True)
-            return
-
-        thread_id = None
-        if runtime is not None and getattr(runtime, "context", None):
-            thread_id = runtime.context.get("thread_id") if isinstance(runtime.context, dict) else None
-
-        try:
-            writer(
-                {
-                    "type": "safety_termination",
-                    "detector": termination.detector,
-                    "reason_field": termination.reason_field,
-                    "reason_value": termination.reason_value,
-                    "suppressed_tool_call_count": len(suppressed_names),
-                    "suppressed_tool_call_names": suppressed_names,
-                    "thread_id": thread_id,
-                }
-            )
-        except Exception:  # noqa: BLE001
-            logger.debug("Failed to emit safety_termination stream event", exc_info=True)
-
-    def _record_audit_event(
-        self,
-        termination: SafetyTermination,
-        message,
-        tool_calls: list[dict],
-        runtime: Runtime,
-    ) -> None:
-        """Write a ``middleware:safety_termination`` record to RunEventStore
-        for post-run auditability.
-
-        The custom stream event in ``_emit_event`` is consumed by live SSE
-        clients and disappears after the run; this event is persisted so an
-        operator can answer "which runs were safety-suppressed today?" from
-        a single SQL query without joining the message body. Worker exposes
-        the run-scoped ``RunJournal`` via ``runtime.context["__run_journal"]``;
-        absent in unit-test / subagent / no-event-store paths, in which case
-        we silently skip.
-
-        Tool **arguments** are deliberately **not** recorded — those are the
-        very content the provider filtered; persisting them would defeat the
-        purpose of the safety filter. Names / count / ids are sufficient for
-        audit and debugging (issue #3028 review).
-        """
-        journal = None
-        if runtime is not None and getattr(runtime, "context", None):
-            context = runtime.context
-            if isinstance(context, dict):
-                journal = context.get("__run_journal")
-        if journal is None:
-            return
-
-        suppressed_names = [tc.get("name") or "unknown" for tc in tool_calls]
-        suppressed_ids = [tc.get("id") for tc in tool_calls if tc.get("id")]
-
-        changes = {
-            "detector": termination.detector,
-            "reason_field": termination.reason_field,
-            "reason_value": termination.reason_value,
-            "suppressed_tool_call_count": len(tool_calls),
-            "suppressed_tool_call_names": suppressed_names,
-            "suppressed_tool_call_ids": suppressed_ids,
-            "message_id": getattr(message, "id", None),
-            "extras": dict(termination.extras) if termination.extras else {},
-        }
-
-        try:
-            journal.record_middleware(
-                tag="safety_termination",
-                name=type(self).__name__,
-                hook="after_model",
-                action="suppress_tool_calls",
-                changes=changes,
-            )
-        except Exception:  # noqa: BLE001
-            # Audit-event persistence must never break agent execution.
-            logger.debug("Failed to record middleware:safety_termination event", exc_info=True)
-
-    # ----- main apply ------------------------------------------------------
-
-    def _apply(self, state: AgentState, runtime: Runtime) -> dict | None:
-        messages = state.get("messages", [])
-        if not messages:
-            return None
-
-        last = messages[-1]
-        if not isinstance(last, AIMessage):
-            return None
-
-        # Issue scope: only intervene when there's something to suppress.
-        # ``content_filter`` without tool_calls is allowed through unchanged
-        # so the partial text response (if any) reaches the user naturally.
-        tool_calls = last.tool_calls
-        if not tool_calls:
-            return None
-
-        termination = self._detect(last)
-        if termination is None:
-            return None
-
-        patched = self._build_suppressed_message(last, termination)
-
-        thread_id = None
-        if runtime is not None and getattr(runtime, "context", None):
-            thread_id = runtime.context.get("thread_id") if isinstance(runtime.context, dict) else None
-
-        logger.warning(
-            "Provider safety termination detected — suppressed %d tool call(s)",
-            len(tool_calls),
-            extra={
-                "thread_id": thread_id,
-                "detector": termination.detector,
-                "reason_field": termination.reason_field,
-                "reason_value": termination.reason_value,
-                "suppressed_tool_call_names": [tc.get("name") for tc in tool_calls],
-            },
-        )
-
-        self._emit_event(termination, [tc.get("name") or "unknown" for tc in tool_calls], runtime)
-        self._record_audit_event(termination, last, list(tool_calls), runtime)
-
-        return {"messages": [patched]}
-
-    # ----- hooks -----------------------------------------------------------
-
-    @override
-    def after_model(self, state: AgentState, runtime: Runtime) -> dict | None:
-        return self._apply(state, runtime)
-
-    @override
-    async def aafter_model(self, state: AgentState, runtime: Runtime) -> dict | None:
-        return self._apply(state, runtime)

backend/packages/harness/deerflow/agents/middlewares/safety_termination_detectors.py

@@ -1,237 +0,0 @@
-"""Detectors for provider-side safety termination signals.
-
-Different LLM providers signal "I stopped this response for safety reasons"
-through different fields with different values. This module defines a small
-strategy interface and three built-in detectors that cover the major
-providers DeerFlow supports today. New providers (Wenxin, Hunyuan, Bedrock
-adapters, in-house gateways, ...) can be added by implementing
-``SafetyTerminationDetector`` and wiring it through
-``config.yaml: safety_finish_reason.detectors``.
-
-The middleware that consumes these detectors lives in
-``safety_finish_reason_middleware.py``.
-"""
-
-from __future__ import annotations
-
-from dataclasses import dataclass, field
-from typing import Any, Protocol, runtime_checkable
-
-from langchain_core.messages import AIMessage
-
-
-@dataclass(frozen=True)
-class SafetyTermination:
-    """A detected safety-related termination signal.
-
-    Attributes:
-        detector: Name of the detector that produced this result. Used for
-            observability so operators can see which provider rule fired.
-        reason_field: The message metadata field that carried the signal
-            (e.g. ``finish_reason``, ``stop_reason``).
-        reason_value: The actual value of that field
-            (e.g. ``content_filter``, ``refusal``, ``SAFETY``).
-        extras: Provider-specific metadata that may help downstream
-            consumers (e.g. Azure OpenAI content_filter_results, Gemini
-            safety_ratings). Detectors are free to populate or skip this.
-    """
-
-    detector: str
-    reason_field: str
-    reason_value: str
-    extras: dict[str, Any] = field(default_factory=dict)
-
-
-@runtime_checkable
-class SafetyTerminationDetector(Protocol):
-    """Strategy interface for provider safety termination detection."""
-
-    name: str
-
-    def detect(self, message: AIMessage) -> SafetyTermination | None:
-        """Return a SafetyTermination if *message* indicates provider safety
-        termination, otherwise return ``None``.
-
-        Implementations must be side-effect free and tolerant of missing or
-        oddly-typed metadata — detectors run on every model response.
-        """
-        ...
-
-
-def _get_metadata_value(message: AIMessage, field_name: str) -> str | None:
-    """Read a string-typed value from either ``response_metadata`` or
-    ``additional_kwargs``.
-
-    LangChain provider adapters are inconsistent about where they stash
-    provider stop signals. Most modern adapters use ``response_metadata``,
-    but some legacy / passthrough paths still surface them via
-    ``additional_kwargs``. We check both, in that order, and only accept
-    string values — Pydantic enums or dicts are ignored so we never raise
-    on malformed inputs.
-    """
-    for container_name in ("response_metadata", "additional_kwargs"):
-        container = getattr(message, container_name, None) or {}
-        if not isinstance(container, dict):
-            continue
-        value = container.get(field_name)
-        if isinstance(value, str) and value:
-            return value
-    return None
-
-
-class OpenAICompatibleContentFilterDetector:
-    """OpenAI-compatible content_filter signal.
-
-    Covers OpenAI, Azure OpenAI, Moonshot/Kimi, DeepSeek, Mistral, vLLM,
-    Qwen (OpenAI-compatible mode), and any other adapter that follows the
-    OpenAI ``finish_reason`` convention.
-
-    Some Chinese providers ship custom OpenAI-compatible gateways that use
-    alternative tokens like ``sensitive`` or ``violation``. Extend the set
-    via the ``finish_reasons`` kwarg in config.
-    """
-
-    name = "openai_compatible_content_filter"
-
-    def __init__(self, finish_reasons: list[str] | tuple[str, ...] | None = None) -> None:
-        configured = finish_reasons if finish_reasons is not None else ("content_filter",)
-        self._finish_reasons: frozenset[str] = frozenset(r.lower() for r in configured)
-
-    def detect(self, message: AIMessage) -> SafetyTermination | None:
-        value = _get_metadata_value(message, "finish_reason")
-        if value is None or value.lower() not in self._finish_reasons:
-            return None
-
-        extras: dict[str, Any] = {}
-        # Azure OpenAI ships a structured content_filter_results block; carry it
-        # through so operators can see *what* was filtered without re-tracing.
-        response_metadata = getattr(message, "response_metadata", None) or {}
-        if isinstance(response_metadata, dict):
-            filter_results = response_metadata.get("content_filter_results")
-            if filter_results:
-                extras["content_filter_results"] = filter_results
-
-        return SafetyTermination(
-            detector=self.name,
-            reason_field="finish_reason",
-            reason_value=value,
-            extras=extras,
-        )
-
-
-class AnthropicRefusalDetector:
-    """Anthropic ``stop_reason == "refusal"`` signal.
-
-    Anthropic models surface safety refusals via a dedicated ``stop_reason``
-    rather than ``finish_reason``. See:
-    https://platform.claude.com/docs/en/test-and-evaluate/strengthen-guardrails/handle-streaming-refusals
-    """
-
-    name = "anthropic_refusal"
-
-    def __init__(self, stop_reasons: list[str] | tuple[str, ...] | None = None) -> None:
-        configured = stop_reasons if stop_reasons is not None else ("refusal",)
-        self._stop_reasons: frozenset[str] = frozenset(r.lower() for r in configured)
-
-    def detect(self, message: AIMessage) -> SafetyTermination | None:
-        value = _get_metadata_value(message, "stop_reason")
-        if value is None or value.lower() not in self._stop_reasons:
-            return None
-        return SafetyTermination(
-            detector=self.name,
-            reason_field="stop_reason",
-            reason_value=value,
-        )
-
-
-class GeminiSafetyDetector:
-    """Gemini / Vertex AI safety-related finish reasons.
-
-    Gemini uses the same ``finish_reason`` field as OpenAI but with an
-    enumerated upper-case taxonomy. The default set covers every Gemini
-    finish_reason that means "the model stopped because the content/image
-    tripped a safety, blocklist, recitation, or PII filter" — i.e. cases
-    where any tool_calls returned alongside are likely truncated/
-    unreliable. Full enum:
-    https://docs.cloud.google.com/python/docs/reference/aiplatform/latest/google.cloud.aiplatform_v1.types.Candidate.FinishReason
-
-    Intentionally **excluded** from the default set:
-    - ``STOP``                       — normal termination.
-    - ``MAX_TOKENS``                 — output length truncation, not safety
-                                       (same root failure mode as
-                                       content_filter, but issue #3028
-                                       scopes it out; expose separately if
-                                       desired).
-    - ``LANGUAGE`` / ``NO_IMAGE``    — capability mismatches, unrelated to
-                                       safety; tool_calls would be absent
-                                       anyway.
-    - ``MALFORMED_FUNCTION_CALL`` /
-      ``UNEXPECTED_TOOL_CALL``       — tool-call protocol errors. The
-                                       tool_calls are *also* unreliable
-                                       here, but the failure category is
-                                       distinct from safety filtering;
-                                       handle in a dedicated detector to
-                                       keep observability records honest.
-    - ``OTHER`` / ``IMAGE_OTHER`` /
-      ``FINISH_REASON_UNSPECIFIED``  — too broad to enable by default;
-                                       opt in via ``finish_reasons=`` if
-                                       your provider abuses these.
-    """
-
-    name = "gemini_safety"
-
-    _DEFAULT_FINISH_REASONS = (
-        # Text safety
-        "SAFETY",
-        "BLOCKLIST",
-        "PROHIBITED_CONTENT",
-        "SPII",
-        "RECITATION",
-        # Image safety (multimodal generation)
-        "IMAGE_SAFETY",
-        "IMAGE_PROHIBITED_CONTENT",
-        "IMAGE_RECITATION",
-    )
-
-    def __init__(self, finish_reasons: list[str] | tuple[str, ...] | None = None) -> None:
-        configured = finish_reasons if finish_reasons is not None else self._DEFAULT_FINISH_REASONS
-        self._finish_reasons: frozenset[str] = frozenset(r.upper() for r in configured)
-
-    def detect(self, message: AIMessage) -> SafetyTermination | None:
-        value = _get_metadata_value(message, "finish_reason")
-        if value is None or value.upper() not in self._finish_reasons:
-            return None
-
-        extras: dict[str, Any] = {}
-        response_metadata = getattr(message, "response_metadata", None) or {}
-        if isinstance(response_metadata, dict):
-            # Gemini surfaces per-category scoring under safety_ratings.
-            ratings = response_metadata.get("safety_ratings")
-            if ratings:
-                extras["safety_ratings"] = ratings
-
-        return SafetyTermination(
-            detector=self.name,
-            reason_field="finish_reason",
-            reason_value=value,
-            extras=extras,
-        )
-
-
-def default_detectors() -> list[SafetyTerminationDetector]:
-    """Built-in detector set used when no custom detectors are configured."""
-    return [
-        OpenAICompatibleContentFilterDetector(),
-        AnthropicRefusalDetector(),
-        GeminiSafetyDetector(),
-    ]
-
-
-__all__ = [
-    "AnthropicRefusalDetector",
-    "GeminiSafetyDetector",
-    "OpenAICompatibleContentFilterDetector",
-    "SafetyTermination",
-    "SafetyTerminationDetector",
-    "default_detectors",
-]

backend/packages/harness/deerflow/agents/middlewares/skill_activation_middleware.py

@@ -1,289 +0,0 @@
-"""Middleware for explicit slash skill activation."""
-
-from __future__ import annotations
-
-import asyncio
-import hashlib
-import html
-import logging
-import uuid
-from collections.abc import Awaitable, Callable
-from dataclasses import dataclass
-from pathlib import Path
-from typing import TYPE_CHECKING, override
-
-from langchain.agents.middleware import AgentMiddleware
-from langchain.agents.middleware.types import ModelRequest, ModelResponse
-from langchain_core.messages import AIMessage, HumanMessage
-
-from deerflow.skills.slash import parse_slash_skill_reference, resolve_slash_skill
-from deerflow.skills.storage import get_or_new_skill_storage
-from deerflow.skills.storage.skill_storage import SkillStorage
-from deerflow.skills.types import SKILL_MD_FILE
-from deerflow.utils.messages import get_original_user_content_text
-
-if TYPE_CHECKING:
-    from deerflow.config.app_config import AppConfig
-
-logger = logging.getLogger(__name__)
-
-_SLASH_SKILL_ACTIVATION_KEY = "slash_skill_activation"
-_SLASH_SKILL_ACTIVATION_TARGET_ID_KEY = "slash_skill_activation_target_id"
-_SUMMARY_MESSAGE_NAME = "summary"
-
-
-@dataclass(frozen=True, slots=True)
-class _Activation:
-    skill_name: str
-    category: str
-    container_file_path: str
-    skill_content: str
-    content_hash: str
-    remaining_text: str
-
-
-@dataclass(frozen=True, slots=True)
-class _ActivationResolution:
-    activation: _Activation | None = None
-    failure_message: str | None = None
-
-
-def is_slash_skill_activation_reminder(message: object) -> bool:
-    """Return whether a message is hidden slash-skill activation context."""
-    return isinstance(message, HumanMessage) and bool(message.additional_kwargs.get(_SLASH_SKILL_ACTIVATION_KEY))
-
-
-def _is_user_activation_target(message: object) -> bool:
-    if not isinstance(message, HumanMessage):
-        return False
-    if message.name == _SUMMARY_MESSAGE_NAME:
-        return False
-    if message.additional_kwargs.get("hide_from_ui"):
-        return False
-    return True
-
-
-class SkillActivationMiddleware(AgentMiddleware):
-    """Inject full SKILL.md content when the user explicitly types /skill-name."""
-
-    def __init__(
-        self,
-        *,
-        available_skills: set[str] | None = None,
-        app_config: AppConfig | None = None,
-    ) -> None:
-        super().__init__()
-        self._available_skills = set(available_skills) if available_skills is not None else None
-        self._app_config = app_config
-
-    def _storage(self) -> SkillStorage:
-        if self._app_config is not None:
-            return get_or_new_skill_storage(app_config=self._app_config)
-        return get_or_new_skill_storage()
-
-    @staticmethod
-    def _read_skill_content(skill_file: Path, skills_root: Path) -> str:
-        if skill_file.name != SKILL_MD_FILE:
-            raise ValueError(f"Expected {SKILL_MD_FILE}, got {skill_file.name}")
-        resolved_root = skills_root.resolve()
-        resolved_file = skill_file.resolve()
-        try:
-            resolved_file.relative_to(resolved_root)
-        except ValueError as exc:
-            raise ValueError("Resolved skill file must stay within the configured skills root.") from exc
-        if not resolved_file.is_file():
-            raise FileNotFoundError(resolved_file)
-        return resolved_file.read_text(encoding="utf-8")
-
-    def _resolve_activation(self, text: str) -> _ActivationResolution | None:
-        reference = parse_slash_skill_reference(text)
-        if reference is None:
-            return None
-
-        storage = self._storage()
-        skills = storage.load_skills(enabled_only=False)
-        skill = next((candidate for candidate in skills if candidate.name == reference.name), None)
-        if skill is None:
-            return _ActivationResolution(failure_message=f"Skill `/{reference.name}` is not installed.")
-        if not skill.enabled:
-            return _ActivationResolution(failure_message=f"Skill `/{reference.name}` is installed but disabled. Enable it before using slash activation.")
-        if self._available_skills is not None and reference.name not in self._available_skills:
-            return _ActivationResolution(failure_message=f"Skill `/{reference.name}` is not available for this agent.")
-
-        resolved = resolve_slash_skill(
-            text,
-            skills,
-            available_skills=self._available_skills,
-            container_base_path=storage.get_container_root(),
-        )
-        if resolved is None:
-            return _ActivationResolution(failure_message=f"Skill `/{reference.name}` could not be resolved.")
-
-        try:
-            skill_content = self._read_skill_content(resolved.skill.skill_file, storage.get_skills_root_path())
-        except (OSError, ValueError):
-            logger.exception("Failed to read slash-activated skill %s", resolved.skill.name)
-            return _ActivationResolution(failure_message=f"Skill `/{reference.name}` could not be loaded safely. Please check the skill installation.")
-
-        content_hash = hashlib.sha256(skill_content.encode("utf-8")).hexdigest()
-        return _ActivationResolution(
-            activation=_Activation(
-                skill_name=resolved.skill.name,
-                category=str(resolved.skill.category),
-                container_file_path=resolved.container_file_path,
-                skill_content=skill_content,
-                content_hash=content_hash,
-                remaining_text=resolved.remaining_text,
-            )
-        )
-
-    @staticmethod
-    def _build_activation_reminder(activation: _Activation) -> str:
-        user_request = activation.remaining_text or ("No additional task text was provided after the slash skill command. Ask the user what they want to do with this skill if the next step is unclear.")
-        escaped_user_request = html.escape(user_request, quote=False)
-        escaped_skill_content = html.escape(activation.skill_content, quote=False)
-        escaped_skill_name = html.escape(activation.skill_name, quote=True)
-        escaped_category = html.escape(activation.category, quote=True)
-        escaped_path = html.escape(activation.container_file_path, quote=True)
-        escaped_content_hash = html.escape(activation.content_hash, quote=True)
-        return f"""<slash_skill_activation>
-The user explicitly activated the `{activation.skill_name}` skill for this turn.
-Treat the task text as:
-<user_request>
-{escaped_user_request}
-</user_request>
-
-Follow this skill before choosing a general workflow. Load supporting resources from the same skill directory only when needed.
-
-<skill name="{escaped_skill_name}" category="{escaped_category}" path="{escaped_path}" sha256="{escaped_content_hash}">
-<skill_content encoding="xml-escaped">
-{escaped_skill_content}
-</skill_content>
-</skill>
-</slash_skill_activation>"""
-
-    @staticmethod
-    def _has_existing_activation_for_target(messages: list, target_index: int, target: HumanMessage) -> bool:
-        if target_index <= 0:
-            return False
-
-        if target.id:
-            for previous in messages[:target_index]:
-                if not is_slash_skill_activation_reminder(previous):
-                    continue
-                target_id = previous.additional_kwargs.get(_SLASH_SKILL_ACTIVATION_TARGET_ID_KEY)
-                if target_id == target.id or previous.id == f"{target.id}__slash_activation":
-                    return True
-
-        previous = messages[target_index - 1]
-        return is_slash_skill_activation_reminder(previous)
-
-    def _find_activation_target(self, messages: list) -> tuple[int, HumanMessage, _ActivationResolution] | None:
-        if not messages:
-            return None
-
-        target_index = next((idx for idx in range(len(messages) - 1, -1, -1) if _is_user_activation_target(messages[idx])), None)
-        if target_index is None:
-            return None
-
-        target = messages[target_index]
-        if target is None:
-            return None
-        if self._has_existing_activation_for_target(messages, target_index, target):
-            return None
-
-        content = get_original_user_content_text(target.content, target.additional_kwargs)
-        resolution = self._resolve_activation(content)
-        if resolution is None:
-            return None
-        return target_index, target, resolution
-
-    @staticmethod
-    def _record_activation(request: ModelRequest, activation: _Activation, *, hook: str) -> None:
-        runtime = getattr(request, "runtime", None)
-        context = getattr(runtime, "context", None)
-        journal = context.get("__run_journal") if isinstance(context, dict) else None
-        if journal is None:
-            return
-        try:
-            journal.record_middleware(
-                "skill_activation",
-                name="SkillActivationMiddleware",
-                hook=hook,
-                action="activate",
-                changes={
-                    "skill_name": activation.skill_name,
-                    "category": activation.category,
-                    "path": activation.container_file_path,
-                    "content_hash": activation.content_hash,
-                },
-            )
-        except Exception:
-            logger.debug("Failed to record slash skill activation audit event", exc_info=True)
-
-    def _prepare_model_request(self, request: ModelRequest, *, hook: str) -> ModelRequest | AIMessage | None:
-        target_and_resolution = self._find_activation_target(list(request.messages))
-        if target_and_resolution is None:
-            return None
-
-        target_index, target, resolution = target_and_resolution
-        if resolution.failure_message:
-            return AIMessage(content=resolution.failure_message)
-
-        activation = resolution.activation
-        if activation is None:
-            return None
-
-        logger.info(
-            "SkillActivationMiddleware: activating slash skill %s category=%s path=%s hash=%s",
-            activation.skill_name,
-            activation.category,
-            activation.container_file_path,
-            activation.content_hash,
-        )
-        self._record_activation(request, activation, hook=hook)
-        activation_msg = self._make_activation_message(target, self._build_activation_reminder(activation))
-        messages = list(request.messages)
-        messages.insert(target_index, activation_msg)
-        return request.override(messages=messages)
-
-    @staticmethod
-    def _make_activation_message(target: HumanMessage, activation_content: str) -> HumanMessage:
-        stable_id = target.id or str(uuid.uuid4())
-        additional_kwargs = {
-            "hide_from_ui": True,
-            _SLASH_SKILL_ACTIVATION_KEY: True,
-        }
-        if target.id:
-            additional_kwargs[_SLASH_SKILL_ACTIVATION_TARGET_ID_KEY] = target.id
-        return HumanMessage(
-            content=activation_content,
-            id=f"{stable_id}__slash_activation",
-            additional_kwargs=additional_kwargs,
-        )
-
-    @override
-    def wrap_model_call(
-        self,
-        request: ModelRequest,
-        handler: Callable[[ModelRequest], ModelResponse],
-    ) -> ModelResponse | AIMessage:
-        prepared = self._prepare_model_request(request, hook="wrap_model_call")
-        if prepared is None:
-            return handler(request)
-        if isinstance(prepared, AIMessage):
-            return prepared
-        return handler(prepared)
-
-    @override
-    async def awrap_model_call(
-        self,
-        request: ModelRequest,
-        handler: Callable[[ModelRequest], Awaitable[ModelResponse]],
-    ) -> ModelResponse | AIMessage:
-        prepared = await asyncio.to_thread(self._prepare_model_request, request, hook="awrap_model_call")
-        if prepared is None:
-            return await handler(request)
-        if isinstance(prepared, AIMessage):
-            return prepared
-        return await handler(prepared)

backend/packages/harness/deerflow/agents/middlewares/subagent_limit_middleware.py

@@ -7,7 +7,6 @@ from langchain.agents import AgentState
 from langchain.agents.middleware import AgentMiddleware
 from langgraph.runtime import Runtime
 
-from deerflow.agents.middlewares.tool_call_metadata import clone_ai_message_with_tool_calls
 from deerflow.subagents.executor import MAX_CONCURRENT_SUBAGENTS
 
 logger = logging.getLogger(__name__)
@@ -64,7 +63,7 @@ class SubagentLimitMiddleware(AgentMiddleware[AgentState]):
         logger.warning(f"Truncated {dropped_count} excess task tool call(s) from model response (limit: {self.max_concurrent})")
 
         # Replace the AIMessage with truncated tool_calls (same id triggers replacement)
-        updated_msg = clone_ai_message_with_tool_calls(last_msg, truncated_tool_calls)
+        updated_msg = last_msg.model_copy(update={"tool_calls": truncated_tool_calls})
         return {"messages": [updated_msg]}
 
     @override

backend/packages/harness/deerflow/agents/middlewares/summarization_middleware.py

@@ -9,15 +9,11 @@ from typing import Any, Protocol, override, runtime_checkable
 
 from langchain.agents import AgentState
 from langchain.agents.middleware import SummarizationMiddleware
-from langchain_core.messages import AIMessage, AnyMessage, HumanMessage, RemoveMessage, ToolMessage, get_buffer_string
+from langchain_core.messages import AIMessage, AnyMessage, HumanMessage, RemoveMessage, ToolMessage
 from langgraph.config import get_config
-from langgraph.constants import TAG_NOSTREAM
 from langgraph.graph.message import REMOVE_ALL_MESSAGES
 from langgraph.runtime import Runtime
 
-from deerflow.agents.middlewares.dynamic_context_middleware import is_dynamic_context_reminder
-from deerflow.agents.middlewares.tool_call_metadata import clone_ai_message_with_tool_calls
-
 logger = logging.getLogger(__name__)
 
 
@@ -82,7 +78,10 @@ def _clone_ai_message(
     content: Any | None = None,
 ) -> AIMessage:
     """Clone an AIMessage while replacing its tool_calls list and optional content."""
-    return clone_ai_message_with_tool_calls(message, tool_calls, content=content)
+    update: dict[str, Any] = {"tool_calls": tool_calls}
+    if content is not None:
+        update["content"] = content
+    return message.model_copy(update=update)
 
 
 @dataclass
@@ -117,74 +116,6 @@ class DeerFlowSummarizationMiddleware(SummarizationMiddleware):
         self._preserve_recent_skill_count = max(0, preserve_recent_skill_count)
         self._preserve_recent_skill_tokens = max(0, preserve_recent_skill_tokens)
         self._preserve_recent_skill_tokens_per_skill = max(0, preserve_recent_skill_tokens_per_skill)
-        # The summary LLM call runs inside a LangGraph middleware hook, so its token
-        # stream would otherwise be captured by the messages-tuple stream callback and
-        # broadcast to the frontend as a phantom AI message. Tag a dedicated model copy
-        # with TAG_NOSTREAM so the streaming handler skips it.
-        # Keep self.model untagged so the parent's profile / ls_params inspection still works.
-        #
-        # Preserve any tags already bound on the model (e.g. "middleware:summarize" set in
-        # lead_agent/agent.py for RunJournal attribution): RunnableBinding.with_config does a
-        # shallow merge that would otherwise overwrite the existing tags list entirely.
-        existing_tags = list((getattr(self.model, "config", None) or {}).get("tags") or [])
-        merged_tags = [*existing_tags, TAG_NOSTREAM] if TAG_NOSTREAM not in existing_tags else existing_tags
-        self._summary_model = self.model.with_config(tags=merged_tags)
-
-    @override
-    def _create_summary(self, messages_to_summarize: list[AnyMessage]) -> str:
-        return self._summarize_with(messages_to_summarize)
-
-    @override
-    async def _acreate_summary(self, messages_to_summarize: list[AnyMessage]) -> str:
-        return await self._asummarize_with(messages_to_summarize)
-
-    def _summarize_with(self, messages_to_summarize: list[AnyMessage]) -> str:
-        """Mirror the parent ``_create_summary`` but invoke the nostream-tagged model.
-
-        We do not swap ``self.model`` at the instance level: the agent/middleware is
-        cached and reused across concurrent runs, so a temporary swap would leak the
-        ``RunnableBinding`` to other coroutines during ``await`` and break parent logic
-        that inspects the raw model (``profile`` / ``_get_ls_params``).
-        """
-        if not messages_to_summarize:
-            return "No previous conversation history."
-        prompt = self._build_summary_prompt(messages_to_summarize)
-        if prompt is None:
-            return "Previous conversation was too long to summarize."
-        try:
-            response = self._summary_model.invoke(
-                prompt,
-                config={"metadata": {"lc_source": "summarization"}},
-            )
-            return response.text.strip()
-        except Exception as e:
-            return f"Error generating summary: {e!s}"
-
-    async def _asummarize_with(self, messages_to_summarize: list[AnyMessage]) -> str:
-        """Async counterpart of :meth:`_summarize_with` using the nostream model."""
-        if not messages_to_summarize:
-            return "No previous conversation history."
-        prompt = self._build_summary_prompt(messages_to_summarize)
-        if prompt is None:
-            return "Previous conversation was too long to summarize."
-        try:
-            response = await self._summary_model.ainvoke(
-                prompt,
-                config={"metadata": {"lc_source": "summarization"}},
-            )
-            return response.text.strip()
-        except Exception as e:
-            return f"Error generating summary: {e!s}"
-
-    def _build_summary_prompt(self, messages_to_summarize: list[AnyMessage]) -> str | None:
-        """Build the summary prompt, returning ``None`` when trimming leaves nothing."""
-        trimmed_messages = self._trim_messages_for_summary(messages_to_summarize)
-        if not trimmed_messages:
-            return None
-        # Format messages to avoid token inflation from metadata when str() is called on
-        # message objects.
-        formatted_messages = get_buffer_string(trimmed_messages)
-        return self.summary_prompt.format(messages=formatted_messages).rstrip()
 
     def before_model(self, state: AgentState, runtime: Runtime) -> dict | None:
         return self._maybe_summarize(state, runtime)
@@ -205,7 +136,6 @@ class DeerFlowSummarizationMiddleware(SummarizationMiddleware):
             return None
 
         messages_to_summarize, preserved_messages = self._partition_with_skill_rescue(messages, cutoff_index)
-        messages_to_summarize, preserved_messages = self._preserve_dynamic_context_reminders(messages_to_summarize, preserved_messages)
         self._fire_hooks(messages_to_summarize, preserved_messages, runtime)
         summary = self._create_summary(messages_to_summarize)
         new_messages = self._build_new_messages(summary)
@@ -231,7 +161,6 @@ class DeerFlowSummarizationMiddleware(SummarizationMiddleware):
             return None
 
         messages_to_summarize, preserved_messages = self._partition_with_skill_rescue(messages, cutoff_index)
-        messages_to_summarize, preserved_messages = self._preserve_dynamic_context_reminders(messages_to_summarize, preserved_messages)
         self._fire_hooks(messages_to_summarize, preserved_messages, runtime)
         summary = await self._acreate_summary(messages_to_summarize)
         new_messages = self._build_new_messages(summary)
@@ -251,24 +180,6 @@ class DeerFlowSummarizationMiddleware(SummarizationMiddleware):
         """
         return [HumanMessage(content=f"Here is a summary of the conversation to date:\n\n{summary}", name="summary")]
 
-    def _preserve_dynamic_context_reminders(
-        self,
-        messages_to_summarize: list[AnyMessage],
-        preserved_messages: list[AnyMessage],
-    ) -> tuple[list[AnyMessage], list[AnyMessage]]:
-        """Keep hidden dynamic-context reminders out of summary compression.
-
-        These reminders carry the current date and optional memory. If summarization
-        removes them, DynamicContextMiddleware can mistake the summary HumanMessage
-        for the first user message and inject the reminder in the wrong place.
-        """
-        reminders = [msg for msg in messages_to_summarize if is_dynamic_context_reminder(msg)]
-        if not reminders:
-            return messages_to_summarize, preserved_messages
-
-        remaining = [msg for msg in messages_to_summarize if not is_dynamic_context_reminder(msg)]
-        return remaining, reminders + preserved_messages
-
     def _partition_with_skill_rescue(
         self,
         messages: list[AnyMessage],

backend/packages/harness/deerflow/agents/middlewares/title_middleware.py

@@ -9,7 +9,6 @@ from langchain.agents.middleware import AgentMiddleware
 from langgraph.config import get_config
 from langgraph.runtime import Runtime
 
-from deerflow.agents.middlewares.dynamic_context_middleware import is_dynamic_context_reminder
 from deerflow.config.title_config import get_title_config
 from deerflow.models import create_chat_model
 
@@ -62,10 +61,6 @@ class TitleMiddleware(AgentMiddleware[TitleMiddlewareState]):
 
         return ""
 
-    @staticmethod
-    def _is_user_message_for_title(message: object) -> bool:
-        return getattr(message, "type", None) == "human" and not is_dynamic_context_reminder(message)
-
     def _should_generate_title(self, state: TitleMiddlewareState) -> bool:
         """Check if we should generate a title for this thread."""
         config = self._get_title_config()
@@ -82,7 +77,7 @@ class TitleMiddleware(AgentMiddleware[TitleMiddlewareState]):
             return False
 
         # Count user and assistant messages
-        user_messages = [m for m in messages if self._is_user_message_for_title(m)]
+        user_messages = [m for m in messages if m.type == "human"]
         assistant_messages = [m for m in messages if m.type == "ai"]
 
         # Generate title after first complete exchange
@@ -96,7 +91,7 @@ class TitleMiddleware(AgentMiddleware[TitleMiddlewareState]):
         config = self._get_title_config()
         messages = state.get("messages", [])
 
-        user_msg_content = next((m.content for m in messages if self._is_user_message_for_title(m)), "")
+        user_msg_content = next((m.content for m in messages if m.type == "human"), "")
         assistant_msg_content = next((m.content for m in messages if m.type == "ai"), "")
 
         user_msg = self._normalize_content(user_msg_content)
@@ -160,11 +155,7 @@ class TitleMiddleware(AgentMiddleware[TitleMiddlewareState]):
         prompt, user_msg = self._build_title_prompt(state)
 
         try:
-            # attach_tracing=False because ``_get_runnable_config()`` inherits
+            model_kwargs = {"thinking_enabled": False}
-            # the graph-level RunnableConfig (set in ``_make_lead_agent``) whose
-            # callbacks already carry tracing handlers; binding them again at
-            # the model level would emit duplicate spans.
-            model_kwargs = {"thinking_enabled": False, "attach_tracing": False}
             if self._app_config is not None:
                 model_kwargs["app_config"] = self._app_config
             if config.model_name:

backend/packages/harness/deerflow/agents/middlewares/todo_middleware.py

@@ -7,26 +7,20 @@ reminder message so the model still knows about the outstanding todo list.
 
 Additionally, this middleware prevents the agent from exiting the loop while
 there are still incomplete todo items. When the model produces a final response
-(no tool calls) but todos are not yet complete, the middleware queues a reminder
+(no tool calls) but todos are not yet complete, the middleware injects a reminder
-for the next model request and jumps back to the model node to force continued
+and jumps back to the model node to force continued engagement.
-engagement. The completion reminder is injected via ``wrap_model_call`` instead
-of being persisted into graph state as a normal user-visible message.
 """
 
 from __future__ import annotations
 
-import threading
-from collections.abc import Awaitable, Callable
 from typing import Any, override
 
 from langchain.agents.middleware import TodoListMiddleware
-from langchain.agents.middleware.todo import Todo
+from langchain.agents.middleware.todo import PlanningState, Todo
-from langchain.agents.middleware.types import ModelCallResult, ModelRequest, ModelResponse, hook_config
+from langchain.agents.middleware.types import hook_config
 from langchain_core.messages import AIMessage, HumanMessage
 from langgraph.runtime import Runtime
 
-from deerflow.agents.thread_state import ThreadState
-
 
 def _todos_in_messages(messages: list[Any]) -> bool:
     """Return True if any AIMessage in *messages* contains a write_todos tool call."""
@@ -61,51 +55,6 @@ def _format_todos(todos: list[Todo]) -> str:
     return "\n".join(lines)
 
 
-def _format_completion_reminder(todos: list[Todo]) -> str:
-    """Format a completion reminder for incomplete todo items."""
-    incomplete = [t for t in todos if t.get("status") != "completed"]
-    incomplete_text = "\n".join(f"- [{t.get('status', 'pending')}] {t.get('content', '')}" for t in incomplete)
-    return (
-        "<system_reminder>\n"
-        "You have incomplete todo items that must be finished before giving your final response:\n\n"
-        f"{incomplete_text}\n\n"
-        "Please continue working on these tasks. Call `write_todos` to mark items as completed "
-        "as you finish them, and only respond when all items are done.\n"
-        "</system_reminder>"
-    )
-
-
-_TOOL_CALL_FINISH_REASONS = {"tool_calls", "function_call"}
-
-
-def _has_tool_call_intent_or_error(message: AIMessage) -> bool:
-    """Return True when an AIMessage is not a clean final answer.
-
-    Todo completion reminders should only fire when the model has produced a
-    plain final response. Provider/tool parsing details have moved across
-    LangChain versions and integrations, so keep all tool-intent/error signals
-    behind this helper instead of checking one concrete field at the call site.
-    """
-    if message.tool_calls:
-        return True
-
-    if getattr(message, "invalid_tool_calls", None):
-        return True
-
-    # Backward/provider compatibility: some integrations preserve raw or legacy
-    # tool-call intent in additional_kwargs even when structured tool_calls is
-    # empty. If this helper changes, update the matching sentinel test
-    # `TestToolCallIntentOrError.test_langchain_ai_message_tool_fields_are_explicitly_handled`;
-    # if that test fails after a LangChain upgrade, review this helper so new
-    # tool-call/error fields are not silently treated as clean final answers.
-    additional_kwargs = getattr(message, "additional_kwargs", {}) or {}
-    if additional_kwargs.get("tool_calls") or additional_kwargs.get("function_call"):
-        return True
-
-    response_metadata = getattr(message, "response_metadata", {}) or {}
-    return response_metadata.get("finish_reason") in _TOOL_CALL_FINISH_REASONS
-
-
 class TodoMiddleware(TodoListMiddleware):
     """Extends TodoListMiddleware with `write_todos` context-loss detection.
 
@@ -115,12 +64,10 @@ class TodoMiddleware(TodoListMiddleware):
     and injects a reminder message so the model can continue tracking progress.
     """
 
-    state_schema = ThreadState
-
     @override
     def before_model(
         self,
-        state: ThreadState,
+        state: PlanningState,
         runtime: Runtime,
     ) -> dict[str, Any] | None:
         """Inject a todo-list reminder when write_todos has left the context window."""
@@ -142,7 +89,6 @@ class TodoMiddleware(TodoListMiddleware):
         formatted = _format_todos(todos)
         reminder = HumanMessage(
             name="todo_reminder",
-            additional_kwargs={"hide_from_ui": True},
             content=(
                 "<system_reminder>\n"
                 "Your todo list from earlier is no longer visible in the current context window, "
@@ -158,7 +104,7 @@ class TodoMiddleware(TodoListMiddleware):
     @override
     async def abefore_model(
         self,
-        state: ThreadState,
+        state: PlanningState,
         runtime: Runtime,
     ) -> dict[str, Any] | None:
         """Async version of before_model."""
@@ -167,106 +113,12 @@ class TodoMiddleware(TodoListMiddleware):
     # Maximum number of completion reminders before allowing the agent to exit.
     # This prevents infinite loops when the agent cannot make further progress.
     _MAX_COMPLETION_REMINDERS = 2
-    # Hard cap for per-run reminder bookkeeping in long-lived middleware instances.
-    _MAX_COMPLETION_REMINDER_KEYS = 4096
-
-    def __init__(self, *args: Any, **kwargs: Any) -> None:
-        super().__init__(*args, **kwargs)
-        self._lock = threading.Lock()
-        self._pending_completion_reminders: dict[tuple[str, str], list[str]] = {}
-        self._completion_reminder_counts: dict[tuple[str, str], int] = {}
-        self._completion_reminder_touch_order: dict[tuple[str, str], int] = {}
-        self._completion_reminder_next_order = 0
-
-    @staticmethod
-    def _get_thread_id(runtime: Runtime) -> str:
-        context = getattr(runtime, "context", None)
-        thread_id = context.get("thread_id") if context else None
-        return str(thread_id) if thread_id else "default"
-
-    @staticmethod
-    def _get_run_id(runtime: Runtime) -> str:
-        context = getattr(runtime, "context", None)
-        run_id = context.get("run_id") if context else None
-        return str(run_id) if run_id else "default"
-
-    def _pending_key(self, runtime: Runtime) -> tuple[str, str]:
-        return self._get_thread_id(runtime), self._get_run_id(runtime)
-
-    def _touch_completion_reminder_key_locked(self, key: tuple[str, str]) -> None:
-        self._completion_reminder_next_order += 1
-        self._completion_reminder_touch_order[key] = self._completion_reminder_next_order
-
-    def _completion_reminder_keys_locked(self) -> set[tuple[str, str]]:
-        keys = set(self._pending_completion_reminders)
-        keys.update(self._completion_reminder_counts)
-        keys.update(self._completion_reminder_touch_order)
-        return keys
-
-    def _drop_completion_reminder_key_locked(self, key: tuple[str, str]) -> None:
-        self._pending_completion_reminders.pop(key, None)
-        self._completion_reminder_counts.pop(key, None)
-        self._completion_reminder_touch_order.pop(key, None)
-
-    def _prune_completion_reminder_state_locked(self, protected_key: tuple[str, str]) -> None:
-        keys = self._completion_reminder_keys_locked()
-        overflow = len(keys) - self._MAX_COMPLETION_REMINDER_KEYS
-        if overflow <= 0:
-            return
-
-        candidates = [key for key in keys if key != protected_key]
-        candidates.sort(key=lambda key: self._completion_reminder_touch_order.get(key, 0))
-        for key in candidates[:overflow]:
-            self._drop_completion_reminder_key_locked(key)
-
-    def _queue_completion_reminder(self, runtime: Runtime, reminder: str) -> None:
-        key = self._pending_key(runtime)
-        with self._lock:
-            self._pending_completion_reminders.setdefault(key, []).append(reminder)
-            self._completion_reminder_counts[key] = self._completion_reminder_counts.get(key, 0) + 1
-            self._touch_completion_reminder_key_locked(key)
-            self._prune_completion_reminder_state_locked(protected_key=key)
-
-    def _completion_reminder_count_for_runtime(self, runtime: Runtime) -> int:
-        key = self._pending_key(runtime)
-        with self._lock:
-            return self._completion_reminder_counts.get(key, 0)
-
-    def _drain_completion_reminders(self, runtime: Runtime) -> list[str]:
-        key = self._pending_key(runtime)
-        with self._lock:
-            reminders = self._pending_completion_reminders.pop(key, [])
-            if reminders or key in self._completion_reminder_counts:
-                self._touch_completion_reminder_key_locked(key)
-            return reminders
-
-    def _clear_other_run_completion_reminders(self, runtime: Runtime) -> None:
-        thread_id, current_run_id = self._pending_key(runtime)
-        with self._lock:
-            for key in self._completion_reminder_keys_locked():
-                if key[0] == thread_id and key[1] != current_run_id:
-                    self._drop_completion_reminder_key_locked(key)
-
-    def _clear_current_run_completion_reminders(self, runtime: Runtime) -> None:
-        key = self._pending_key(runtime)
-        with self._lock:
-            self._drop_completion_reminder_key_locked(key)
-
-    @override
-    def before_agent(self, state: ThreadState, runtime: Runtime) -> dict[str, Any] | None:
-        self._clear_other_run_completion_reminders(runtime)
-        return None
-
-    @override
-    async def abefore_agent(self, state: ThreadState, runtime: Runtime) -> dict[str, Any] | None:
-        self._clear_other_run_completion_reminders(runtime)
-        return None
 
     @hook_config(can_jump_to=["model"])
     @override
     def after_model(
         self,
-        state: ThreadState,
+        state: PlanningState,
         runtime: Runtime,
     ) -> dict[str, Any] | None:
         """Prevent premature agent exit when todo items are still incomplete.
@@ -285,12 +137,10 @@ class TodoMiddleware(TodoListMiddleware):
         if base_result is not None:
             return base_result
 
-        # 2. Only intervene when the agent wants to exit cleanly. Tool-call
+        # 2. Only intervene when the agent wants to exit (no tool calls).
-        # intent or tool-call parse errors should be handled by the tool path
-        # instead of being masked by todo reminders.
         messages = state.get("messages") or []
         last_ai = next((m for m in reversed(messages) if isinstance(m, AIMessage)), None)
-        if not last_ai or _has_tool_call_intent_or_error(last_ai):
+        if not last_ai or last_ai.tool_calls:
             return None
 
         # 3. Allow exit when all todos are completed or there are no todos.
@@ -299,65 +149,31 @@ class TodoMiddleware(TodoListMiddleware):
             return None
 
         # 4. Enforce a reminder cap to prevent infinite re-engagement loops.
-        if self._completion_reminder_count_for_runtime(runtime) >= self._MAX_COMPLETION_REMINDERS:
+        if _completion_reminder_count(messages) >= self._MAX_COMPLETION_REMINDERS:
             return None
 
-        # 5. Queue a reminder for the next model request and jump back. We must
+        # 5. Inject a reminder and force the agent back to the model.
-        # not persist this control prompt as a normal HumanMessage, otherwise it
+        incomplete = [t for t in todos if t.get("status") != "completed"]
-        # can leak into user-visible message streams and saved transcripts.
+        incomplete_text = "\n".join(f"- [{t.get('status', 'pending')}] {t.get('content', '')}" for t in incomplete)
-        self._queue_completion_reminder(runtime, _format_completion_reminder(todos))
+        reminder = HumanMessage(
-        return {"jump_to": "model"}
+            name="todo_completion_reminder",
+            content=(
+                "<system_reminder>\n"
+                "You have incomplete todo items that must be finished before giving your final response:\n\n"
+                f"{incomplete_text}\n\n"
+                "Please continue working on these tasks. Call `write_todos` to mark items as completed "
+                "as you finish them, and only respond when all items are done.\n"
+                "</system_reminder>"
+            ),
+        )
+        return {"jump_to": "model", "messages": [reminder]}
 
     @override
     @hook_config(can_jump_to=["model"])
     async def aafter_model(
         self,
-        state: ThreadState,
+        state: PlanningState,
         runtime: Runtime,
     ) -> dict[str, Any] | None:
         """Async version of after_model."""
         return self.after_model(state, runtime)
-
-    @staticmethod
-    def _format_pending_completion_reminders(reminders: list[str]) -> str:
-        return "\n\n".join(dict.fromkeys(reminders))
-
-    def _augment_request(self, request: ModelRequest) -> ModelRequest:
-        reminders = self._drain_completion_reminders(request.runtime)
-        if not reminders:
-            return request
-        new_messages = [
-            *request.messages,
-            HumanMessage(
-                content=self._format_pending_completion_reminders(reminders),
-                name="todo_completion_reminder",
-                additional_kwargs={"hide_from_ui": True},
-            ),
-        ]
-        return request.override(messages=new_messages)
-
-    @override
-    def wrap_model_call(
-        self,
-        request: ModelRequest,
-        handler: Callable[[ModelRequest], ModelResponse],
-    ) -> ModelCallResult:
-        return handler(self._augment_request(request))
-
-    @override
-    async def awrap_model_call(
-        self,
-        request: ModelRequest,
-        handler: Callable[[ModelRequest], Awaitable[ModelResponse]],
-    ) -> ModelCallResult:
-        return await handler(self._augment_request(request))
-
-    @override
-    def after_agent(self, state: ThreadState, runtime: Runtime) -> dict[str, Any] | None:
-        self._clear_current_run_completion_reminders(runtime)
-        return None
-
-    @override
-    async def aafter_agent(self, state: ThreadState, runtime: Runtime) -> dict[str, Any] | None:
-        self._clear_current_run_completion_reminders(runtime)
-        return None

backend/packages/harness/deerflow/agents/middlewares/token_usage_middleware.py

@@ -1,358 +1,37 @@
-"""Middleware for logging token usage and annotating step attribution."""
+"""Middleware for logging LLM token usage."""
-
-from __future__ import annotations
 
 import logging
-from collections import defaultdict
+from typing import override
-from typing import Any, override
 
 from langchain.agents import AgentState
 from langchain.agents.middleware import AgentMiddleware
-from langchain.agents.middleware.todo import Todo
-from langchain_core.messages import AIMessage, ToolMessage
 from langgraph.runtime import Runtime
 
 logger = logging.getLogger(__name__)
 
-TOKEN_USAGE_ATTRIBUTION_KEY = "token_usage_attribution"
-
-
-def _string_arg(value: Any) -> str | None:
-    if isinstance(value, str):
-        normalized = value.strip()
-        return normalized or None
-    return None
-
-
-def _normalize_todos(value: Any) -> list[Todo]:
-    if not isinstance(value, list):
-        return []
-
-    normalized: list[Todo] = []
-    for item in value:
-        if not isinstance(item, dict):
-            continue
-
-        todo: Todo = {}
-        content = _string_arg(item.get("content"))
-        status = item.get("status")
-
-        if content is not None:
-            todo["content"] = content
-        if status in {"pending", "in_progress", "completed"}:
-            todo["status"] = status
-
-        normalized.append(todo)
-
-    return normalized
-
-
-def _todo_action_kind(previous: Todo | None, current: Todo) -> str:
-    status = current.get("status")
-    previous_content = previous.get("content") if previous else None
-    current_content = current.get("content")
-
-    if previous is None:
-        if status == "completed":
-            return "todo_complete"
-        if status == "in_progress":
-            return "todo_start"
-        return "todo_update"
-
-    if previous_content != current_content:
-        return "todo_update"
-
-    if status == "completed":
-        return "todo_complete"
-    if status == "in_progress":
-        return "todo_start"
-    return "todo_update"
-
-
-def _build_todo_actions(previous_todos: list[Todo], next_todos: list[Todo]) -> list[dict[str, Any]]:
-    # This is the single source of truth for precise write_todos token
-    # attribution. The frontend intentionally falls back to a generic
-    # "Update to-do list" label when this metadata is missing or malformed.
-    previous_by_content: dict[str, list[tuple[int, Todo]]] = defaultdict(list)
-    matched_previous_indices: set[int] = set()
-
-    for index, todo in enumerate(previous_todos):
-        content = todo.get("content")
-        if isinstance(content, str) and content:
-            previous_by_content[content].append((index, todo))
-
-    actions: list[dict[str, Any]] = []
-
-    for index, todo in enumerate(next_todos):
-        content = todo.get("content")
-        if not isinstance(content, str) or not content:
-            continue
-
-        previous_match: Todo | None = None
-        content_matches = previous_by_content.get(content)
-        if content_matches:
-            while content_matches and content_matches[0][0] in matched_previous_indices:
-                content_matches.pop(0)
-            if content_matches:
-                previous_index, previous_match = content_matches.pop(0)
-                matched_previous_indices.add(previous_index)
-
-        if previous_match is None and index < len(previous_todos) and index not in matched_previous_indices:
-            previous_match = previous_todos[index]
-            matched_previous_indices.add(index)
-
-        if previous_match is not None:
-            previous_content = previous_match.get("content")
-            previous_status = previous_match.get("status")
-            if previous_content == content and previous_status == todo.get("status"):
-                continue
-
-        actions.append(
-            {
-                "kind": _todo_action_kind(previous_match, todo),
-                "content": content,
-            }
-        )
-
-    for index, todo in enumerate(previous_todos):
-        if index in matched_previous_indices:
-            continue
-
-        content = todo.get("content")
-        if not isinstance(content, str) or not content:
-            continue
-
-        actions.append(
-            {
-                "kind": "todo_remove",
-                "content": content,
-            }
-        )
-
-    return actions
-
-
-def _describe_tool_call(tool_call: dict[str, Any], todos: list[Todo]) -> list[dict[str, Any]]:
-    name = _string_arg(tool_call.get("name")) or "unknown"
-    args = tool_call.get("args") if isinstance(tool_call.get("args"), dict) else {}
-    tool_call_id = _string_arg(tool_call.get("id"))
-
-    if name == "write_todos":
-        next_todos = _normalize_todos(args.get("todos"))
-        actions = _build_todo_actions(todos, next_todos)
-        if not actions:
-            return [
-                {
-                    "kind": "tool",
-                    "tool_name": name,
-                    "tool_call_id": tool_call_id,
-                }
-            ]
-        return [
-            {
-                **action,
-                "tool_call_id": tool_call_id,
-            }
-            for action in actions
-        ]
-
-    if name == "task":
-        return [
-            {
-                "kind": "subagent",
-                "description": _string_arg(args.get("description")),
-                "subagent_type": _string_arg(args.get("subagent_type")),
-                "tool_call_id": tool_call_id,
-            }
-        ]
-
-    if name in {"web_search", "image_search"}:
-        query = _string_arg(args.get("query"))
-        return [
-            {
-                "kind": "search",
-                "tool_name": name,
-                "query": query,
-                "tool_call_id": tool_call_id,
-            }
-        ]
-
-    if name == "present_files":
-        return [
-            {
-                "kind": "present_files",
-                "tool_call_id": tool_call_id,
-            }
-        ]
-
-    if name == "ask_clarification":
-        return [
-            {
-                "kind": "clarification",
-                "tool_call_id": tool_call_id,
-            }
-        ]
-
-    return [
-        {
-            "kind": "tool",
-            "tool_name": name,
-            "description": _string_arg(args.get("description")),
-            "tool_call_id": tool_call_id,
-        }
-    ]
-
-
-def _infer_step_kind(message: AIMessage, actions: list[dict[str, Any]]) -> str:
-    if actions:
-        first_kind = actions[0].get("kind")
-        if len(actions) == 1 and first_kind in {"todo_start", "todo_complete", "todo_update", "todo_remove"}:
-            return "todo_update"
-        if len(actions) == 1 and first_kind == "subagent":
-            return "subagent_dispatch"
-        return "tool_batch"
-
-    if message.content:
-        return "final_answer"
-    return "thinking"
-
-
-def _has_tool_call(message: AIMessage, tool_call_id: str) -> bool:
-    """Return True if the AIMessage contains a tool_call with the given id."""
-    for tc in message.tool_calls or []:
-        if isinstance(tc, dict):
-            if tc.get("id") == tool_call_id:
-                return True
-        elif hasattr(tc, "id") and tc.id == tool_call_id:
-            return True
-    return False
-
-
-def _build_attribution(message: AIMessage, todos: list[Todo]) -> dict[str, Any]:
-    tool_calls = getattr(message, "tool_calls", None) or []
-    actions: list[dict[str, Any]] = []
-    current_todos = list(todos)
-
-    for raw_tool_call in tool_calls:
-        if not isinstance(raw_tool_call, dict):
-            continue
-
-        described_actions = _describe_tool_call(raw_tool_call, current_todos)
-        actions.extend(described_actions)
-
-        if raw_tool_call.get("name") == "write_todos":
-            args = raw_tool_call.get("args") if isinstance(raw_tool_call.get("args"), dict) else {}
-            current_todos = _normalize_todos(args.get("todos"))
-
-    tool_call_ids: list[str] = []
-    for tool_call in tool_calls:
-        if not isinstance(tool_call, dict):
-            continue
-
-        tool_call_id = _string_arg(tool_call.get("id"))
-        if tool_call_id is not None:
-            tool_call_ids.append(tool_call_id)
-
-    return {
-        # Schema changes should remain additive where possible so older
-        # frontends can ignore unknown fields and fall back safely.
-        "version": 1,
-        "kind": _infer_step_kind(message, actions),
-        "shared_attribution": len(actions) > 1,
-        "tool_call_ids": tool_call_ids,
-        "actions": actions,
-    }
-
 
 class TokenUsageMiddleware(AgentMiddleware):
-    """Logs token usage from model responses and annotates the AI step."""
+    """Logs token usage from model response usage_metadata."""
-
-    def _apply(self, state: AgentState) -> dict | None:
-        messages = state.get("messages", [])
-        if not messages:
-            return None
-
-        # Annotate subagent token usage onto the AIMessage that dispatched it.
-        # When a task tool completes, its usage is cached by tool_call_id.  Detect
-        # the ToolMessage → search backward for the corresponding AIMessage → merge.
-        # Walk backward through consecutive ToolMessages before the new AIMessage
-        # so that multiple concurrent task tool calls all get their subagent tokens
-        # written back to the same dispatch message (merging into one update).
-        state_updates: dict[int, AIMessage] = {}
-        if len(messages) >= 2:
-            from deerflow.tools.builtins.task_tool import pop_cached_subagent_usage
-
-            idx = len(messages) - 2
-            while idx >= 0:
-                tool_msg = messages[idx]
-                if not isinstance(tool_msg, ToolMessage) or not tool_msg.tool_call_id:
-                    break
-
-                subagent_usage = pop_cached_subagent_usage(tool_msg.tool_call_id)
-                if subagent_usage:
-                    # Search backward from the ToolMessage to find the AIMessage
-                    # that dispatched it.  A single model response can dispatch
-                    # multiple task tool calls, so we can't assume a fixed offset.
-                    dispatch_idx = idx - 1
-                    while dispatch_idx >= 0:
-                        candidate = messages[dispatch_idx]
-                        if isinstance(candidate, AIMessage) and _has_tool_call(candidate, tool_msg.tool_call_id):
-                            # Accumulate into an existing update for the same
-                            # AIMessage (multiple task calls in one response),
-                            # or merge fresh from the original message.
-                            existing_update = state_updates.get(dispatch_idx)
-                            prev = existing_update.usage_metadata if existing_update else (getattr(candidate, "usage_metadata", None) or {})
-                            merged = {
-                                **prev,
-                                "input_tokens": prev.get("input_tokens", 0) + subagent_usage["input_tokens"],
-                                "output_tokens": prev.get("output_tokens", 0) + subagent_usage["output_tokens"],
-                                "total_tokens": prev.get("total_tokens", 0) + subagent_usage["total_tokens"],
-                            }
-                            state_updates[dispatch_idx] = candidate.model_copy(update={"usage_metadata": merged})
-                            break
-                        dispatch_idx -= 1
-                idx -= 1
-
-        last = messages[-1]
-        if not isinstance(last, AIMessage):
-            if state_updates:
-                return {"messages": [state_updates[idx] for idx in sorted(state_updates)]}
-            return None
-
-        usage = getattr(last, "usage_metadata", None)
-        if usage:
-            input_token_details = usage.get("input_token_details") or {}
-            output_token_details = usage.get("output_token_details") or {}
-            detail_parts = []
-            if input_token_details:
-                detail_parts.append(f"input_token_details={input_token_details}")
-            if output_token_details:
-                detail_parts.append(f"output_token_details={output_token_details}")
-            detail_suffix = f" {' '.join(detail_parts)}" if detail_parts else ""
-            logger.info(
-                "LLM token usage: input=%s output=%s total=%s%s",
-                usage.get("input_tokens", "?"),
-                usage.get("output_tokens", "?"),
-                usage.get("total_tokens", "?"),
-                detail_suffix,
-            )
-
-        todos = state.get("todos") or []
-        attribution = _build_attribution(last, todos if isinstance(todos, list) else [])
-        additional_kwargs = dict(getattr(last, "additional_kwargs", {}) or {})
-
-        if additional_kwargs.get(TOKEN_USAGE_ATTRIBUTION_KEY) == attribution:
-            return {"messages": [state_updates[idx] for idx in sorted(state_updates)]} if state_updates else None
-
-        additional_kwargs[TOKEN_USAGE_ATTRIBUTION_KEY] = attribution
-        updated_msg = last.model_copy(update={"additional_kwargs": additional_kwargs})
-        state_updates[len(messages) - 1] = updated_msg
-        return {"messages": [state_updates[idx] for idx in sorted(state_updates)]}
 
     @override
     def after_model(self, state: AgentState, runtime: Runtime) -> dict | None:
-        return self._apply(state)
+        return self._log_usage(state)
 
     @override
     async def aafter_model(self, state: AgentState, runtime: Runtime) -> dict | None:
-        return self._apply(state)
+        return self._log_usage(state)
+
+    def _log_usage(self, state: AgentState) -> None:
+        messages = state.get("messages", [])
+        if not messages:
+            return None
+        last = messages[-1]
+        usage = getattr(last, "usage_metadata", None)
+        if usage:
+            logger.info(
+                "LLM token usage: input=%s output=%s total=%s",
+                usage.get("input_tokens", "?"),
+                usage.get("output_tokens", "?"),
+                usage.get("total_tokens", "?"),
+            )
+        return None

backend/packages/harness/deerflow/agents/middlewares/tool_call_metadata.py

@@ -1,50 +0,0 @@
-"""Helpers for keeping AIMessage tool-call metadata consistent."""
-
-from __future__ import annotations
-
-from typing import Any
-
-from langchain_core.messages import AIMessage
-
-
-def _raw_tool_call_id(raw_tool_call: Any) -> str | None:
-    if not isinstance(raw_tool_call, dict):
-        return None
-
-    raw_id = raw_tool_call.get("id")
-    return raw_id if isinstance(raw_id, str) and raw_id else None
-
-
-def clone_ai_message_with_tool_calls(
-    message: AIMessage,
-    tool_calls: list[dict[str, Any]],
-    *,
-    content: Any | None = None,
-) -> AIMessage:
-    """Clone an AIMessage while keeping raw provider tool-call metadata in sync."""
-    kept_ids = {tc["id"] for tc in tool_calls if isinstance(tc.get("id"), str) and tc["id"]}
-
-    update: dict[str, Any] = {"tool_calls": tool_calls}
-    if content is not None:
-        update["content"] = content
-
-    additional_kwargs = dict(getattr(message, "additional_kwargs", {}) or {})
-    raw_tool_calls = additional_kwargs.get("tool_calls")
-    if isinstance(raw_tool_calls, list):
-        synced_raw_tool_calls = [raw_tc for raw_tc in raw_tool_calls if _raw_tool_call_id(raw_tc) in kept_ids]
-        if synced_raw_tool_calls:
-            additional_kwargs["tool_calls"] = synced_raw_tool_calls
-        else:
-            additional_kwargs.pop("tool_calls", None)
-
-    if not tool_calls:
-        additional_kwargs.pop("function_call", None)
-
-    update["additional_kwargs"] = additional_kwargs
-
-    response_metadata = dict(getattr(message, "response_metadata", {}) or {})
-    if not tool_calls and response_metadata.get("finish_reason") == "tool_calls":
-        response_metadata["finish_reason"] = "stop"
-    update["response_metadata"] = response_metadata
-
-    return message.model_copy(update=update)