Address Copilot IM channel feedback

The default `make up` started the Gateway with `--workers 4`, but run state
(RunManager and the stream bridge) is held in-process and nginx uses no sticky
sessions. With the default config, same-run requests scatter across workers that
each keep their own run state, breaking run cancellation (409), SSE reconnect
(hangs on heartbeats), multitask de-duplication, and IM channels (duplicate
replies). The shared cross-worker stream bridge does not exist yet.

Default GATEWAY_WORKERS to 1 so the out-of-the-box deployment is correct,
document the single-worker boundary in the README, and add a regression test
pinning the default while keeping it overridable. This is a stop-gap, not a
multi-worker implementation; the full fix (shared run state + stream bridge) is
tracked in #3191.

Refs #3239, #3260

.env.example

@@ -21,6 +21,7 @@ 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

.github/labeler.yml

@@ -1,72 +0,0 @@
-# Path-based PR auto-labeling config for actions/labeler@v5.
-# Each key is a label (must exist — see .github/labels.yml); the globs decide
-# when it is applied. A PR can match several areas, which is expected.
-
-"area:frontend":
-  - changed-files:
-      - any-glob-to-any-file:
-          - "frontend/**"
-
-"area:backend":
-  - changed-files:
-      - any-glob-to-any-file:
-          - "backend/app/**"
-          - "backend/packages/harness/deerflow/runtime/**"
-          - "backend/packages/harness/deerflow/persistence/**"
-          - "backend/packages/harness/deerflow/config/**"
-          - "backend/packages/harness/deerflow/tools/**"
-          - "backend/packages/harness/deerflow/guardrails/**"
-          - "backend/packages/harness/deerflow/tracing/**"
-          - "backend/packages/harness/deerflow/models/**"
-          - "backend/packages/harness/deerflow/utils/**"
-          - "backend/packages/harness/deerflow/uploads/**"
-
-"area:agents":
-  - changed-files:
-      - any-glob-to-any-file:
-          - "backend/packages/harness/deerflow/agents/**"
-          - "backend/packages/harness/deerflow/subagents/**"
-          - "backend/packages/harness/deerflow/reflection/**"
-          - "backend/langgraph.json"
-          - "backend/**/prompts/**"
-
-"area:sandbox":
-  - changed-files:
-      - any-glob-to-any-file:
-          - "docker/**"
-          - "backend/packages/harness/deerflow/sandbox/**"
-          - "backend/Dockerfile"
-          - "frontend/Dockerfile"
-
-"area:skills":
-  - changed-files:
-      - any-glob-to-any-file:
-          - "skills/**"
-          - "backend/packages/harness/deerflow/skills/**"
-          - "frontend/src/core/skills/**"
-
-"area:mcp":
-  - changed-files:
-      - any-glob-to-any-file:
-          - "backend/packages/harness/deerflow/mcp/**"
-          - "frontend/src/core/mcp/**"
-
-"area:ci":
-  - changed-files:
-      - any-glob-to-any-file:
-          - ".github/**"
-          - "scripts/**"
-
-"area:docs":
-  - changed-files:
-      - any-glob-to-any-file:
-          - "docs/**"
-          - "**/*.md"
-
-"area:deps":
-  - changed-files:
-      - any-glob-to-any-file:
-          - "backend/pyproject.toml"
-          - "backend/uv.lock"
-          - "frontend/package.json"
-          - "frontend/pnpm-lock.yaml"

.github/workflows/issue-triage.yml

@@ -1,44 +0,0 @@
-name: Issue Triage
-
-# Ensures every newly opened issue carries `needs-triage`, even blank or
-# API-created ones that bypass the issue templates. Creates the label if it is
-# somehow missing, so the workflow is self-healing.
-
-on:
-  issues:
-    types: [opened]
-
-permissions:
-  issues: write
-
-jobs:
-  needs-triage:
-    runs-on: ubuntu-latest
-    steps:
-      - name: Add needs-triage label
-        uses: actions/github-script@v7
-        with:
-          script: |
-            const { owner, repo } = context.repo;
-            const issue_number = context.payload.issue.number;
-
-            const current = (context.payload.issue.labels || []).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}.`);

.github/workflows/pr-labeler.yml

@@ -1,28 +0,0 @@
-name: PR Labeler
-
-# Applies area:* labels based on which files a PR changes (see .github/labeler.yml).
-# Uses pull_request_target so it also works on fork PRs. SAFE: actions/labeler
-# only reads the changed-file list via the API — it never checks out or runs PR code.
-
-on:
-  pull_request_target:
-    types: [opened, synchronize, reopened, ready_for_review]
-
-permissions:
-  contents: read
-  pull-requests: write
-
-concurrency:
-  group: pr-labeler-${{ github.event.pull_request.number }}
-  cancel-in-progress: true
-
-jobs:
-  label:
-    if: github.event.pull_request.draft == false
-    runs-on: ubuntu-latest
-    steps:
-      - name: Apply area labels
-        uses: actions/labeler@v5
-        with:
-          configuration-path: .github/labeler.yml
-          sync-labels: true

.github/workflows/pr-triage.yml

@@ -1,164 +0,0 @@
-name: PR Triage
-
-# Two responsibilities, both pure-metadata (no PR code is checked out or run):
-#   1. On open/sync: apply size/* + risk:* labels, and needs-validation when the
-#      PR touches the front/back contract surface (backend API, SSE, agents, or
-#      the frontend streaming client). A `skip-validation` label opts out.
-#   2. On maintainer review: apply the `reviewing` label.
-#
-# All labels are managed within their own namespace — labels outside size/*,
-# risk:*, needs-validation and reviewing are never touched here.
-
-on:
-  pull_request_target:
-    types: [opened, synchronize, reopened, ready_for_review]
-  pull_request_review:
-    types: [submitted]
-
-permissions:
-  contents: read
-  pull-requests: write
-
-concurrency:
-  group: pr-triage-${{ github.event.pull_request.number }}
-  cancel-in-progress: false
-
-jobs:
-  size-and-risk:
-    if: github.event_name == 'pull_request_target' && github.event.pull_request.draft == false
-    runs-on: ubuntu-latest
-    steps:
-      - name: Label size, risk and validation need
-        uses: actions/github-script@v7
-        with:
-          script: |
-            const pr = context.payload.pull_request;
-            const { owner, repo } = context.repo;
-            const prNumber = pr.number;
-
-            // ---- size, from additions + deletions ----
-            const churn = (pr.additions || 0) + (pr.deletions || 0);
-            const sizeLabel =
-              churn < 20 ? 'size/XS' :
-              churn < 100 ? 'size/S' :
-              churn < 300 ? 'size/M' :
-              churn < 700 ? 'size/L' : 'size/XL';
-
-            // ---- changed paths ----
-            const files = await github.paginate(github.rest.pulls.listFiles, {
-              owner, repo, pull_number: prNumber, per_page: 100,
-            });
-            const paths = files.map(f => f.filename);
-
-            const matches = (re) => paths.some(p => re.test(p));
-
-            const docsOnly = paths.length > 0 && paths.every(p =>
-              /\.(md|mdx|txt)$/i.test(p) || p.startsWith('docs/') ||
-              /\.(png|jpe?g|gif|svg|webp|ico)$/i.test(p));
-
-            const highRisk = matches(
-              /^backend\/app\/gateway\//) || matches(
-              /^backend\/packages\/harness\/deerflow\/(agents|subagents|sandbox)\//) || matches(
-              /(^|\/)langgraph\.json$/) || matches(
-              /(^|\/)(auth|authz|security)/i) || matches(
-              /(pyproject\.toml|uv\.lock|package\.json|pnpm-lock\.yaml)$/) || matches(
-              /^docker\//) || matches(
-              /^\.github\/workflows\//);
-
-            const riskLabel = docsOnly ? 'risk:low' : (highRisk ? 'risk:high' : 'risk:medium');
-
-            // needs-validation: front/back contract surface
-            const contractSurface =
-              matches(/^backend\/app\/gateway\//) ||
-              matches(/^backend\/packages\/harness\/deerflow\/(agents|subagents)\//) ||
-              matches(/(^|\/)langgraph\.json$/) ||
-              matches(/^frontend\/src\/core\/(api|threads|messages)\//);
-
-            const current = (pr.labels || []).map(l => l.name);
-            const hasSkip = current.includes('skip-validation');
-
-            const desired = [sizeLabel, riskLabel];
-            if (contractSurface && !hasSkip) desired.push('needs-validation');
-
-            const managed = (name) =>
-              name.startsWith('size/') || name.startsWith('risk:') || name === 'needs-validation';
-
-            const toRemove = current.filter(l => managed(l) && !desired.includes(l));
-            const toAdd = desired.filter(l => !current.includes(l));
-
-            for (const name of toRemove) {
-              try {
-                await github.rest.issues.removeLabel({ owner, repo, issue_number: prNumber, name });
-              } catch (e) {
-                if (e.status !== 404) throw e;
-              }
-            }
-            if (toAdd.length) {
-              await github.rest.issues.addLabels({ owner, repo, issue_number: prNumber, labels: toAdd });
-            }
-            core.info(`size=${sizeLabel} risk=${riskLabel} churn=${churn} ` +
-              `validation=${desired.includes('needs-validation')} ` +
-              `(+${toAdd.join(',') || '-'} / -${toRemove.join(',') || '-'})`);
-
-  first-time:
-    if: github.event_name == 'pull_request_target' && github.event.action == 'opened'
-    runs-on: ubuntu-latest
-    steps:
-      - name: Label first-time contributors
-        uses: actions/github-script@v7
-        with:
-          script: |
-            const pr = context.payload.pull_request;
-            const { owner, repo } = context.repo;
-            const assoc = pr.author_association;
-            const isBot = pr.user.type === 'Bot';
-            core.info(`author=${pr.user.login} association=${assoc} bot=${isBot}`);
-
-            // FIRST_TIME_CONTRIBUTOR = no prior merged commit to this repo;
-            // FIRST_TIMER = no prior commit anywhere on GitHub. Either counts.
-            if (isBot || !['FIRST_TIME_CONTRIBUTOR', 'FIRST_TIMER'].includes(assoc)) {
-              core.info('Not a first-time contributor; skipping.');
-              return;
-            }
-            await github.rest.issues.addLabels({
-              owner, repo, issue_number: pr.number, labels: ['first-time-contributor'],
-            });
-            core.info(`Added first-time-contributor to #${pr.number}.`);
-
-  reviewing:
-    if: github.event_name == 'pull_request_review'
-    runs-on: ubuntu-latest
-    steps:
-      - name: Add reviewing label for maintainer reviews
-        uses: actions/github-script@v7
-        with:
-          script: |
-            const { owner, repo } = context.repo;
-            const prNumber = context.payload.pull_request.number;
-            const reviewer = context.payload.review.user.login;
-
-            const { data: perm } = await github.rest.repos.getCollaboratorPermissionLevel({
-              owner, repo, username: reviewer,
-            });
-            if (!['admin', 'write', 'maintain'].includes(perm.permission)) {
-              core.info(`Reviewer ${reviewer} (${perm.permission}) is not a maintainer; skipping.`);
-              return;
-            }
-
-            const { data: labels } = await github.rest.issues.listLabelsOnIssue({
-              owner, repo, issue_number: prNumber,
-            });
-            if (labels.some(l => l.name === 'reviewing')) {
-              core.info('Already labeled reviewing; skipping.');
-              return;
-            }
-            try {
-              await github.rest.issues.addLabels({
-                owner, repo, issue_number: prNumber, labels: ['reviewing'],
-              });
-              core.info(`Added "reviewing" (reviewer ${reviewer}).`);
-            } catch (e) {
-              // 403 is expected for review events on some fork PR contexts.
-              if (e.status === 403) core.info('No permission to label (expected on some fork PRs).');
-              else throw e;
-            }

.github/workflows/triage.yml

@@ -0,0 +1,223 @@
+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}.`);

README.md

@@ -247,6 +247,9 @@ 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.
 
+> [!IMPORTANT]
+> The Gateway holds run state (RunManager and the stream bridge) in process, so production defaults to a single Gateway worker (`GATEWAY_WORKERS=1`). Raising the worker count without a shared cross-worker stream bridge — which is not yet available — breaks run cancellation, SSE reconnects, request de-duplication, and IM channels, because nginx uses no sticky sessions and each worker keeps its own run state. Scale a single worker up with more CPU/RAM (or move the database and sandbox onto dedicated tiers) instead of raising `GATEWAY_WORKERS`.
+
 See [CONTRIBUTING.md](CONTRIBUTING.md) for detailed Docker development guide.
 
 #### Option 2: Local Development
@@ -340,6 +343,8 @@ See the [MCP Server Guide](backend/docs/MCP_SERVER.md) for detailed instructions
 
 DeerFlow supports receiving tasks from messaging apps. Channels auto-start when configured — no public IP required for any of them.
 
+DeerFlow can also expose user-owned IM channel connections in the workspace UI. When `channel_connections` is enabled, logged-in users can bind Telegram, Slack, or Discord from the sidebar / Settings > Channels. It reuses the existing outbound `channels.*` transports, so no public IP or provider callback URL is required. Incoming IM messages then run under the connected DeerFlow user account. See [IM Channel Connections](backend/docs/IM_CHANNEL_CONNECTIONS.md) for setup and security notes.
+
 | Channel | Transport | Difficulty |
 |---------|-----------|------------|
 | Telegram | Bot API (long-polling) | Easy |
@@ -585,6 +590,8 @@ 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.

backend/.gitignore

@@ -24,5 +24,10 @@ 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

@@ -192,7 +192,7 @@ 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
@@ -202,16 +202,17 @@ 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. **SummarizationMiddleware** - Context reduction when approaching token limits (optional, if enabled)
+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
-10. **TodoListMiddleware** - Task tracking with `write_todos` tool (optional, if plan_mode)
+10. **SummarizationMiddleware** - Context reduction when approaching token limits (optional, if enabled)
-11. **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
+11. **TodoListMiddleware** - Task tracking with `write_todos` tool (optional, if plan_mode)
-12. **TitleMiddleware** - Auto-generates thread title after first complete exchange and normalizes structured message content before prompting the title model
+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
-13. **MemoryMiddleware** - Queues conversations for async memory update (filters to user + final AI responses)
+13. **TitleMiddleware** - Auto-generates thread title after first complete exchange and normalizes structured message content before prompting the title model
-14. **ViewImageMiddleware** - Injects base64 image data before LLM call (conditional on vision support)
+14. **MemoryMiddleware** - Queues conversations for async memory update (filters to user + final AI responses)
-15. **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`)
+15. **ViewImageMiddleware** - Injects base64 image data before LLM call (conditional on vision support)
-16. **SubagentLimitMiddleware** - Truncates excess `task` tool calls from model response to enforce `MAX_CONCURRENT_SUBAGENTS` limit (optional, if `subagent_enabled`)
+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`)
-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
+17. **SubagentLimitMiddleware** - Truncates excess `task` tool calls from model response to enforce `MAX_CONCURRENT_SUBAGENTS` limit (optional, if `subagent_enabled`)
-18. **ClarificationMiddleware** - Intercepts `ask_clarification` tool calls, interrupts via `Command(goto=END)` (must be last)
+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
+19. **ClarificationMiddleware** - Intercepts `ask_clarification` tool calls, interrupts via `Command(goto=END)` (must be last)
 
 ### Configuration System
 
@@ -348,6 +349,7 @@ 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`)
@@ -367,8 +369,7 @@ 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, Discord, DingTalk) to the DeerFlow agent via Gateway's LangGraph-compatible API.
-
 
 **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.
 
@@ -378,18 +379,21 @@ Bridges external messaging platforms (Feishu, Slack, Telegram, DingTalk) to the
 - `manager.py` - Core dispatcher: creates threads via `client.threads.create()`, routes commands, keeps Slack/Telegram on `client.runs.wait()`, and uses `client.runs.stream(["messages-tuple", "values"])` for Feishu incremental outbound updates
 - `base.py` - Abstract `Channel` base class (start/stop/send lifecycle)
 - `service.py` - Manages lifecycle of all configured channels from `config.yaml`
-- `slack.py` / `feishu.py` / `telegram.py` / `dingtalk.py` - Platform-specific implementations (`feishu.py` tracks the running card `message_id` in memory and patches the same card in place; `dingtalk.py` optionally uses AI Card streaming for in-place updates when `card_template_id` is configured)
+- `slack.py` / `feishu.py` / `telegram.py` / `discord.py` / `dingtalk.py` - Platform-specific implementations (`feishu.py` tracks the running card `message_id` in memory and patches the same card in place; `dingtalk.py` optionally uses AI Card streaming for in-place updates when `card_template_id` is configured)
+- `app/gateway/routers/channel_connections.py` - Browser-facing user connection and disconnect APIs
+- `deerflow.persistence.channel_connections` - SQL-backed user-owned connection, optional credential, connect state, and conversation store
 
 **Message Flow**:
 1. External platform -> Channel impl -> `MessageBus.publish_inbound()`
 2. `ChannelManager._dispatch_loop()` consumes from queue
-3. For chat: look up/create thread through Gateway's LangGraph-compatible API
+3. For user-owned channel connections, incoming messages carry `connection_id`, `owner_user_id`, and `workspace_id`; `owner_user_id` becomes the DeerFlow run `user_id`, while the raw platform user id remains `channel_user_id`
-4. Feishu chat: `runs.stream()` → accumulate AI text → publish multiple outbound updates (`is_final=False`) → publish final outbound (`is_final=True`)
+4. For chat: look up/create thread through Gateway's LangGraph-compatible API
-5. Slack/Telegram chat: `runs.wait()` → extract final response → publish outbound
+5. Feishu chat: `runs.stream()` → accumulate AI text → publish multiple outbound updates (`is_final=False`) → publish final outbound (`is_final=True`)
-6. Feishu channel sends one running reply card up front, then patches the same card for each outbound update (card JSON sets `config.update_multi=true` for Feishu's patch API requirement)
+6. Slack/Telegram chat: `runs.wait()` → extract final response → publish outbound
-7. DingTalk AI Card mode (when `card_template_id` configured): `runs.stream()` → create card with initial text → stream updates via `PUT /v1.0/card/streaming` → finalize on `is_final=True`. Falls back to `sampleMarkdown` if card creation or streaming fails
+7. Feishu channel sends one running reply card up front, then patches the same card for each outbound update (card JSON sets `config.update_multi=true` for Feishu's patch API requirement)
-8. For commands (`/new`, `/status`, `/models`, `/memory`, `/help`): handle locally or query Gateway API
+8. DingTalk AI Card mode (when `card_template_id` configured): `runs.stream()` → create card with initial text → stream updates via `PUT /v1.0/card/streaming` → finalize on `is_final=True`. Falls back to `sampleMarkdown` if card creation or streaming fails
-9. Outbound → channel callbacks → platform reply
+9. For commands (`/new`, `/status`, `/models`, `/memory`, `/help`): handle locally or query Gateway API
+10. Outbound → channel callbacks → platform reply
 
 **Configuration** (`config.yaml` -> `channels`):
 - `langgraph_url` - LangGraph-compatible Gateway API base URL (default: `http://localhost:8001/api`)
@@ -397,6 +401,16 @@ Bridges external messaging platforms (Feishu, Slack, Telegram, DingTalk) to the
 - In Docker Compose, IM channels run inside the `gateway` container, so `localhost` points back to that container. Use `http://gateway:8001/api` for `langgraph_url` and `http://gateway:8001` for `gateway_url`, or set `DEER_FLOW_CHANNELS_LANGGRAPH_URL` / `DEER_FLOW_CHANNELS_GATEWAY_URL`.
 - Per-channel configs: `feishu` (app_id, app_secret), `slack` (bot_token, app_token), `telegram` (bot_token), `dingtalk` (client_id, client_secret, optional `card_template_id` for AI Card streaming)
 
+**User-owned channel connections** (`config.yaml` -> `channel_connections`):
+- Disabled by default. It is a user-binding layer on top of the existing `channels.*` runtime config, not a replacement for provider bot credentials.
+- No public IP, OAuth callback URL, or provider webhook route is required by the current implementation.
+- Telegram uses a deep-link `/start <code>` flow over the existing long-polling worker. Slack uses `/connect <code>` over the existing Socket Mode worker. Discord uses `/connect <code>` over the existing Gateway worker.
+- Frontend APIs: `GET /api/channels/providers`, `GET /api/channels/connections`, `POST /api/channels/{provider}/connect`, and `DELETE /api/channels/connections/{connection_id}`.
+- Browser APIs remain protected by normal Gateway auth/CSRF. Provider messages arrive through the already-configured channel workers.
+- Slack replies use the configured operator bot token from `channels.slack` unless a future provider-token flow stores per-connection credentials.
+- Telegram, Slack, and Discord workers resolve incoming platform identities to connection records before reaching `ChannelManager`.
+- See `backend/docs/IM_CHANNEL_CONNECTIONS.md` for provider setup and operational notes.
+
 
 ### Memory System (`packages/harness/deerflow/agents/memory/`)
 
@@ -493,7 +507,7 @@ 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

backend/app/channels/commands.py

@@ -18,3 +18,10 @@ 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 KNOWN_CHANNEL_COMMANDS
+from app.channels.commands import is_known_channel_command
 from app.channels.message_bus import InboundMessage, InboundMessageType, MessageBus, OutboundMessage, ResolvedAttachment
 
 logger = logging.getLogger(__name__)
@@ -59,9 +59,7 @@ def _normalize_allowed_users(allowed_users: Any) -> set[str]:
 
 
 def _is_dingtalk_command(text: str) -> bool:
-    if not text.startswith("/"):
+    return is_known_channel_command(text)
-        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

@@ -10,13 +10,24 @@ from pathlib import Path
 from typing import Any
 
 from app.channels.base import Channel
-from app.channels.message_bus import InboundMessageType, MessageBus, OutboundMessage, ResolvedAttachment
+from app.channels.commands import is_known_channel_command
+from app.channels.message_bus import InboundMessage, InboundMessageType, MessageBus, OutboundMessage, ResolvedAttachment
 
 logger = logging.getLogger(__name__)
 
 _DISCORD_MAX_MESSAGE_LEN = 2000
 
 
+def _extract_connect_code(text: str) -> str | None:
+    parts = text.strip().split()
+    if len(parts) < 2:
+        return None
+    command = parts[0].lower()
+    if command in {"/connect", "connect"}:
+        return parts[1]
+    return None
+
+
 class DiscordChannel(Channel):
     """Discord bot channel.
 
@@ -69,6 +80,7 @@ class DiscordChannel(Channel):
         self._discord_loop: asyncio.AbstractEventLoop | None = None
         self._main_loop: asyncio.AbstractEventLoop | None = None
         self._discord_module = None
+        self._connection_repo = config.get("connection_repo")
 
     async def start(self) -> None:
         if self._running:
@@ -286,6 +298,10 @@ class DiscordChannel(Channel):
             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)
 
+        connect_code = _extract_connect_code(text)
+        if connect_code and await self._bind_connection_from_connect_code(message, connect_code):
+            return
+
         # --- Determine thread/channel routing and typing target ---
         thread_id = None
         chat_id = None
@@ -300,7 +316,7 @@ class DiscordChannel(Channel):
 
             # If this is a known active thread, process normally
             if thread_id in self._active_thread_ids:
-                msg_type = InboundMessageType.COMMAND if text.startswith("/") else InboundMessageType.CHAT
+                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),
@@ -314,6 +330,7 @@ class DiscordChannel(Channel):
                     },
                 )
                 inbound.topic_id = thread_id
+                inbound = await self._attach_connection_identity(inbound, guild_id=str(guild.id) if guild else None)
                 self._publish(inbound)
                 # Start typing indicator in the thread
                 if typing_target:
@@ -407,7 +424,7 @@ class DiscordChannel(Channel):
             chat_id = channel_id
             typing_target = message.channel  # Type into the channel
 
-        msg_type = InboundMessageType.COMMAND if text.startswith("/") else InboundMessageType.CHAT
+        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),
@@ -421,6 +438,7 @@ class DiscordChannel(Channel):
             },
         )
         inbound.topic_id = thread_id
+        inbound = await self._attach_connection_identity(inbound, guild_id=str(guild.id) if guild else None)
 
         # Start typing indicator in the correct target (thread or channel)
         if typing_target:
@@ -435,6 +453,76 @@ class DiscordChannel(Channel):
             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)
 
+    async def _attach_connection_identity(self, inbound: InboundMessage, guild_id: str | None = None) -> InboundMessage:
+        if self._connection_repo is None:
+            return inbound
+
+        connection = None
+        if guild_id:
+            connection = await self._connection_repo.find_connection_by_external_identity(
+                provider="discord",
+                external_account_id=inbound.user_id,
+                workspace_id=guild_id,
+            )
+        if connection is None:
+            connection = await self._connection_repo.find_connection_by_external_identity(
+                provider="discord",
+                external_account_id=inbound.user_id,
+                workspace_id=None,
+            )
+        if connection is None:
+            return inbound
+
+        inbound.connection_id = connection["id"]
+        inbound.owner_user_id = connection["owner_user_id"]
+        inbound.workspace_id = connection.get("workspace_id")
+        return inbound
+
+    async def _bind_connection_from_connect_code(self, message, code: str) -> bool:
+        if self._connection_repo is None or not code:
+            return False
+
+        state = await self._connection_repo.consume_oauth_state(provider="discord", state=code)
+        if state is None:
+            await self._send_connection_reply(message, "Discord connection code is invalid or expired.")
+            return True
+
+        guild = getattr(message, "guild", None)
+        channel = getattr(message, "channel", None)
+        author = getattr(message, "author", None)
+        user_id = str(getattr(author, "id", "") or "")
+        if not user_id:
+            await self._send_connection_reply(message, "Discord connection could not be completed from this message.")
+            return True
+
+        guild_id = str(getattr(guild, "id", "") or "") or None
+        await self._connection_repo.upsert_connection(
+            owner_user_id=state["owner_user_id"],
+            provider="discord",
+            external_account_id=user_id,
+            external_account_name=getattr(author, "display_name", None) or getattr(author, "name", None),
+            workspace_id=guild_id,
+            workspace_name=getattr(guild, "name", None) if guild is not None else None,
+            metadata={
+                "guild_id": guild_id,
+                "channel_id": str(getattr(channel, "id", "") or ""),
+            },
+            status="connected",
+        )
+        await self._send_connection_reply(message, "Discord connected to DeerFlow.")
+        return True
+
+    @staticmethod
+    async def _send_connection_reply(message, text: str) -> None:
+        channel = getattr(message, "channel", None)
+        send = getattr(channel, "send", None)
+        if send is None:
+            return
+        try:
+            await send(text)
+        except Exception:
+            logger.exception("[Discord] failed to send connection reply")
+
     def _run_client(self) -> None:
         self._discord_loop = asyncio.new_event_loop()
         asyncio.set_event_loop(self._discord_loop)

backend/app/channels/feishu.py

@@ -11,7 +11,7 @@ import time
 from typing import Any, Literal
 
 from app.channels.base import Channel
-from app.channels.commands import KNOWN_CHANNEL_COMMANDS
+from app.channels.commands import is_known_channel_command
 from app.channels.message_bus import (
     PENDING_CLARIFICATION_METADATA_KEY,
     RESOLVED_FROM_PENDING_CLARIFICATION_METADATA_KEY,
@@ -30,9 +30,7 @@ PENDING_CLARIFICATION_TTL_SECONDS = 30 * 60
 
 
 def _is_feishu_command(text: str) -> bool:
-    if not text.startswith("/"):
+    return is_known_channel_command(text)
-        return False
-    return text.split(maxsplit=1)[0].lower() in KNOWN_CHANNEL_COMMANDS
 
 
 class FeishuChannel(Channel):

backend/app/channels/manager.py

@@ -8,6 +8,7 @@ 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
 
@@ -26,8 +27,13 @@ from app.channels.message_bus import (
 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__)
 
@@ -124,6 +130,16 @@ 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
@@ -410,6 +426,46 @@ 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.
 
@@ -614,6 +670,7 @@ class ChannelManager:
         assistant_id: str = DEFAULT_ASSISTANT_ID,
         default_session: dict[str, Any] | None = None,
         channel_sessions: dict[str, Any] | None = None,
+        connection_repo: Any | None = None,
     ) -> None:
         self.bus = bus
         self.store = store
@@ -623,7 +680,9 @@ class ChannelManager:
         self._assistant_id = assistant_id
         self._default_session = _as_dict(default_session)
         self._channel_sessions = dict(channel_sessions or {})
+        self._connection_repo = connection_repo
         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
@@ -671,12 +730,16 @@ class ChannelManager:
         configurable["checkpoint_ns"] = ""
         configurable["thread_id"] = thread_id
 
-        # ``user_id`` drives user-scoped filesystem buckets that only accept
+        # ``user_id`` drives DeerFlow-owned memory, files, and thread buckets.
-        # ``[A-Za-z0-9_-]``, so normalize the channel id and keep the raw value
+        # For browser-connected IM channels, prefer the DeerFlow account that
-        # under ``channel_user_id`` for platform-facing lookups.
+        # owns the connection. Preserve the raw platform user under
+        # ``channel_user_id`` for platform-facing lookups and audits.
         run_context_identity: dict[str, Any] = {"thread_id": thread_id}
-        if msg.user_id:
+        if msg.owner_user_id:
+            run_context_identity["user_id"] = make_safe_user_id(msg.owner_user_id)
+        elif msg.user_id:
             run_context_identity["user_id"] = make_safe_user_id(msg.user_id)
+        if msg.user_id:
             run_context_identity["channel_user_id"] = msg.user_id
 
         run_context = _merge_dicts(
@@ -696,6 +759,21 @@ 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):
@@ -713,6 +791,11 @@ 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:
@@ -782,6 +865,14 @@ 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)",
@@ -792,10 +883,27 @@ class ChannelManager:
 
     # -- chat handling -----------------------------------------------------
 
-    async def _create_thread(self, client, msg: InboundMessage) -> str:
+    async def _lookup_thread_id(self, msg: InboundMessage) -> str | None:
-        """Create a new thread through Gateway and store the mapping."""
+        if msg.connection_id and self._connection_repo is not None:
-        thread = await client.threads.create()
+            return await self._connection_repo.get_thread_id(
-        thread_id = thread["thread_id"]
+                msg.connection_id,
+                msg.chat_id,
+                msg.topic_id,
+            )
+        return self.store.get_thread_id(msg.channel_name, msg.chat_id, topic_id=msg.topic_id)
+
+    async def _store_thread_id(self, msg: InboundMessage, thread_id: str) -> None:
+        if msg.connection_id and msg.owner_user_id and self._connection_repo is not None:
+            await self._connection_repo.set_thread_id(
+                connection_id=msg.connection_id,
+                owner_user_id=msg.owner_user_id,
+                provider=msg.channel_name,
+                external_conversation_id=msg.chat_id,
+                external_topic_id=msg.topic_id,
+                thread_id=thread_id,
+            )
+            return
+
         self.store.set_thread_id(
             msg.channel_name,
             msg.chat_id,
@@ -803,6 +911,12 @@ class ChannelManager:
             topic_id=msg.topic_id,
             user_id=msg.user_id,
         )
+
+    async def _create_thread(self, client, msg: InboundMessage) -> str:
+        """Create a new thread through Gateway and store the mapping."""
+        thread = await client.threads.create()
+        thread_id = thread["thread_id"]
+        await self._store_thread_id(msg, thread_id)
         logger.info("[Manager] new thread created through Gateway: thread_id=%s for chat_id=%s topic_id=%s", thread_id, msg.chat_id, msg.topic_id)
         return thread_id
 
@@ -812,7 +926,7 @@ class ChannelManager:
         # Look up existing DeerFlow thread.
         # topic_id may be None (e.g. Telegram private chats) — the store
         # handles this by using the "channel:chat_id" key without a topic suffix.
-        thread_id = self.store.get_thread_id(msg.channel_name, msg.chat_id, topic_id=msg.topic_id)
+        thread_id = await self._lookup_thread_id(msg)
         if thread_id:
             logger.info("[Manager] reusing thread: thread_id=%s for topic_id=%s", thread_id, msg.topic_id)
 
@@ -836,9 +950,11 @@ 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(
@@ -848,6 +964,7 @@ class ChannelManager:
                 assistant_id,
                 run_config,
                 run_context,
+                human_message,
             )
             return
 
@@ -856,7 +973,7 @@ class ChannelManager:
             result = await client.runs.wait(
                 thread_id,
                 assistant_id,
-                input={"messages": [{"role": "human", "content": msg.text}]},
+                input={"messages": [human_message]},
                 config=run_config,
                 context=run_context,
                 multitask_strategy="reject",
@@ -896,6 +1013,8 @@ class ChannelManager:
             artifacts=artifacts,
             attachments=attachments,
             thread_ts=msg.thread_ts,
+            connection_id=msg.connection_id,
+            owner_user_id=msg.owner_user_id,
             metadata=_response_metadata(msg.metadata, pending_clarification=pending_clarification),
         )
         logger.info("[Manager] publishing outbound message to bus: channel=%s, chat_id=%s", msg.channel_name, msg.chat_id)
@@ -909,6 +1028,7 @@ 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])
 
@@ -924,7 +1044,7 @@ class ChannelManager:
             async for chunk in client.runs.stream(
                 thread_id,
                 assistant_id,
-                input={"messages": [{"role": "human", "content": msg.text}]},
+                input={"messages": [human_message]},
                 config=run_config,
                 context=run_context,
                 stream_mode=["messages-tuple", "values"],
@@ -958,6 +1078,8 @@ class ChannelManager:
                         text=latest_text,
                         is_final=False,
                         thread_ts=msg.thread_ts,
+                        connection_id=msg.connection_id,
+                        owner_user_id=msg.owner_user_id,
                         metadata=_response_metadata(msg.metadata),
                     )
                 )
@@ -1004,6 +1126,8 @@ class ChannelManager:
                     attachments=attachments,
                     is_final=True,
                     thread_ts=msg.thread_ts,
+                    connection_id=msg.connection_id,
+                    owner_user_id=msg.owner_user_id,
                     metadata=_response_metadata(msg.metadata, pending_clarification=pending_clarification),
                 )
             )
@@ -1011,11 +1135,20 @@ class ChannelManager:
     # -- command handling --------------------------------------------------
 
     async def _handle_command(self, msg: InboundMessage) -> None:
-        text = msg.text.strip()
+        raw_text = msg.text
+        text = raw_text.strip()
         parts = text.split(maxsplit=1)
-        command = parts[0].lower().lstrip("/")
+        reply: str | None = None
+        if not parts:
+            command = None
+            reply = _unknown_command_reply()
+        else:
+            command = parts[0].lower().removeprefix("/")
 
-        if command == "bootstrap":
+        if reply is None and not raw_text.startswith("/"):
+            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"
@@ -1023,27 +1156,21 @@ class ChannelManager:
             await self._handle_chat(chat_msg, extra_context={"is_bootstrap": True})
             return
 
-        if command == "new":
+        if reply is None and command == "new":
             # Create a new thread through Gateway
             client = self._get_client()
             thread = await client.threads.create()
             new_thread_id = thread["thread_id"]
-            self.store.set_thread_id(
+            await self._store_thread_id(msg, new_thread_id)
-                msg.channel_name,
-                msg.chat_id,
-                new_thread_id,
-                topic_id=msg.topic_id,
-                user_id=msg.user_id,
-            )
             reply = "New conversation started."
-        elif command == "status":
+        elif reply is None and command == "status":
-            thread_id = self.store.get_thread_id(msg.channel_name, msg.chat_id, topic_id=msg.topic_id)
+            thread_id = await self._lookup_thread_id(msg)
             reply = f"Active thread: {thread_id}" if thread_id else "No active conversation."
-        elif command == "models":
+        elif reply is None and command == "models":
             reply = await self._fetch_gateway("/api/models", "models")
-        elif command == "memory":
+        elif reply is None and command == "memory":
             reply = await self._fetch_gateway("/api/memory", "memory")
-        elif command == "help":
+        elif reply is None and command == "help":
             reply = (
                 "Available commands:\n"
                 "/bootstrap — Start a bootstrap session (enables agent setup)\n"
@@ -1051,18 +1178,36 @@ 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"
             )
-        else:
+        elif reply is None:
-            available = " | ".join(sorted(KNOWN_CHANNEL_COMMANDS))
+            slash_resolution = await asyncio.to_thread(
-            reply = f"Unknown command: /{command}. Available commands: {available}"
+                lambda: _resolve_slash_skill_command(
+                    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) or "",
+            thread_id=await self._lookup_thread_id(msg) or "",
             text=reply,
             thread_ts=msg.thread_ts,
+            connection_id=msg.connection_id,
+            owner_user_id=msg.owner_user_id,
             metadata=_slim_metadata(msg.metadata),
         )
         await self.bus.publish_outbound(outbound)
@@ -1098,9 +1243,11 @@ 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) or "",
+            thread_id=await self._lookup_thread_id(msg) or "",
             text=error_text,
             thread_ts=msg.thread_ts,
+            connection_id=msg.connection_id,
+            owner_user_id=msg.owner_user_id,
             metadata=_slim_metadata(msg.metadata),
         )
         await self.bus.publish_outbound(outbound)

backend/app/channels/message_bus.py

@@ -44,6 +44,12 @@ class InboundMessage:
             Messages sharing the same ``topic_id`` within a ``chat_id`` will
             reuse the same DeerFlow thread.  When ``None``, each message
             creates a new thread (one-shot Q&A).
+        connection_id: Optional DeerFlow channel connection id. When present,
+            conversation mapping is scoped by the connection instead of the
+            legacy global ``channel_name:chat_id[:topic_id]`` key.
+        owner_user_id: DeerFlow user id that owns the channel connection.
+            Platform user ids stay in ``user_id``.
+        workspace_id: Optional external workspace/guild/team id.
         files: Optional list of file attachments (platform-specific dicts).
         metadata: Arbitrary extra data from the channel.
         created_at: Unix timestamp when the message was created.
@@ -56,6 +62,9 @@ class InboundMessage:
     msg_type: InboundMessageType = InboundMessageType.CHAT
     thread_ts: str | None = None
     topic_id: str | None = None
+    connection_id: str | None = None
+    owner_user_id: str | None = None
+    workspace_id: str | None = None
     files: list[dict[str, Any]] = field(default_factory=list)
     metadata: dict[str, Any] = field(default_factory=dict)
     created_at: float = field(default_factory=time.time)
@@ -95,6 +104,9 @@ class OutboundMessage:
         is_final: Whether this is the final message in the response stream.
         thread_ts: Optional platform thread identifier for threaded replies.
         metadata: Arbitrary extra data.
+        connection_id: Optional DeerFlow channel connection id used for
+            connection-specific outbound credentials.
+        owner_user_id: DeerFlow user id that owns the channel connection.
         created_at: Unix timestamp.
     """
 
@@ -106,6 +118,8 @@ class OutboundMessage:
     attachments: list[ResolvedAttachment] = field(default_factory=list)
     is_final: bool = True
     thread_ts: str | None = None
+    connection_id: str | None = None
+    owner_user_id: str | None = None
     metadata: dict[str, Any] = field(default_factory=dict)
     created_at: float = field(default_factory=time.time)
 

backend/app/channels/service.py

@@ -52,6 +52,31 @@ def _resolve_service_url(config: dict[str, Any], config_key: str, env_key: str,
     return default
 
 
+def _merge_channel_connection_runtime_config(channels_config: dict[str, Any], app_config: AppConfig) -> None:
+    connection_config = getattr(app_config, "channel_connections", None)
+    if connection_config is None or not getattr(connection_config, "enabled", False):
+        return
+
+
+def _make_connection_repo(app_config: AppConfig):
+    connection_config = getattr(app_config, "channel_connections", None)
+    if connection_config is None or not getattr(connection_config, "enabled", False):
+        return None
+
+    try:
+        from deerflow.persistence.channel_connections import ChannelConnectionRepository
+        from deerflow.persistence.engine import get_session_factory
+    except Exception:
+        logger.exception("Failed to import channel connection repository")
+        return None
+
+    session_factory = get_session_factory()
+    if session_factory is None:
+        logger.warning("Channel connections are enabled but database persistence is not available")
+        return None
+    return ChannelConnectionRepository(session_factory)
+
+
 class ChannelService:
     """Manages the lifecycle of all configured IM channels.
 
@@ -59,9 +84,10 @@ class ChannelService:
     instantiates enabled channels, and starts the ChannelManager dispatcher.
     """
 
-    def __init__(self, channels_config: dict[str, Any] | None = None) -> None:
+    def __init__(self, channels_config: dict[str, Any] | None = None, *, connection_repo: Any | None = None) -> None:
         self.bus = MessageBus()
         self.store = ChannelStore()
+        self._connection_repo = connection_repo
         config = dict(channels_config or {})
         langgraph_url = _resolve_service_url(config, "langgraph_url", _CHANNELS_LANGGRAPH_URL_ENV, DEFAULT_LANGGRAPH_URL)
         gateway_url = _resolve_service_url(config, "gateway_url", _CHANNELS_GATEWAY_URL_ENV, DEFAULT_GATEWAY_URL)
@@ -74,6 +100,7 @@ class ChannelService:
             gateway_url=gateway_url,
             default_session=default_session if isinstance(default_session, dict) else None,
             channel_sessions=channel_sessions,
+            connection_repo=connection_repo,
         )
         self._channels: dict[str, Any] = {}  # name -> Channel instance
         self._config = config
@@ -90,8 +117,9 @@ class ChannelService:
         # extra fields are allowed by AppConfig (extra="allow")
         extra = app_config.model_extra or {}
         if "channels" in extra:
-            channels_config = extra["channels"]
+            channels_config = dict(extra["channels"] or {})
-        return cls(channels_config=channels_config)
+        _merge_channel_connection_runtime_config(channels_config, app_config)
+        return cls(channels_config=channels_config, connection_repo=_make_connection_repo(app_config))
 
     async def start(self) -> None:
         """Start the manager and all enabled channels."""
@@ -169,6 +197,8 @@ class ChannelService:
         try:
             config = dict(config)
             config["channel_store"] = self.store
+            if self._connection_repo is not None:
+                config["connection_repo"] = self._connection_repo
             channel = channel_cls(bus=self.bus, config=config)
             self._channels[name] = channel
             await channel.start()

backend/app/channels/slack.py

@@ -9,6 +9,7 @@ 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__)
@@ -32,6 +33,30 @@ 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()
+
+
+def _extract_connect_code(text: str) -> str | None:
+    parts = text.strip().split()
+    if len(parts) < 2:
+        return None
+    command = parts[0].lower()
+    if command in {"/connect", "connect"}:
+        return parts[1]
+    return None
+
+
 class SlackChannel(Channel):
     """Slack IM channel using Socket Mode (WebSocket, no public IP).
 
@@ -49,6 +74,10 @@ class SlackChannel(Channel):
         self._web_client = None
         self._loop: asyncio.AbstractEventLoop | None = None
         self._allowed_users = _normalize_allowed_users(config.get("allowed_users", []))
+        self._connection_repo = config.get("connection_repo")
+        self._web_client_factory = config.get("web_client_factory")
+        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:
@@ -63,15 +92,28 @@ class SlackChannel(Channel):
             return
 
         self._SocketModeResponse = SocketModeResponse
+        if self._web_client_factory is None:
+            self._web_client_factory = WebClient
 
         bot_token = self.config.get("bot_token", "")
         app_token = self.config.get("app_token", "")
 
+        if self._connection_repo is not None and self.config.get("event_delivery") == "http":
+            if not bot_token:
+                logger.error("Slack HTTP Events mode requires bot_token")
+                return
+            await self._initialize_operator_web_client(str(bot_token))
+            self._loop = asyncio.get_event_loop()
+            self._running = True
+            self.bus.subscribe_outbound(self._on_outbound)
+            logger.info("Slack channel started in HTTP Events mode")
+            return
+
         if not bot_token or not app_token:
             logger.error("Slack channel requires bot_token and app_token")
             return
 
-        self._web_client = WebClient(token=bot_token)
+        await self._initialize_operator_web_client(str(bot_token))
         self._socket_client = SocketModeClient(
             app_token=app_token,
             web_client=self._web_client,
@@ -96,7 +138,8 @@ class SlackChannel(Channel):
         logger.info("Slack channel stopped")
 
     async def send(self, msg: OutboundMessage, *, _max_retries: int = 3) -> None:
-        if not self._web_client:
+        web_client = await self._get_web_client_for_message(msg)
+        if not web_client:
             return
 
         kwargs: dict[str, Any] = {
@@ -109,11 +152,12 @@ class SlackChannel(Channel):
         last_exc: Exception | None = None
         for attempt in range(_max_retries):
             try:
-                await asyncio.to_thread(self._web_client.chat_postMessage, **kwargs)
+                await asyncio.to_thread(web_client.chat_postMessage, **kwargs)
                 # Add a completion reaction to the thread root
                 if msg.thread_ts:
                     await asyncio.to_thread(
-                        self._add_reaction,
+                        self._add_reaction_with_client,
+                        web_client,
                         msg.chat_id,
                         msg.thread_ts,
                         "white_check_mark",
@@ -137,7 +181,8 @@ class SlackChannel(Channel):
         if msg.thread_ts:
             try:
                 await asyncio.to_thread(
-                    self._add_reaction,
+                    self._add_reaction_with_client,
+                    web_client,
                     msg.chat_id,
                     msg.thread_ts,
                     "x",
@@ -149,7 +194,8 @@ class SlackChannel(Channel):
         raise last_exc
 
     async def send_file(self, msg: OutboundMessage, attachment: ResolvedAttachment) -> bool:
-        if not self._web_client:
+        web_client = await self._get_web_client_for_message(msg)
+        if not web_client:
             return False
 
         try:
@@ -162,7 +208,7 @@ class SlackChannel(Channel):
             if msg.thread_ts:
                 kwargs["thread_ts"] = msg.thread_ts
 
-            await asyncio.to_thread(self._web_client.files_upload_v2, **kwargs)
+            await asyncio.to_thread(web_client.files_upload_v2, **kwargs)
             logger.info("[Slack] file uploaded: %s to channel=%s", attachment.filename, msg.chat_id)
             return True
         except Exception:
@@ -171,12 +217,38 @@ class SlackChannel(Channel):
 
     # -- internal ----------------------------------------------------------
 
-    def _add_reaction(self, channel_id: str, timestamp: str, emoji: str) -> None:
+    async def _initialize_operator_web_client(self, bot_token: str) -> None:
-        """Add an emoji reaction to a message (best-effort, non-blocking)."""
+        self._web_client = self._web_client_factory(token=bot_token)
-        if not self._web_client:
+        if self._bot_user_id is not None:
             return
         try:
-            self._web_client.reactions_add(
+            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)
+
+    async def _get_web_client_for_message(self, msg: OutboundMessage):
+        if msg.connection_id and self._connection_repo is not None:
+            credentials = await self._connection_repo.get_credentials(msg.connection_id)
+            access_token = credentials.get("access_token") if credentials else None
+            if not access_token:
+                return self._web_client
+            if self._web_client_factory is None:
+                from slack_sdk import WebClient
+
+                self._web_client_factory = WebClient
+            return self._web_client_factory(token=access_token)
+        return self._web_client
+
+    @staticmethod
+    def _add_reaction_with_client(web_client, channel_id: str, timestamp: str, emoji: str) -> None:
+        try:
+            web_client.reactions_add(
                 channel=channel_id,
                 timestamp=timestamp,
                 name=emoji,
@@ -185,6 +257,12 @@ class SlackChannel(Channel):
             if "already_reacted" not in str(exc):
                 logger.warning("[Slack] failed to add reaction %s: %s", emoji, exc)
 
+    def _add_reaction(self, channel_id: str, timestamp: str, emoji: str) -> None:
+        """Add an emoji reaction to a message (best-effort, non-blocking)."""
+        if not self._web_client:
+            return
+        self._add_reaction_with_client(self._web_client, channel_id, timestamp, emoji)
+
     def _send_running_reply(self, channel_id: str, thread_ts: str) -> None:
         """Send a 'Working on it......' reply in the thread (called from SDK thread)."""
         if not self._web_client:
@@ -210,17 +288,26 @@ 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", "")
 
             # Handle message events (DM or @mention)
             if etype in ("message", "app_mention"):
-                self._handle_message_event(event)
+                self._handle_message_event(
+                    event,
+                    team_id=req.payload.get("team_id") or req.payload.get("team") or event.get("team"),
+                )
 
         except Exception:
             logger.exception("Error processing Slack event")
 
-    def _handle_message_event(self, event: dict) -> None:
+    def _handle_message_event(self, event: dict, *, team_id: str | None = None) -> None:
         # Ignore bot messages
         if event.get("bot_id") or event.get("subtype"):
             return
@@ -233,13 +320,28 @@ 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
 
+        connect_code = _extract_connect_code(text)
+        if connect_code:
+            if self._loop and self._loop.is_running():
+                asyncio.run_coroutine_threadsafe(
+                    self._bind_connection_from_connect_code(
+                        event=event,
+                        team_id=str(team_id or event.get("team") or ""),
+                        code=connect_code,
+                    ),
+                    self._loop,
+                )
+            return
+
         channel_id = event.get("channel", "")
         thread_ts = event.get("thread_ts") or event.get("ts", "")
 
-        if text.startswith("/"):
+        if is_known_channel_command(text):
             msg_type = InboundMessageType.COMMAND
         else:
             msg_type = InboundMessageType.CHAT
@@ -261,4 +363,73 @@ class SlackChannel(Channel):
             self._add_reaction(channel_id, event.get("ts", thread_ts), "eyes")
             # Send "running" reply first (fire-and-forget from SDK thread)
             self._send_running_reply(channel_id, thread_ts)
-            asyncio.run_coroutine_threadsafe(self.bus.publish_inbound(inbound), self._loop)
+            if self._connection_repo is None:
+                asyncio.run_coroutine_threadsafe(self.bus.publish_inbound(inbound), self._loop)
+            else:
+                asyncio.run_coroutine_threadsafe(self._publish_inbound_with_connection(inbound, team_id=team_id), self._loop)
+
+    async def _publish_inbound_with_connection(self, inbound, *, team_id: str | None = None) -> None:
+        inbound = await self._attach_connection_identity(inbound, team_id=team_id)
+        await self.bus.publish_inbound(inbound)
+
+    async def _attach_connection_identity(self, inbound, *, team_id: str | None = None):
+        if self._connection_repo is None:
+            return inbound
+
+        workspace_id = str(team_id or inbound.metadata.get("team_id") or "")
+        if not workspace_id:
+            return inbound
+
+        connection = await self._connection_repo.find_connection_by_external_identity(
+            provider="slack",
+            external_account_id=inbound.user_id,
+            workspace_id=workspace_id,
+        )
+        if connection is None:
+            return inbound
+
+        inbound.connection_id = connection["id"]
+        inbound.owner_user_id = connection["owner_user_id"]
+        inbound.workspace_id = connection.get("workspace_id")
+        return inbound
+
+    async def _bind_connection_from_connect_code(self, *, event: dict, team_id: str, code: str) -> bool:
+        if self._connection_repo is None or not code:
+            return False
+
+        channel_id = str(event.get("channel") or "")
+        thread_ts = str(event.get("thread_ts") or event.get("ts") or "")
+        state = await self._connection_repo.consume_oauth_state(provider="slack", state=code)
+        if state is None:
+            self._post_connection_reply(channel_id, "Slack connection code is invalid or expired.", thread_ts)
+            return True
+
+        user_id = str(event.get("user") or "")
+        if not user_id or not team_id:
+            self._post_connection_reply(channel_id, "Slack connection could not be completed from this message.", thread_ts)
+            return True
+
+        await self._connection_repo.upsert_connection(
+            owner_user_id=state["owner_user_id"],
+            provider="slack",
+            external_account_id=user_id,
+            workspace_id=team_id,
+            metadata={
+                "team_id": team_id,
+                "channel_id": channel_id,
+            },
+            status="connected",
+        )
+        self._post_connection_reply(channel_id, "Slack connected to DeerFlow.", thread_ts)
+        return True
+
+    def _post_connection_reply(self, channel_id: str, text: str, thread_ts: str | None = None) -> None:
+        if not self._web_client or not channel_id:
+            return
+        kwargs: dict[str, Any] = {"channel": channel_id, "text": text}
+        if thread_ts:
+            kwargs["thread_ts"] = thread_ts
+        try:
+            self._web_client.chat_postMessage(**kwargs)
+        except Exception:
+            logger.exception("[Slack] failed to send connection reply in channel=%s", channel_id)

backend/app/channels/telegram.py

@@ -35,6 +35,7 @@ class TelegramChannel(Channel):
                 pass
         # chat_id -> last sent message_id for threaded replies
         self._last_bot_message: dict[str, int] = {}
+        self._connection_repo = config.get("connection_repo")
 
     async def start(self) -> None:
         if self._running:
@@ -60,12 +61,17 @@ 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))
 
@@ -171,6 +177,26 @@ class TelegramChannel(Channel):
             logger.exception("[Telegram] failed to send file: %s", attachment.filename)
             return False
 
+    async def process_webhook_update(self, payload: dict[str, Any]) -> bool:
+        if not self._application:
+            return False
+        try:
+            from telegram import Update
+        except ImportError:
+            logger.error("python-telegram-bot is not installed. Install it with: uv add python-telegram-bot")
+            return False
+
+        update = Update.de_json(payload, self._application.bot)
+        if update is None:
+            return False
+
+        if self._tg_loop and self._tg_loop.is_running():
+            future = asyncio.run_coroutine_threadsafe(self._application.process_update(update), self._tg_loop)
+            await asyncio.wrap_future(future)
+        else:
+            await self._application.process_update(update)
+        return True
+
     # -- helpers -----------------------------------------------------------
 
     async def _send_running_reply(self, chat_id: str, reply_to_message_id: int) -> None:
@@ -228,10 +254,99 @@ class TelegramChannel(Channel):
             return True
         return user_id in self._allowed_users
 
+    @staticmethod
+    def _telegram_display_name(user) -> str:
+        full_name = getattr(user, "full_name", None)
+        if isinstance(full_name, str) and full_name:
+            return full_name
+        username = getattr(user, "username", None)
+        if isinstance(username, str) and username:
+            return username
+        return str(getattr(user, "id", ""))
+
+    async def _bind_connection_from_start_token(self, update, state_token: str) -> bool:
+        if self._connection_repo is None or not state_token:
+            return False
+
+        state = await self._connection_repo.consume_oauth_state(provider="telegram", state=state_token)
+        if state is None:
+            await update.message.reply_text("Telegram connection link is invalid or expired.")
+            return True
+
+        owner_user_id = state["owner_user_id"]
+        user_id = str(update.effective_user.id)
+        chat_id = str(update.effective_chat.id)
+        connection = await self._connection_repo.upsert_connection(
+            owner_user_id=owner_user_id,
+            provider="telegram",
+            external_account_id=user_id,
+            external_account_name=self._telegram_display_name(update.effective_user),
+            workspace_id=chat_id,
+            workspace_name=None,
+            metadata={
+                "chat_id": chat_id,
+                "chat_type": update.effective_chat.type,
+                "telegram_username": getattr(update.effective_user, "username", None),
+            },
+            status="connected",
+        )
+        logger.info("[Telegram] bound chat=%s user=%s to DeerFlow user=%s connection=%s", chat_id, user_id, owner_user_id, connection["id"])
+        await update.message.reply_text("Telegram connected to DeerFlow.")
+        return True
+
+    async def _attach_connection_identity(self, inbound: InboundMessage) -> InboundMessage:
+        if self._connection_repo is None:
+            return inbound
+
+        connection = await self._connection_repo.find_connection_by_external_identity(
+            provider="telegram",
+            external_account_id=inbound.user_id,
+            workspace_id=inbound.chat_id,
+        )
+        if connection is None:
+            return inbound
+
+        inbound.connection_id = connection["id"]
+        inbound.owner_user_id = connection["owner_user_id"]
+        inbound.workspace_id = connection.get("workspace_id")
+        return inbound
+
+    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):
             return
+        args = getattr(context, "args", []) if context is not None else []
+        if args:
+            handled = await self._bind_connection_from_start_token(update, str(args[0]))
+            if handled:
+                return
         await update.message.reply_text("Welcome to DeerFlow! Send me a message to start a conversation.\nType /help for available commands.")
 
     async def _process_incoming_with_reply(self, chat_id: str, msg_id: int, inbound: InboundMessage) -> None:
@@ -243,7 +358,7 @@ class TelegramChannel(Channel):
         if not self._check_user(update.effective_user.id):
             return
 
-        text = update.message.text
+        text = self._strip_bot_username_from_leading_command(update.message.text.strip(), self._get_bot_username(context))
         chat_id = str(update.effective_chat.id)
         user_id = str(update.effective_user.id)
         msg_id = str(update.message.message_id)
@@ -267,6 +382,7 @@ class TelegramChannel(Channel):
             thread_ts=msg_id,
         )
         inbound.topic_id = topic_id
+        inbound = await self._attach_connection_identity(inbound)
 
         if self._main_loop and self._main_loop.is_running():
             fut = asyncio.run_coroutine_threadsafe(self._process_incoming_with_reply(chat_id, update.message.message_id, inbound), self._main_loop)
@@ -279,7 +395,7 @@ class TelegramChannel(Channel):
         if not self._check_user(update.effective_user.id):
             return
 
-        text = update.message.text.strip()
+        text = self._strip_bot_username_from_leading_command(update.message.text.strip(), self._get_bot_username(context))
         if not text:
             return
 
@@ -309,6 +425,7 @@ class TelegramChannel(Channel):
             thread_ts=msg_id,
         )
         inbound.topic_id = topic_id
+        inbound = await self._attach_connection_identity(inbound)
 
         if self._main_loop and self._main_loop.is_running():
             fut = asyncio.run_coroutine_threadsafe(self._process_incoming_with_reply(chat_id, update.message.message_id, inbound), self._main_loop)

backend/app/channels/wechat.py

@@ -22,6 +22,7 @@ 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__)
@@ -620,7 +621,7 @@ class WechatChannel(Channel):
             chat_id=chat_id,
             user_id=chat_id,
             text=text,
-            msg_type=InboundMessageType.COMMAND if text.startswith("/") else InboundMessageType.CHAT,
+            msg_type=InboundMessageType.COMMAND if is_known_channel_command(text) else InboundMessageType.CHAT,
             thread_ts=thread_ts,
             files=files,
             metadata={

backend/app/channels/wecom.py

@@ -8,6 +8,7 @@ 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,
@@ -270,7 +271,7 @@ class WeComChannel(Channel):
 
         user_id = (body.get("from") or {}).get("userid")
 
-        inbound_type = InboundMessageType.COMMAND if text.startswith("/") else InboundMessageType.CHAT
+        inbound_type = InboundMessageType.COMMAND if is_known_channel_command(text) 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

@@ -6,6 +6,7 @@ from contextlib import asynccontextmanager
 from fastapi import FastAPI
 from fastapi.middleware.cors import CORSMiddleware
 
+from app.gateway.auth_disabled import warn_if_auth_disabled_enabled
 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
@@ -15,6 +16,7 @@ from app.gateway.routers import (
     artifacts,
     assistants_compat,
     auth,
+    channel_connections,
     channels,
     feedback,
     mcp,
@@ -172,6 +174,7 @@ async def lifespan(app: FastAPI) -> AsyncGenerator[None, None]:
         startup_config = get_app_config()
         apply_logging_level(startup_config.log_level)
         logger.info("Configuration loaded successfully")
+        warn_if_auth_disabled_enabled()
     except Exception as e:
         error_msg = f"Failed to load configuration during gateway startup: {e}"
         logger.exception(error_msg)
@@ -376,6 +379,9 @@ This gateway provides runtime endpoints for agent runs plus custom endpoints for
     # Suggestions API is mounted at /api/threads/{thread_id}/suggestions
     app.include_router(suggestions.router)
 
+    # User-facing IM channel connection API is mounted at /api/channels
+    app.include_router(channel_connections.router)
+
     # Channels API is mounted at /api/channels
     app.include_router(channels.router)
 

backend/app/gateway/auth_disabled.py

@@ -0,0 +1,54 @@
+"""Shared helpers for local/E2E auth-disabled mode."""
+
+from __future__ import annotations
+
+import logging
+import os
+from types import SimpleNamespace
+
+AUTH_DISABLED_ENV_VAR = "DEER_FLOW_AUTH_DISABLED"
+AUTH_DISABLED_USER_ID = "e2e-user"
+AUTH_DISABLED_USER_EMAIL = "e2e@test.local"
+
+AUTH_SOURCE_SESSION = "session"
+AUTH_SOURCE_INTERNAL = "internal"
+AUTH_SOURCE_AUTH_DISABLED = "auth_disabled"
+
+_PRODUCTION_ENV_VARS: tuple[str, ...] = ("DEER_FLOW_ENV", "ENVIRONMENT")
+_PRODUCTION_ENV_VALUES: frozenset[str] = frozenset({"prod", "production"})
+
+logger = logging.getLogger(__name__)
+
+
+def is_explicit_production_environment() -> bool:
+    return any(os.environ.get(name, "").strip().lower() in _PRODUCTION_ENV_VALUES for name in _PRODUCTION_ENV_VARS)
+
+
+def is_auth_disabled_requested() -> bool:
+    return os.environ.get(AUTH_DISABLED_ENV_VAR) == "1"
+
+
+def is_auth_disabled() -> bool:
+    return is_auth_disabled_requested() and not is_explicit_production_environment()
+
+
+def warn_if_auth_disabled_enabled() -> None:
+    if not is_auth_disabled():
+        return
+
+    logger.warning(
+        "%s=1 is active: authentication is bypassed and anonymous requests run as synthetic admin user %r. Do not enable this in shared or production deployments.",
+        AUTH_DISABLED_ENV_VAR,
+        AUTH_DISABLED_USER_ID,
+    )
+
+
+def get_auth_disabled_user():
+    return SimpleNamespace(
+        id=AUTH_DISABLED_USER_ID,
+        email=AUTH_DISABLED_USER_EMAIL,
+        password_hash=None,
+        system_role="admin",
+        needs_setup=False,
+        token_version=0,
+    )

backend/app/gateway/auth_middleware.py

@@ -17,6 +17,13 @@ from starlette.responses import JSONResponse
 from starlette.types import ASGIApp
 
 from app.gateway.auth.errors import AuthErrorCode, AuthErrorResponse
+from app.gateway.auth_disabled import (
+    AUTH_SOURCE_AUTH_DISABLED,
+    AUTH_SOURCE_INTERNAL,
+    AUTH_SOURCE_SESSION,
+    get_auth_disabled_user,
+    is_auth_disabled,
+)
 from app.gateway.authz import _ALL_PERMISSIONS, AuthContext
 from app.gateway.internal_auth import INTERNAL_AUTH_HEADER_NAME, get_internal_user, is_valid_internal_auth_token
 from deerflow.runtime.user_context import reset_current_user, set_current_user
@@ -80,8 +87,38 @@ class AuthMiddleware(BaseHTTPMiddleware):
         if is_valid_internal_auth_token(request.headers.get(INTERNAL_AUTH_HEADER_NAME)):
             internal_user = get_internal_user()
 
+        auth_source = AUTH_SOURCE_SESSION
+        access_token = request.cookies.get("access_token")
+
         # Non-public path: require session cookie
-        if internal_user is None and not request.cookies.get("access_token"):
+        if internal_user is not None:
+            user = internal_user
+            auth_source = AUTH_SOURCE_INTERNAL
+        elif access_token:
+            # Strict JWT validation: reject junk/expired tokens with 401
+            # right here instead of silently passing through. This closes
+            # the "junk cookie bypass" gap (AUTH_TEST_PLAN test 7.5.8):
+            # without this, non-isolation routes like /api/models would
+            # accept any cookie-shaped string as authentication.
+            #
+            # We call the *strict* resolver so that fine-grained error
+            # codes (token_expired, token_invalid, user_not_found, …)
+            # propagate from AuthErrorCode, not get flattened into one
+            # generic code. BaseHTTPMiddleware doesn't let HTTPException
+            # bubble up, so we catch and render it as JSONResponse here.
+            from app.gateway.deps import get_current_user_from_request
+
+            try:
+                user = await get_current_user_from_request(request)
+            except HTTPException as exc:
+                if not is_auth_disabled():
+                    return JSONResponse(status_code=exc.status_code, content={"detail": exc.detail})
+                user = get_auth_disabled_user()
+                auth_source = AUTH_SOURCE_AUTH_DISABLED
+        elif is_auth_disabled():
+            user = get_auth_disabled_user()
+            auth_source = AUTH_SOURCE_AUTH_DISABLED
+        else:
             return JSONResponse(
                 status_code=401,
                 content={
@@ -92,32 +129,12 @@ class AuthMiddleware(BaseHTTPMiddleware):
                 },
             )
 
-        # Strict JWT validation: reject junk/expired tokens with 401
-        # right here instead of silently passing through. This closes
-        # the "junk cookie bypass" gap (AUTH_TEST_PLAN test 7.5.8):
-        # without this, non-isolation routes like /api/models would
-        # accept any cookie-shaped string as authentication.
-        #
-        # We call the *strict* resolver so that fine-grained error
-        # codes (token_expired, token_invalid, user_not_found, …)
-        # propagate from AuthErrorCode, not get flattened into one
-        # generic code. BaseHTTPMiddleware doesn't let HTTPException
-        # bubble up, so we catch and render it as JSONResponse here.
-        from app.gateway.deps import get_current_user_from_request
-
-        if internal_user is not None:
-            user = internal_user
-        else:
-            try:
-                user = await get_current_user_from_request(request)
-            except HTTPException as exc:
-                return JSONResponse(status_code=exc.status_code, content={"detail": exc.detail})
-
         # Stamp both request.state.user (for the contextvar pattern)
         # and request.state.auth (so @require_permission's "auth is
         # None" branch short-circuits instead of running the entire
         # JWT-decode + DB-lookup pipeline a second time per request).
         request.state.user = user
+        request.state.auth_source = auth_source
         request.state.auth = AuthContext(user=user, permissions=_ALL_PERMISSIONS)
         token = set_current_user(user)
         try:

backend/app/gateway/csrf_middleware.py

@@ -14,6 +14,8 @@ from starlette.middleware.base import BaseHTTPMiddleware
 from starlette.responses import JSONResponse
 from starlette.types import ASGIApp
 
+from app.gateway.auth_disabled import is_auth_disabled
+
 CSRF_COOKIE_NAME = "csrf_token"
 CSRF_HEADER_NAME = "X-CSRF-Token"
 CSRF_TOKEN_LENGTH = 64  # bytes
@@ -38,6 +40,9 @@ def should_check_csrf(request: Request) -> bool:
     if request.method not in ("POST", "PUT", "DELETE", "PATCH"):
         return False
 
+    if is_auth_disabled():
+        return False
+
     path = request.url.path.rstrip("/")
     # Exempt /api/v1/auth/me endpoint
     if path == "/api/v1/auth/me":

backend/app/gateway/deps.py

@@ -331,6 +331,17 @@ async def get_current_user_from_request(request: Request):
 
     Raises HTTPException 401 if not authenticated.
     """
+    state = getattr(request, "state", None)
+    state_user = getattr(state, "user", None)
+    from app.gateway.auth_disabled import AUTH_SOURCE_AUTH_DISABLED, AUTH_SOURCE_INTERNAL, AUTH_SOURCE_SESSION
+
+    if state_user is not None and getattr(state, "auth_source", None) in {
+        AUTH_SOURCE_SESSION,
+        AUTH_SOURCE_AUTH_DISABLED,
+        AUTH_SOURCE_INTERNAL,
+    }:
+        return state_user
+
     from app.gateway.auth import decode_token
     from app.gateway.auth.errors import AuthErrorCode, AuthErrorResponse, TokenError, token_error_to_code
 

backend/app/gateway/langgraph_auth.py

@@ -20,6 +20,7 @@ from langgraph_sdk import Auth
 
 from app.gateway.auth.errors import TokenError
 from app.gateway.auth.jwt import decode_token
+from app.gateway.auth_disabled import AUTH_DISABLED_USER_ID, is_auth_disabled
 from app.gateway.deps import get_local_provider
 
 auth = Auth()
@@ -38,6 +39,9 @@ def _check_csrf(request) -> None:
     if method.upper() not in _CSRF_METHODS:
         return
 
+    if is_auth_disabled():
+        return
+
     cookie_token = request.cookies.get("csrf_token")
     header_token = request.headers.get("x-csrf-token")
 
@@ -66,6 +70,9 @@ async def authenticate(request):
     # are rejected early, even if the cookie carries a valid JWT.
     _check_csrf(request)
 
+    if is_auth_disabled():
+        return AUTH_DISABLED_USER_ID
+
     token = request.cookies.get("access_token")
     if not token:
         raise Auth.exceptions.HTTPException(

backend/app/gateway/routers/agents.py

@@ -1,5 +1,6 @@
 """CRUD API for custom agents."""
 
+import asyncio
 import logging
 import re
 import shutil
@@ -213,48 +214,61 @@ async def create_agent_endpoint(request: AgentCreateRequest) -> AgentResponse:
     user_id = get_effective_user_id()
     paths = get_paths()
 
-    agent_dir = paths.user_agent_dir(user_id, normalized_name)
+    def _create_agent() -> AgentResponse | None:
-    legacy_dir = 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 agent_dir.exists() or legacy_dir.exists():
+        if legacy_dir.exists():
-        raise HTTPException(status_code=409, detail=f"Agent '{normalized_name}' already 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:
-        agent_dir.mkdir(parents=True, exist_ok=True)
+        response = await asyncio.to_thread(_create_agent)
-
-        # 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 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)}")
 
+    if response is None:
+        raise HTTPException(status_code=409, detail=f"Agent '{normalized_name}' already exists")
+
+    return response
+
 
 @router.put(
     "/agents/{name}",
@@ -428,19 +442,30 @@ async def delete_agent(name: str) -> None:
     name = _normalize_agent_name(name)
     user_id = get_effective_user_id()
     paths = get_paths()
-    agent_dir = paths.user_agent_dir(user_id, name)
 
-    if not agent_dir.exists():
+    def _remove_agent_dir() -> tuple[str, str]:
-        if paths.agent_dir(name).exists():
+        # Runs in a worker thread: resolving the base dir, probing the directory
-            raise HTTPException(
+        # (`exists`), and removing it (`rmtree`) are all blocking filesystem IO
-                status_code=409,
+        # that must stay off the event loop.
-                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."),
+        agent_dir = paths.user_agent_dir(user_id, name)
-            )
+        if not agent_dir.exists():
-        raise HTTPException(status_code=404, detail=f"Agent '{name}' not found")
+            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:
-        shutil.rmtree(agent_dir)
+        outcome, agent_dir = await asyncio.to_thread(_remove_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/auth.py

@@ -341,9 +341,19 @@ async def change_password(request: Request, response: Response, body: ChangePass
     - Re-issues session cookie with new token_version
     """
     from app.gateway.auth.password import hash_password_async, verify_password_async
+    from app.gateway.auth_disabled import AUTH_SOURCE_AUTH_DISABLED
 
     user = await get_current_user_from_request(request)
 
+    if getattr(request.state, "auth_source", None) == AUTH_SOURCE_AUTH_DISABLED:
+        raise HTTPException(
+            status_code=status.HTTP_400_BAD_REQUEST,
+            detail=AuthErrorResponse(
+                code=AuthErrorCode.INVALID_CREDENTIALS,
+                message="Password changes are not available when DEER_FLOW_AUTH_DISABLED=1.",
+            ).model_dump(),
+        )
+
     if user.password_hash is None:
         raise HTTPException(status_code=status.HTTP_400_BAD_REQUEST, detail=AuthErrorResponse(code=AuthErrorCode.INVALID_CREDENTIALS, message="OAuth users cannot change password").model_dump())
 

backend/app/gateway/routers/channel_connections.py

@@ -0,0 +1,296 @@
+"""Browser-facing APIs for user-owned IM channel bindings."""
+
+from __future__ import annotations
+
+import secrets
+from datetime import UTC, datetime, timedelta
+from typing import Any
+
+from fastapi import APIRouter, HTTPException, Request, Response
+from pydantic import BaseModel, Field
+
+from deerflow.config.channel_connections_config import ChannelConnectionsConfig
+from deerflow.persistence.channel_connections import ChannelConnectionRepository
+from deerflow.persistence.engine import get_session_factory
+
+router = APIRouter(prefix="/api/channels", tags=["channel-connections"])
+
+_STATE_TTL_SECONDS = 600
+
+
+class ChannelProviderResponse(BaseModel):
+    provider: str
+    display_name: str
+    enabled: bool
+    configured: bool
+    connectable: bool
+    unavailable_reason: str | None = None
+    auth_mode: str
+    connection_status: str
+
+
+class ChannelProvidersResponse(BaseModel):
+    enabled: bool
+    providers: list[ChannelProviderResponse]
+
+
+class ChannelConnectionResponse(BaseModel):
+    id: str
+    provider: str
+    status: str
+    external_account_id: str | None = None
+    external_account_name: str | None = None
+    workspace_id: str | None = None
+    workspace_name: str | None = None
+    scopes: list[str] = Field(default_factory=list)
+    metadata: dict[str, Any] = Field(default_factory=dict)
+
+
+class ChannelConnectionsResponse(BaseModel):
+    connections: list[ChannelConnectionResponse]
+
+
+class ChannelConnectResponse(BaseModel):
+    provider: str
+    mode: str
+    url: str | None = None
+    code: str
+    instruction: str
+    expires_in: int
+
+
+_PROVIDER_META: dict[str, dict[str, str]] = {
+    "telegram": {"display_name": "Telegram", "auth_mode": "deep_link"},
+    "slack": {"display_name": "Slack", "auth_mode": "binding_code"},
+    "discord": {"display_name": "Discord", "auth_mode": "binding_code"},
+}
+
+_RUNTIME_REQUIREMENTS: dict[str, tuple[str, ...]] = {
+    "telegram": ("bot_token",),
+    "slack": ("bot_token", "app_token"),
+    "discord": ("bot_token",),
+}
+
+
+def _get_user_id(request: Request) -> str:
+    user = getattr(request.state, "user", None)
+    if user is None:
+        raise HTTPException(status_code=401, detail="Authentication required")
+    return str(user.id)
+
+
+def _get_app_config():
+    from deerflow.config.app_config import get_app_config
+
+    return get_app_config()
+
+
+def _get_channel_connections_config(request: Request) -> ChannelConnectionsConfig:
+    config = getattr(request.app.state, "channel_connections_config", None)
+    if isinstance(config, ChannelConnectionsConfig):
+        return config
+    return _get_app_config().channel_connections
+
+
+def _get_channels_config(request: Request) -> dict[str, Any]:
+    state_config = getattr(request.app.state, "channels_config", None)
+    if isinstance(state_config, dict):
+        return state_config
+
+    app_config = _get_app_config()
+    extra = app_config.model_extra or {}
+    channels_config = extra.get("channels")
+    return dict(channels_config) if isinstance(channels_config, dict) else {}
+
+
+def _get_repository(request: Request, config: ChannelConnectionsConfig) -> ChannelConnectionRepository:
+    repo = getattr(request.app.state, "channel_connection_repo", None)
+    if isinstance(repo, ChannelConnectionRepository):
+        return repo
+
+    sf = get_session_factory()
+    if sf is None:
+        raise HTTPException(status_code=503, detail="Channel connection persistence is not available")
+
+    repo = ChannelConnectionRepository(sf)
+    request.app.state.channel_connection_repo = repo
+    return repo
+
+
+def _provider_config(config: ChannelConnectionsConfig, provider: str):
+    provider_config = getattr(config, provider, None)
+    if provider_config is None:
+        raise HTTPException(status_code=404, detail="Unknown channel provider")
+    return provider_config
+
+
+def _runtime_channel_configured(provider: str, channels_config: dict[str, Any]) -> bool:
+    runtime_config = channels_config.get(provider)
+    if not isinstance(runtime_config, dict) or not runtime_config.get("enabled", False):
+        return False
+    return all(str(runtime_config.get(key) or "").strip() for key in _RUNTIME_REQUIREMENTS[provider])
+
+
+def _runtime_unavailable_reason(provider: str) -> str:
+    keys = " and ".join(f"channels.{provider}.{key}" for key in _RUNTIME_REQUIREMENTS[provider])
+    return f"Enable and configure channels.{provider} with {keys}."
+
+
+def _provider_unavailable_reason(
+    config: ChannelConnectionsConfig,
+    channels_config: dict[str, Any],
+    provider: str,
+) -> str | None:
+    provider_config = _provider_config(config, provider)
+    if not provider_config.enabled:
+        return None
+    if not provider_config.configured:
+        if provider == "telegram":
+            return "Configure channel_connections.telegram.bot_username for Telegram deep links."
+        return f"Configure channel_connections.{provider}."
+    if not _runtime_channel_configured(provider, channels_config):
+        return _runtime_unavailable_reason(provider)
+    return None
+
+
+def _provider_status(
+    config: ChannelConnectionsConfig,
+    channels_config: dict[str, Any],
+    provider: str,
+) -> tuple[dict[str, bool], str | None]:
+    declared = config.provider_status(provider)
+    unavailable_reason = _provider_unavailable_reason(config, channels_config, provider)
+    configured = declared["configured"] and _runtime_channel_configured(provider, channels_config)
+    return {"enabled": declared["enabled"], "configured": configured}, unavailable_reason
+
+
+def _new_binding_code() -> str:
+    return secrets.token_urlsafe(16)
+
+
+async def _create_state(
+    repo: ChannelConnectionRepository,
+    *,
+    owner_user_id: str,
+    provider: str,
+) -> str:
+    state = _new_binding_code()
+    await repo.create_oauth_state(
+        owner_user_id=owner_user_id,
+        provider=provider,
+        state=state,
+        expires_at=datetime.now(UTC) + timedelta(seconds=_STATE_TTL_SECONDS),
+    )
+    return state
+
+
+def _connect_instruction(provider: str, code: str) -> str:
+    if provider == "telegram":
+        return f"Send /start {code} to the DeerFlow Telegram bot."
+    if provider == "slack":
+        return f"Send /connect {code} to the DeerFlow Slack bot."
+    if provider == "discord":
+        return f"Send /connect {code} to the DeerFlow Discord bot."
+    raise HTTPException(status_code=404, detail="Unknown channel provider")
+
+
+def _connect_url(config: ChannelConnectionsConfig, provider: str, code: str) -> str | None:
+    if provider == "telegram":
+        provider_config = _provider_config(config, provider)
+        return f"https://t.me/{provider_config.bot_username}?start={code}"
+    if provider in {"slack", "discord"}:
+        return None
+    raise HTTPException(status_code=404, detail="Unknown channel provider")
+
+
+@router.get("/providers", response_model=ChannelProvidersResponse)
+async def get_channel_providers(request: Request) -> ChannelProvidersResponse:
+    config = _get_channel_connections_config(request)
+    channels_config = _get_channels_config(request)
+    repo = None
+    if config.enabled:
+        try:
+            repo = _get_repository(request, config)
+        except HTTPException as exc:
+            if exc.status_code != 503:
+                raise
+    owner_user_id = _get_user_id(request)
+    connections = await repo.list_connections(owner_user_id) if repo is not None else []
+    by_provider: dict[str, dict[str, Any]] = {}
+    for item in connections:
+        by_provider.setdefault(item["provider"], item)
+
+    providers: list[ChannelProviderResponse] = []
+    for provider, meta in _PROVIDER_META.items():
+        status, unavailable_reason = _provider_status(config, channels_config, provider)
+        connection = by_provider.get(provider)
+        providers.append(
+            ChannelProviderResponse(
+                provider=provider,
+                display_name=meta["display_name"],
+                enabled=status["enabled"],
+                configured=status["configured"],
+                connectable=status["enabled"] and status["configured"] and unavailable_reason is None,
+                unavailable_reason=unavailable_reason,
+                auth_mode=meta["auth_mode"],
+                connection_status=connection["status"] if connection else "not_connected",
+            )
+        )
+    return ChannelProvidersResponse(enabled=config.enabled, providers=providers)
+
+
+@router.get("/connections", response_model=ChannelConnectionsResponse)
+async def get_channel_connections(request: Request) -> ChannelConnectionsResponse:
+    config = _get_channel_connections_config(request)
+    if not config.enabled:
+        return ChannelConnectionsResponse(connections=[])
+    repo = _get_repository(request, config)
+    rows = await repo.list_connections(_get_user_id(request))
+    return ChannelConnectionsResponse(connections=[ChannelConnectionResponse(**row) for row in rows])
+
+
+@router.delete("/connections/{connection_id}", status_code=204)
+async def disconnect_channel_connection(connection_id: str, request: Request) -> Response:
+    config = _get_channel_connections_config(request)
+    if not config.enabled:
+        raise HTTPException(status_code=400, detail="Channel connections are disabled")
+
+    repo = _get_repository(request, config)
+    disconnected = await repo.disconnect_connection(
+        connection_id=connection_id,
+        owner_user_id=_get_user_id(request),
+    )
+    if not disconnected:
+        raise HTTPException(status_code=404, detail="Channel connection not found")
+    return Response(status_code=204)
+
+
+@router.post("/{provider}/connect", response_model=ChannelConnectResponse)
+async def connect_channel_provider(provider: str, request: Request) -> ChannelConnectResponse:
+    config = _get_channel_connections_config(request)
+    channels_config = _get_channels_config(request)
+    if not config.enabled:
+        raise HTTPException(status_code=400, detail="Channel connections are disabled")
+
+    status, unavailable_reason = _provider_status(config, channels_config, provider)
+    if not status["enabled"]:
+        raise HTTPException(status_code=400, detail="Channel provider is not enabled")
+    if unavailable_reason:
+        raise HTTPException(status_code=400, detail=unavailable_reason)
+    if not status["configured"]:
+        raise HTTPException(status_code=400, detail="Channel provider is not configured")
+
+    repo = _get_repository(request, config)
+    code = await _create_state(
+        repo,
+        owner_user_id=_get_user_id(request),
+        provider=provider,
+    )
+    return ChannelConnectResponse(
+        provider=provider,
+        mode=_PROVIDER_META[provider]["auth_mode"],
+        url=_connect_url(config, provider, code),
+        code=code,
+        instruction=_connect_instruction(provider, code),
+        expires_in=_STATE_TTL_SECONDS,
+    )

backend/docs/IM_CHANNEL_CONNECTIONS.md

@@ -0,0 +1,84 @@
+# IM Channel Connections
+
+DeerFlow supports user-owned IM channel bindings for Telegram, Slack, and Discord. The feature reuses the existing `channels.*` runtime configuration, so it works in local and private deployments with the same outbound transports already supported by DeerFlow.
+
+No public IP, OAuth callback URL, or provider webhook is required in this implementation.
+
+## Configuration
+
+Configure the actual IM bots under the existing `channels` block:
+
+```yaml
+channels:
+  telegram:
+    enabled: true
+    bot_token: $TELEGRAM_BOT_TOKEN
+
+  slack:
+    enabled: true
+    bot_token: $SLACK_BOT_TOKEN
+    app_token: $SLACK_APP_TOKEN
+
+  discord:
+    enabled: true
+    bot_token: $DISCORD_BOT_TOKEN
+```
+
+Then enable user bindings in `channel_connections`:
+
+```yaml
+channel_connections:
+  enabled: true
+
+  telegram:
+    enabled: true
+    bot_username: $TELEGRAM_BOT_USERNAME
+
+  slack:
+    enabled: true
+
+  discord:
+    enabled: true
+```
+
+`channel_connections` does not duplicate provider secrets. It only controls the browser-facing connect UI and stores per-user binding records. Telegram needs `bot_username` only so the frontend can open a deep link.
+
+## Connect Flow
+
+Telegram:
+
+- The frontend creates a short one-time code.
+- The Connect button opens `https://t.me/<bot_username>?start=<code>`.
+- The existing Telegram long-polling worker receives `/start <code>` and binds that Telegram chat/user to the current DeerFlow user.
+
+Slack:
+
+- The frontend creates a short one-time code.
+- The UI shows `Send /connect <code> to the DeerFlow Slack bot.`
+- The existing Slack Socket Mode worker receives the message and binds the Slack user/team to the current DeerFlow user.
+
+Discord:
+
+- The frontend creates a short one-time code.
+- The UI shows `Send /connect <code> to the DeerFlow Discord bot.`
+- The existing Discord Gateway worker receives the message and binds the Discord user/guild to the current DeerFlow user.
+
+Codes expire after 10 minutes and are single-use.
+
+## Runtime Model
+
+Connection records live in SQL tables under `deerflow.persistence.channel_connections`:
+
+- `channel_connections`: owner user, provider identity, workspace/guild/team, status, metadata.
+- `channel_oauth_states`: one-time connect codes and Telegram deep-link state.
+- `channel_conversations`: connection-scoped IM conversation to DeerFlow thread mapping.
+- `channel_credentials`: reserved for future provider-token flows, not used by the local/private binding flow.
+
+Incoming messages that resolve to a connection carry `connection_id`, `owner_user_id`, and `workspace_id`. `ChannelManager` uses `owner_user_id` as the DeerFlow run user id and preserves the raw platform user id as `channel_user_id`.
+
+## Security Notes
+
+- Browser APIs remain authenticated and CSRF-protected.
+- Connect codes are random, short-lived, and single-use.
+- Provider bot tokens remain in `channels.*` and are never returned to the browser.
+- This implementation does not add public provider callback or webhook routes.

backend/docs/REPLAY_E2E.md

@@ -50,18 +50,22 @@ gateway's own run/event stores using the request's auth context, so the real
 ## How replay works
 
 `tests/replay_provider.py::ReplayChatModel` returns recorded assistant turns keyed
-by a **normalized hash of the conversation** (human / ai / tool messages — role,
+by a **normalized hash of the model caller + conversation**. The conversation is
-text, tool-call name+args; with `<system-reminder>`, dates, UUIDs, tmp paths
+human / ai / tool messages — role, text, tool-call name+args; with
-stripped). A miss raises loudly rather than passing silently.
+`<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. (This mirrors how open-design's
+stable contract that identifies a recorded turn. The caller still stays in the
-mock picker keys on the user prompt, not the system internals.) Combined with
+key so two different model users with identical conversation text do not compete
-pinning skills + extensions empty and disabling memory/summarization
+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**.
 

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/lead_agent/agent.py

@@ -49,6 +49,8 @@ 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."""
@@ -265,21 +267,31 @@ 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 middleware chain based on runtime configuration.
+    """Build the lead-agent 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.
@@ -293,6 +305,13 @@ def _build_middlewares(
 
     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:
@@ -360,7 +379,7 @@ def _build_middlewares(
 
 def _available_skill_names(agent_config, is_bootstrap: bool) -> set[str] | None:
     if is_bootstrap:
-        return {"bootstrap"}
+        return set(_BOOTSTRAP_SKILL_NAMES)
     if agent_config and agent_config.skills is not None:
         return set(agent_config.skills)
     return None
@@ -466,17 +485,25 @@ def _make_lead_agent(config: RunnableConfig, *, app_config: AppConfig):
 
     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),
             tools=final_tools,
-            middleware=_build_middlewares(config, model_name=model_name, app_config=resolved_app_config, deferred_setup=setup),
+            middleware=build_middlewares(
+                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"]),
+                available_skills=set(_BOOTSTRAP_SKILL_NAMES),
                 app_config=resolved_app_config,
                 deferred_names=setup.deferred_names,
             ),
@@ -493,12 +520,19 @@ def _make_lead_agent(config: RunnableConfig, *, app_config: AppConfig):
     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),
         tools=final_tools,
-        middleware=_build_middlewares(config, model_name=model_name, agent_name=agent_name, app_config=resolved_app_config, deferred_setup=setup),
+        middleware=build_middlewares(
+            config,
+            model_name=model_name,
+            agent_name=agent_name,
+            available_skills=available_skills,
+            app_config=resolved_app_config,
+            deferred_setup=setup,
+        ),
         system_prompt=apply_prompt_template(
             subagent_enabled=subagent_enabled,
             max_concurrent_subagents=max_concurrent_subagents,
             agent_name=agent_name,
-            available_skills=set(agent_config.skills) if agent_config and agent_config.skills is not None else None,
+            available_skills=available_skills,
             app_config=resolved_app_config,
             deferred_names=setup.deferred_names,
         ),

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

@@ -625,6 +625,11 @@ 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}

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

@@ -0,0 +1,289 @@
+"""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/uploads_middleware.py

@@ -13,6 +13,7 @@ from langgraph.runtime import Runtime
 from deerflow.config.paths import Paths, get_paths
 from deerflow.runtime.user_context import get_effective_user_id
 from deerflow.utils.file_conversion import extract_outline
+from deerflow.utils.messages import ORIGINAL_USER_CONTENT_KEY, message_content_to_text
 
 logger = logging.getLogger(__name__)
 
@@ -265,6 +266,8 @@ class UploadsMiddleware(AgentMiddleware[UploadsMiddlewareState]):
 
         # Extract original content - handle both string and list formats
         original_content = last_message.content
+        additional_kwargs = dict(last_message.additional_kwargs or {})
+        additional_kwargs.setdefault(ORIGINAL_USER_CONTENT_KEY, message_content_to_text(original_content))
         if isinstance(original_content, str):
             # Simple case: string content, just prepend files message
             updated_content = f"{files_message}\n\n{original_content}"
@@ -285,7 +288,7 @@ class UploadsMiddleware(AgentMiddleware[UploadsMiddlewareState]):
             content=updated_content,
             id=last_message.id,
             name=last_message.name,
-            additional_kwargs=last_message.additional_kwargs,
+            additional_kwargs=additional_kwargs,
         )
 
         messages[last_message_index] = updated_message

backend/packages/harness/deerflow/client.py

@@ -33,7 +33,7 @@ from langchain.agents.middleware import AgentMiddleware
 from langchain_core.messages import AIMessage, HumanMessage, SystemMessage, ToolMessage
 from langchain_core.runnables import RunnableConfig
 
-from deerflow.agents.lead_agent.agent import _build_middlewares
+from deerflow.agents.lead_agent.agent import build_middlewares
 from deerflow.agents.lead_agent.prompt import apply_prompt_template
 from deerflow.agents.thread_state import ThreadState
 from deerflow.config.agents_config import AGENT_NAME_PATTERN
@@ -247,7 +247,15 @@ class DeerFlowClient:
             # Attaching them again on the model would emit duplicate spans.
             "model": create_chat_model(name=model_name, thinking_enabled=thinking_enabled, attach_tracing=False),
             "tools": final_tools,
-            "middleware": _build_middlewares(config, model_name=model_name, agent_name=self._agent_name, custom_middlewares=self._middlewares, deferred_setup=deferred_setup),
+            "middleware": build_middlewares(
+                config,
+                model_name=model_name,
+                agent_name=self._agent_name,
+                available_skills=self._available_skills,
+                custom_middlewares=self._middlewares,
+                app_config=self._app_config,
+                deferred_setup=deferred_setup,
+            ),
             "system_prompt": apply_prompt_template(
                 subagent_enabled=subagent_enabled,
                 max_concurrent_subagents=max_concurrent_subagents,

backend/packages/harness/deerflow/community/jina_ai/jina_client.py

@@ -9,7 +9,7 @@ _api_key_warned = False
 
 
 class JinaClient:
-    async def crawl(self, url: str, return_format: str = "html", timeout: int = 10) -> str:
+    async def crawl(self, url: str, return_format: str = "html", timeout: int = 10, proxy: str | None = None, trust_env: bool = True) -> str:
         global _api_key_warned
         headers = {
             "Content-Type": "application/json",
@@ -23,7 +23,10 @@ class JinaClient:
             logger.warning("Jina API key is not set. Provide your own key to access a higher rate limit. See https://jina.ai/reader for more information.")
         data = {"url": url}
         try:
-            async with httpx.AsyncClient() as client:
+            client_kwargs: dict[str, object] = {"trust_env": trust_env}
+            if proxy:
+                client_kwargs["proxy"] = proxy
+            async with httpx.AsyncClient(**client_kwargs) as client:
                 response = await client.post("https://r.jina.ai/", headers=headers, json=data, timeout=timeout)
 
             if response.status_code != 200:

backend/packages/harness/deerflow/community/jina_ai/tools.py

@@ -9,6 +9,38 @@ from deerflow.utils.readability import ReadabilityExtractor
 readability_extractor = ReadabilityExtractor()
 
 
+def _coerce_bool(value: object, default: bool) -> bool:
+    if isinstance(value, bool):
+        return value
+    if isinstance(value, str):
+        normalized = value.strip().lower()
+        if normalized in {"1", "true", "yes", "on"}:
+            return True
+        if normalized in {"0", "false", "no", "off"}:
+            return False
+    return default
+
+
+def _coerce_timeout(value: object, default: int) -> int:
+    if isinstance(value, bool):
+        return default
+    if isinstance(value, int):
+        return value
+    if isinstance(value, str):
+        try:
+            return int(value)
+        except ValueError:
+            return default
+    return default
+
+
+def _coerce_proxy(value: object) -> str | None:
+    if not isinstance(value, str):
+        return None
+    proxy = value.strip()
+    return proxy or None
+
+
 @tool("web_fetch", parse_docstring=True)
 async def web_fetch_tool(url: str) -> str:
     """Fetch the contents of a web page at a given URL.
@@ -22,10 +54,14 @@ async def web_fetch_tool(url: str) -> str:
     """
     jina_client = JinaClient()
     timeout = 10
+    proxy = None
+    trust_env = True
     config = get_app_config().get_tool_config("web_fetch")
-    if config is not None and "timeout" in config.model_extra:
+    if config is not None:
-        timeout = config.model_extra.get("timeout")
+        timeout = _coerce_timeout(config.model_extra.get("timeout"), timeout)
-    html_content = await jina_client.crawl(url, return_format="html", timeout=timeout)
+        proxy = _coerce_proxy(config.model_extra.get("proxy"))
+        trust_env = _coerce_bool(config.model_extra.get("trust_env"), trust_env)
+    html_content = await jina_client.crawl(url, return_format="html", timeout=timeout, proxy=proxy, trust_env=trust_env)
     if isinstance(html_content, str) and html_content.startswith("Error:"):
         return html_content
     article = await asyncio.to_thread(readability_extractor.extract_article, html_content)

backend/packages/harness/deerflow/config/app_config.py

@@ -7,10 +7,11 @@ from typing import Any, Self
 
 import yaml
 from dotenv import load_dotenv
-from pydantic import BaseModel, ConfigDict, Field
+from pydantic import BaseModel, ConfigDict, Field, field_validator
 
 from deerflow.config.acp_config import ACPAgentConfig, load_acp_config_from_dict
 from deerflow.config.agents_api_config import AgentsApiConfig, load_agents_api_config_from_dict
+from deerflow.config.channel_connections_config import ChannelConnectionsConfig
 from deerflow.config.checkpointer_config import CheckpointerConfig, load_checkpointer_config_from_dict
 from deerflow.config.database_config import DatabaseConfig
 from deerflow.config.extensions_config import ExtensionsConfig
@@ -116,6 +117,7 @@ class AppConfig(BaseModel):
     subagents: SubagentsAppConfig = Field(default_factory=SubagentsAppConfig, description="Subagent runtime configuration")
     guardrails: GuardrailsConfig = Field(default_factory=GuardrailsConfig, description="Guardrail middleware configuration")
     circuit_breaker: CircuitBreakerConfig = Field(default_factory=CircuitBreakerConfig, description="LLM circuit breaker configuration")
+    channel_connections: ChannelConnectionsConfig = Field(default_factory=ChannelConnectionsConfig, description="User-facing IM channel connection configuration")
     loop_detection: LoopDetectionConfig = Field(default_factory=LoopDetectionConfig, description="Loop detection middleware configuration")
     safety_finish_reason: SafetyFinishReasonConfig = Field(default_factory=SafetyFinishReasonConfig, description="Provider safety-filter finish_reason interception middleware configuration")
     model_config = ConfigDict(extra="allow")
@@ -148,6 +150,21 @@ class AppConfig(BaseModel):
         ),
     )
 
+    @field_validator("models", "tools", "tool_groups", mode="before")
+    @classmethod
+    def _coerce_null_list_sections(cls, value: Any) -> Any:
+        """Treat a present-but-empty config section as an empty list.
+
+        Commenting out every entry under a top-level YAML key — e.g. ``models:``
+        with only comments beneath it, exactly as shipped in
+        ``config.example.yaml`` — makes PyYAML parse the value as ``None``.
+        Without this, the documented ``cp config.example.yaml config.yaml``
+        first-run flow crashes with an opaque ``Input should be a valid list``
+        pydantic error. Coercing ``None`` to ``[]`` keeps that flow working and
+        matches the field's own ``default_factory=list``.
+        """
+        return [] if value is None else value
+
     @classmethod
     def resolve_config_path(cls, config_path: str | None = None) -> Path:
         """Resolve the config file path.
@@ -209,6 +226,11 @@ class AppConfig(BaseModel):
         config_data["extensions"] = extensions_config.model_dump()
 
         result = cls.model_validate(config_data)
+        if not result.models:
+            logger.warning(
+                "No models are configured in %s. Add at least one entry under `models:` (see the commented examples in config.example.yaml) or run `make setup`.",
+                resolved_path,
+            )
         acp_agents = cls._validate_acp_agents(config_data.get("acp_agents", {}))
         cls._apply_singleton_configs(result, acp_agents)
         return result

backend/packages/harness/deerflow/config/channel_connections_config.py

@@ -0,0 +1,49 @@
+"""Configuration for user-owned IM channel connections."""
+
+from __future__ import annotations
+
+from pydantic import BaseModel, Field
+
+
+class SlackChannelConnectionConfig(BaseModel):
+    enabled: bool = False
+
+    @property
+    def configured(self) -> bool:
+        return True
+
+
+class TelegramChannelConnectionConfig(BaseModel):
+    enabled: bool = False
+    bot_username: str = ""
+
+    @property
+    def configured(self) -> bool:
+        return bool(self.bot_username)
+
+
+class DiscordChannelConnectionConfig(BaseModel):
+    enabled: bool = False
+
+    @property
+    def configured(self) -> bool:
+        return True
+
+
+class ChannelConnectionsConfig(BaseModel):
+    """Top-level config for browser-connectable IM channels."""
+
+    enabled: bool = False
+    slack: SlackChannelConnectionConfig = Field(default_factory=SlackChannelConnectionConfig)
+    telegram: TelegramChannelConnectionConfig = Field(default_factory=TelegramChannelConnectionConfig)
+    discord: DiscordChannelConnectionConfig = Field(default_factory=DiscordChannelConnectionConfig)
+
+    def provider_status(self, provider: str) -> dict[str, bool]:
+        config = getattr(self, provider, None)
+        if config is None:
+            return {"enabled": False, "configured": False}
+        enabled = bool(config.enabled)
+        return {
+            "enabled": enabled,
+            "configured": enabled and bool(config.configured),
+        }

backend/packages/harness/deerflow/config/paths.py

@@ -47,7 +47,7 @@ def make_safe_user_id(raw: str) -> str:
     sanitized = _UNSAFE_USER_ID_CHAR_RE.sub("-", raw)
     if sanitized == raw:
         return raw
-    digest = hashlib.sha1(raw.encode("utf-8")).hexdigest()[:_SAFE_USER_ID_DIGEST_HEX_LEN]
+    digest = hashlib.sha256(raw.encode("utf-8")).hexdigest()[:_SAFE_USER_ID_DIGEST_HEX_LEN]
     return f"{sanitized}-{digest}"
 
 

backend/packages/harness/deerflow/config/sandbox_config.py

@@ -4,7 +4,20 @@ from pydantic import BaseModel, ConfigDict, Field
 class VolumeMountConfig(BaseModel):
     """Configuration for a volume mount."""
 
-    host_path: str = Field(..., description="Path on the host machine")
+    host_path: str = Field(
+        ...,
+        description=(
+            "Source path for the mount. Resolution depends on the active provider: "
+            "``LocalSandboxProvider`` checks this path from the gateway process — in "
+            "``make dev`` that is the host machine, but in Docker deployments "
+            "(``make up`` / docker-compose) it is the path *inside* the "
+            "``deer-flow-gateway`` container, so the host directory must also be "
+            "bind-mounted into the gateway service for the mount to take effect. "
+            "``AioSandboxProvider`` (DooD) passes this value straight to ``docker -v`` "
+            "for the sandbox container, where it is resolved by the host Docker daemon "
+            "from the host machine's perspective."
+        ),
+    )
     container_path: str = Field(..., description="Path inside the container")
     read_only: bool = Field(default=False, description="Whether the mount is read-only")
 

backend/packages/harness/deerflow/models/patched_stepfun.py

@@ -0,0 +1,175 @@
+"""Patched ChatOpenAI adapter for StepFun reasoning models.
+
+StepFun returns ``reasoning`` (or ``reasoning_content`` with deepseek-style) in
+both streaming deltas and non-streaming responses. Standard ``ChatOpenAI``
+ignores these non-standard fields, so reasoning content is silently dropped.
+This adapter captures reasoning from all response paths and replays it on
+historical assistant messages for multi-turn tool-call conversations.
+"""
+
+from __future__ import annotations
+
+from collections.abc import Mapping
+from typing import Any
+
+from langchain_core.language_models import LanguageModelInput
+from langchain_core.messages import AIMessage, AIMessageChunk
+from langchain_core.outputs import ChatGeneration, ChatGenerationChunk, ChatResult
+from langchain_openai import ChatOpenAI
+
+from deerflow.models.assistant_payload_replay import (
+    restore_assistant_payloads,
+    restore_reasoning_content,
+)
+
+_MISSING = object()
+
+
+def _extract_reasoning(value: Any) -> str | object:
+    """Return reasoning content from a dict/Pydantic object.
+
+    StepFun may return reasoning via ``reasoning`` (default) or
+    ``reasoning_content`` (deepseek-style). Check both fields.
+    """
+    if isinstance(value, Mapping):
+        # Check reasoning_content first (deepseek-style), then reasoning (default)
+        for field in ("reasoning_content", "reasoning"):
+            if field in value and value[field] is not None:
+                return value[field]
+        return _MISSING
+
+    # Pydantic / SDK object attributes
+    for field in ("reasoning_content", "reasoning"):
+        attr = getattr(value, field, _MISSING)
+        if attr is not _MISSING and attr is not None:
+            return attr
+
+    # Some SDK versions store extra fields in model_extra
+    model_extra = getattr(value, "model_extra", None)
+    if isinstance(model_extra, Mapping):
+        for field in ("reasoning_content", "reasoning"):
+            if field in model_extra and model_extra[field] is not None:
+                return model_extra[field]
+
+    return _MISSING
+
+
+def _with_reasoning_content(message: AIMessage | AIMessageChunk, reasoning: str) -> AIMessage | AIMessageChunk:
+    """Return a copy of *message* with reasoning_content stored in additional_kwargs."""
+    additional_kwargs = dict(message.additional_kwargs)
+    if additional_kwargs.get("reasoning_content") != reasoning:
+        additional_kwargs["reasoning_content"] = reasoning
+    return message.model_copy(update={"additional_kwargs": additional_kwargs})
+
+
+def _get_typed_choice_message(response: Any, index: int) -> Any:
+    """Extract the SDK-typed choice message at *index*, if available."""
+    choices = getattr(response, "choices", None)
+    if choices is None:
+        return None
+    try:
+        return choices[index].message
+    except (AttributeError, IndexError, TypeError):
+        return None
+
+
+class PatchedChatStepFun(ChatOpenAI):
+    """ChatOpenAI with full reasoning support for StepFun models.
+
+    Captures ``reasoning`` / ``reasoning_content`` from both streaming and
+    non-streaming responses and replays it on historical assistant messages in
+    multi-turn tool-call conversations.
+    """
+
+    @classmethod
+    def is_lc_serializable(cls) -> bool:
+        return True
+
+    @property
+    def lc_secrets(self) -> dict[str, str]:
+        return {"api_key": "STEPFUN_API_KEY", "openai_api_key": "STEPFUN_API_KEY"}
+
+    # --- Request payload replay ---
+
+    def _get_request_payload(
+        self,
+        input_: LanguageModelInput,
+        *,
+        stop: list[str] | None = None,
+        **kwargs: Any,
+    ) -> dict:
+        """Restore ``reasoning_content`` on historical assistant messages."""
+        original_messages = self._convert_input(input_).to_messages()
+        payload = super()._get_request_payload(input_, stop=stop, **kwargs)
+
+        restore_assistant_payloads(
+            payload.get("messages", []),
+            original_messages,
+            restore_reasoning_content,
+        )
+
+        return payload
+
+    # --- Streaming reasoning capture ---
+
+    def _convert_chunk_to_generation_chunk(
+        self,
+        chunk: dict,
+        default_chunk_class: type,
+        base_generation_info: dict | None,
+    ) -> ChatGenerationChunk | None:
+        """Capture ``reasoning`` / ``reasoning_content`` from streaming deltas."""
+        generation_chunk = super()._convert_chunk_to_generation_chunk(
+            chunk,
+            default_chunk_class,
+            base_generation_info,
+        )
+        if generation_chunk is None:
+            return None
+
+        choices = chunk.get("choices", [])
+        if choices:
+            delta = choices[0].get("delta") or {}
+            reasoning = _extract_reasoning(delta)
+            if reasoning is not _MISSING and isinstance(generation_chunk.message, AIMessageChunk):
+                generation_chunk = ChatGenerationChunk(
+                    message=_with_reasoning_content(generation_chunk.message, reasoning),
+                    generation_info=generation_chunk.generation_info,
+                )
+
+        return generation_chunk
+
+    # --- Non-streaming reasoning capture ---
+
+    def _create_chat_result(
+        self,
+        response: dict | Any,
+        generation_info: dict | None = None,
+    ) -> ChatResult:
+        """Extract ``reasoning`` / ``reasoning_content`` from non-streaming responses."""
+        result = super()._create_chat_result(response, generation_info)
+        response_dict = response if isinstance(response, dict) else response.model_dump()
+        choices = response_dict.get("choices", [])
+
+        patched_generations: list[ChatGeneration] | None = None
+        for index, generation in enumerate(result.generations):
+            choice = choices[index] if index < len(choices) else {}
+            choice_message = choice.get("message", {}) if isinstance(choice, Mapping) else {}
+            reasoning = _extract_reasoning(choice_message)
+
+            if reasoning is _MISSING and not isinstance(response, dict):
+                reasoning = _extract_reasoning(_get_typed_choice_message(response, index))
+
+            message = generation.message
+            if reasoning is not _MISSING and isinstance(message, AIMessage):
+                if patched_generations is None:
+                    patched_generations = list(result.generations)
+                patched_generations[index] = ChatGeneration(
+                    message=_with_reasoning_content(message, reasoning),
+                    generation_info=generation.generation_info,
+                )
+
+        return ChatResult(
+            generations=patched_generations or result.generations,
+            llm_output=result.llm_output,
+        )

backend/packages/harness/deerflow/persistence/channel_connections/__init__.py

@@ -0,0 +1,21 @@
+"""User-owned IM channel connection persistence."""
+
+from deerflow.persistence.channel_connections.model import (
+    ChannelConnectionRow,
+    ChannelConversationRow,
+    ChannelCredentialRow,
+    ChannelOAuthStateRow,
+)
+from deerflow.persistence.channel_connections.sql import (
+    ChannelConnectionRepository,
+    ChannelCredentialCipher,
+)
+
+__all__ = [
+    "ChannelConnectionRepository",
+    "ChannelConnectionRow",
+    "ChannelConversationRow",
+    "ChannelCredentialCipher",
+    "ChannelCredentialRow",
+    "ChannelOAuthStateRow",
+]

backend/packages/harness/deerflow/persistence/channel_connections/model.py

@@ -0,0 +1,111 @@
+"""ORM models for user-owned IM channel connections."""
+
+from __future__ import annotations
+
+from datetime import UTC, datetime
+
+from sqlalchemy import JSON, DateTime, ForeignKey, Index, Integer, String, Text, UniqueConstraint
+from sqlalchemy.orm import Mapped, mapped_column
+
+from deerflow.persistence.base import Base
+
+
+def _utc_now() -> datetime:
+    return datetime.now(UTC)
+
+
+class ChannelConnectionRow(Base):
+    __tablename__ = "channel_connections"
+
+    id: Mapped[str] = mapped_column(String(64), primary_key=True)
+    owner_user_id: Mapped[str] = mapped_column(String(64), nullable=False, index=True)
+    provider: Mapped[str] = mapped_column(String(32), nullable=False, index=True)
+    status: Mapped[str] = mapped_column(String(32), nullable=False, default="connected")
+
+    external_account_id: Mapped[str] = mapped_column(String(128), nullable=False, default="")
+    external_account_name: Mapped[str | None] = mapped_column(String(256), nullable=True)
+    workspace_id: Mapped[str] = mapped_column(String(128), nullable=False, default="")
+    workspace_name: Mapped[str | None] = mapped_column(String(256), nullable=True)
+    bot_user_id: Mapped[str | None] = mapped_column(String(128), nullable=True)
+
+    scopes_json: Mapped[list] = mapped_column(JSON, default=list)
+    capabilities_json: Mapped[dict] = mapped_column(JSON, default=dict)
+    metadata_json: Mapped[dict] = mapped_column(JSON, default=dict)
+
+    created_at: Mapped[datetime] = mapped_column(DateTime(timezone=True), nullable=False, default=_utc_now)
+    updated_at: Mapped[datetime] = mapped_column(DateTime(timezone=True), nullable=False, default=_utc_now, onupdate=_utc_now)
+    last_seen_at: Mapped[datetime | None] = mapped_column(DateTime(timezone=True), nullable=True)
+    last_error_at: Mapped[datetime | None] = mapped_column(DateTime(timezone=True), nullable=True)
+
+    __table_args__ = (
+        UniqueConstraint(
+            "owner_user_id",
+            "provider",
+            "external_account_id",
+            "workspace_id",
+            name="uq_channel_connection_owner_provider_identity",
+        ),
+        Index("idx_channel_connections_event_lookup", "provider", "workspace_id", "bot_user_id"),
+    )
+
+
+class ChannelCredentialRow(Base):
+    __tablename__ = "channel_credentials"
+
+    connection_id: Mapped[str] = mapped_column(
+        String(64),
+        ForeignKey("channel_connections.id", ondelete="CASCADE"),
+        primary_key=True,
+    )
+    encrypted_access_token: Mapped[str | None] = mapped_column(Text, nullable=True)
+    encrypted_refresh_token: Mapped[str | None] = mapped_column(Text, nullable=True)
+    token_type: Mapped[str | None] = mapped_column(String(32), nullable=True)
+    expires_at: Mapped[datetime | None] = mapped_column(DateTime(timezone=True), nullable=True)
+    refresh_expires_at: Mapped[datetime | None] = mapped_column(DateTime(timezone=True), nullable=True)
+    encrypted_extra_json: Mapped[str | None] = mapped_column(Text, nullable=True)
+    version: Mapped[int] = mapped_column(Integer, nullable=False, default=1)
+    updated_at: Mapped[datetime] = mapped_column(DateTime(timezone=True), nullable=False, default=_utc_now, onupdate=_utc_now)
+
+
+class ChannelOAuthStateRow(Base):
+    __tablename__ = "channel_oauth_states"
+
+    state_hash: Mapped[str] = mapped_column(String(128), primary_key=True)
+    owner_user_id: Mapped[str] = mapped_column(String(64), nullable=False, index=True)
+    provider: Mapped[str] = mapped_column(String(32), nullable=False, index=True)
+    code_verifier_encrypted: Mapped[str | None] = mapped_column(Text, nullable=True)
+    nonce_hash: Mapped[str | None] = mapped_column(String(128), nullable=True)
+    redirect_after: Mapped[str | None] = mapped_column(Text, nullable=True)
+    requested_scopes_json: Mapped[list] = mapped_column(JSON, default=list)
+    metadata_json: Mapped[dict] = mapped_column(JSON, default=dict)
+    expires_at: Mapped[datetime] = mapped_column(DateTime(timezone=True), nullable=False)
+    consumed_at: Mapped[datetime | None] = mapped_column(DateTime(timezone=True), nullable=True)
+    created_at: Mapped[datetime] = mapped_column(DateTime(timezone=True), nullable=False, default=_utc_now)
+
+
+class ChannelConversationRow(Base):
+    __tablename__ = "channel_conversations"
+
+    id: Mapped[str] = mapped_column(String(64), primary_key=True)
+    connection_id: Mapped[str] = mapped_column(
+        String(64),
+        ForeignKey("channel_connections.id", ondelete="CASCADE"),
+        nullable=False,
+        index=True,
+    )
+    owner_user_id: Mapped[str] = mapped_column(String(64), nullable=False, index=True)
+    provider: Mapped[str] = mapped_column(String(32), nullable=False, index=True)
+    external_conversation_id: Mapped[str] = mapped_column(String(128), nullable=False)
+    external_topic_id: Mapped[str] = mapped_column(String(128), nullable=False, default="")
+    thread_id: Mapped[str] = mapped_column(String(64), nullable=False, index=True)
+    created_at: Mapped[datetime] = mapped_column(DateTime(timezone=True), nullable=False, default=_utc_now)
+    updated_at: Mapped[datetime] = mapped_column(DateTime(timezone=True), nullable=False, default=_utc_now, onupdate=_utc_now)
+
+    __table_args__ = (
+        UniqueConstraint(
+            "connection_id",
+            "external_conversation_id",
+            "external_topic_id",
+            name="uq_channel_conversation_connection_external",
+        ),
+    )

backend/packages/harness/deerflow/persistence/channel_connections/sql.py

@@ -0,0 +1,346 @@
+"""SQL repository for user-owned IM channel connections."""
+
+from __future__ import annotations
+
+import base64
+import hashlib
+import json
+import uuid
+from datetime import UTC, datetime
+from typing import Any
+
+from cryptography.fernet import Fernet
+from sqlalchemy import select
+from sqlalchemy.ext.asyncio import AsyncSession, async_sessionmaker
+
+from deerflow.persistence.channel_connections.model import (
+    ChannelConnectionRow,
+    ChannelConversationRow,
+    ChannelCredentialRow,
+    ChannelOAuthStateRow,
+)
+from deerflow.utils.time import coerce_iso
+
+
+class ChannelCredentialCipher:
+    """Encrypts provider credentials before they are persisted."""
+
+    def __init__(self, fernet: Fernet) -> None:
+        self._fernet = fernet
+
+    @classmethod
+    def from_key(cls, key: str) -> ChannelCredentialCipher:
+        digest = hashlib.sha256(key.encode("utf-8")).digest()
+        return cls(Fernet(base64.urlsafe_b64encode(digest)))
+
+    def encrypt_text(self, value: str | None) -> str | None:
+        if value is None:
+            return None
+        return "fernet:v1:" + self._fernet.encrypt(value.encode("utf-8")).decode("ascii")
+
+    def decrypt_text(self, value: str | None) -> str | None:
+        if value is None:
+            return None
+        token = value.removeprefix("fernet:v1:")
+        return self._fernet.decrypt(token.encode("ascii")).decode("utf-8")
+
+
+class ChannelConnectionRepository:
+    """Persistence facade for channel connections, credentials, and conversations."""
+
+    def __init__(
+        self,
+        session_factory: async_sessionmaker[AsyncSession],
+        *,
+        cipher: ChannelCredentialCipher | None = None,
+    ) -> None:
+        self.session_factory = session_factory
+        self._cipher = cipher
+
+    async def close(self) -> None:
+        from deerflow.persistence.engine import close_engine
+
+        await close_engine()
+
+    @staticmethod
+    def _new_id() -> str:
+        return uuid.uuid4().hex
+
+    @staticmethod
+    def _normalize_optional_identity(value: str | None) -> str:
+        return value or ""
+
+    @staticmethod
+    def _coerce_datetime(value: datetime | None) -> datetime | None:
+        if value is None or value.tzinfo is not None:
+            return value
+        return value.replace(tzinfo=UTC)
+
+    def _encrypt_optional_secret(self, value: str | None) -> str | None:
+        if value is None:
+            return None
+        if self._cipher is None:
+            raise RuntimeError("channel connection encryption key is required")
+        return self._cipher.encrypt_text(value)
+
+    @staticmethod
+    def _connection_to_dict(row: ChannelConnectionRow) -> dict[str, Any]:
+        data = row.to_dict()
+        data["external_account_id"] = data["external_account_id"] or None
+        data["workspace_id"] = data["workspace_id"] or None
+        data["scopes"] = data.pop("scopes_json") or []
+        data["capabilities"] = data.pop("capabilities_json") or {}
+        data["metadata"] = data.pop("metadata_json") or {}
+        for key in ("created_at", "updated_at", "last_seen_at", "last_error_at"):
+            value = data.get(key)
+            if isinstance(value, datetime):
+                data[key] = coerce_iso(value)
+        return data
+
+    async def upsert_connection(
+        self,
+        *,
+        owner_user_id: str,
+        provider: str,
+        external_account_id: str | None = None,
+        external_account_name: str | None = None,
+        workspace_id: str | None = None,
+        workspace_name: str | None = None,
+        bot_user_id: str | None = None,
+        scopes: list[str] | None = None,
+        capabilities: dict[str, Any] | None = None,
+        metadata: dict[str, Any] | None = None,
+        status: str = "connected",
+    ) -> dict[str, Any]:
+        external_account_id_value = self._normalize_optional_identity(external_account_id)
+        workspace_id_value = self._normalize_optional_identity(workspace_id)
+        async with self.session_factory() as session:
+            stmt = select(ChannelConnectionRow).where(
+                ChannelConnectionRow.owner_user_id == owner_user_id,
+                ChannelConnectionRow.provider == provider,
+                ChannelConnectionRow.external_account_id == external_account_id_value,
+                ChannelConnectionRow.workspace_id == workspace_id_value,
+            )
+            row = (await session.execute(stmt)).scalar_one_or_none()
+            if row is None:
+                row = ChannelConnectionRow(
+                    id=self._new_id(),
+                    owner_user_id=owner_user_id,
+                    provider=provider,
+                    external_account_id=external_account_id_value,
+                    workspace_id=workspace_id_value,
+                )
+                session.add(row)
+
+            row.status = status
+            row.external_account_name = external_account_name
+            row.workspace_name = workspace_name
+            row.bot_user_id = bot_user_id
+            row.scopes_json = list(scopes or [])
+            row.capabilities_json = dict(capabilities or {})
+            row.metadata_json = dict(metadata or {})
+            await session.commit()
+            await session.refresh(row)
+            return self._connection_to_dict(row)
+
+    async def list_connections(self, owner_user_id: str) -> list[dict[str, Any]]:
+        async with self.session_factory() as session:
+            result = await session.execute(select(ChannelConnectionRow).where(ChannelConnectionRow.owner_user_id == owner_user_id).order_by(ChannelConnectionRow.updated_at.desc(), ChannelConnectionRow.id.desc()))
+            return [self._connection_to_dict(row) for row in result.scalars()]
+
+    async def disconnect_connection(self, *, connection_id: str, owner_user_id: str) -> bool:
+        async with self.session_factory() as session:
+            row = await session.get(ChannelConnectionRow, connection_id)
+            if row is None or row.owner_user_id != owner_user_id:
+                return False
+
+            row.status = "revoked"
+            credential = await session.get(ChannelCredentialRow, connection_id)
+            if credential is not None:
+                await session.delete(credential)
+            await session.commit()
+            return True
+
+    async def store_credentials(
+        self,
+        connection_id: str,
+        *,
+        access_token: str | None,
+        refresh_token: str | None = None,
+        token_type: str | None = None,
+        expires_at: datetime | None = None,
+        refresh_expires_at: datetime | None = None,
+        extra: dict[str, Any] | None = None,
+    ) -> None:
+        if self._cipher is None:
+            raise RuntimeError("channel connection encryption key is required")
+        async with self.session_factory() as session:
+            row = await session.get(ChannelCredentialRow, connection_id)
+            if row is None:
+                row = ChannelCredentialRow(connection_id=connection_id)
+                session.add(row)
+            row.encrypted_access_token = self._cipher.encrypt_text(access_token)
+            row.encrypted_refresh_token = self._cipher.encrypt_text(refresh_token)
+            row.token_type = token_type
+            row.expires_at = expires_at
+            row.refresh_expires_at = refresh_expires_at
+            row.encrypted_extra_json = self._cipher.encrypt_text(json.dumps(extra or {}, ensure_ascii=False))
+            row.version = (row.version or 0) + 1
+            await session.commit()
+
+    async def get_credentials(self, connection_id: str) -> dict[str, Any] | None:
+        if self._cipher is None:
+            return None
+        async with self.session_factory() as session:
+            row = await session.get(ChannelCredentialRow, connection_id)
+            if row is None:
+                return None
+            extra_raw = self._cipher.decrypt_text(row.encrypted_extra_json)
+            return {
+                "connection_id": row.connection_id,
+                "access_token": self._cipher.decrypt_text(row.encrypted_access_token),
+                "refresh_token": self._cipher.decrypt_text(row.encrypted_refresh_token),
+                "token_type": row.token_type,
+                "expires_at": self._coerce_datetime(row.expires_at),
+                "refresh_expires_at": self._coerce_datetime(row.refresh_expires_at),
+                "extra": json.loads(extra_raw) if extra_raw else {},
+            }
+
+    @staticmethod
+    def hash_state(state: str) -> str:
+        return hashlib.sha256(state.encode("utf-8")).hexdigest()
+
+    async def create_oauth_state(
+        self,
+        *,
+        owner_user_id: str,
+        provider: str,
+        state: str,
+        expires_at: datetime,
+        code_verifier: str | None = None,
+        nonce_hash: str | None = None,
+        redirect_after: str | None = None,
+        requested_scopes: list[str] | None = None,
+        metadata: dict[str, Any] | None = None,
+    ) -> None:
+        row = ChannelOAuthStateRow(
+            state_hash=self.hash_state(state),
+            owner_user_id=owner_user_id,
+            provider=provider,
+            code_verifier_encrypted=self._encrypt_optional_secret(code_verifier),
+            nonce_hash=nonce_hash,
+            redirect_after=redirect_after,
+            requested_scopes_json=list(requested_scopes or []),
+            metadata_json=dict(metadata or {}),
+            expires_at=expires_at,
+        )
+        async with self.session_factory() as session:
+            session.add(row)
+            await session.commit()
+
+    async def count_oauth_states(self, *, owner_user_id: str, provider: str) -> int:
+        async with self.session_factory() as session:
+            result = await session.execute(
+                select(ChannelOAuthStateRow).where(
+                    ChannelOAuthStateRow.owner_user_id == owner_user_id,
+                    ChannelOAuthStateRow.provider == provider,
+                )
+            )
+            return len(list(result.scalars()))
+
+    async def consume_oauth_state(
+        self,
+        *,
+        provider: str,
+        state: str,
+        now: datetime | None = None,
+    ) -> dict[str, Any] | None:
+        current_time = now or datetime.now(UTC)
+        async with self.session_factory() as session:
+            row = await session.get(ChannelOAuthStateRow, self.hash_state(state))
+            if row is None or row.provider != provider or row.consumed_at is not None:
+                return None
+            expires_at = self._coerce_datetime(row.expires_at)
+            if expires_at is not None and expires_at < current_time:
+                return None
+
+            row.consumed_at = current_time
+            await session.commit()
+            return {
+                "owner_user_id": row.owner_user_id,
+                "provider": row.provider,
+                "requested_scopes": row.requested_scopes_json or [],
+                "metadata": row.metadata_json or {},
+                "redirect_after": row.redirect_after,
+            }
+
+    async def find_connection_by_external_identity(
+        self,
+        *,
+        provider: str,
+        external_account_id: str,
+        workspace_id: str | None = None,
+    ) -> dict[str, Any] | None:
+        async with self.session_factory() as session:
+            result = await session.execute(
+                select(ChannelConnectionRow)
+                .where(
+                    ChannelConnectionRow.provider == provider,
+                    ChannelConnectionRow.external_account_id == self._normalize_optional_identity(external_account_id),
+                    ChannelConnectionRow.workspace_id == self._normalize_optional_identity(workspace_id),
+                    ChannelConnectionRow.status == "connected",
+                )
+                .order_by(ChannelConnectionRow.updated_at.desc(), ChannelConnectionRow.id.desc())
+                .limit(1)
+            )
+            row = result.scalar_one_or_none()
+            return self._connection_to_dict(row) if row is not None else None
+
+    async def set_thread_id(
+        self,
+        *,
+        connection_id: str,
+        owner_user_id: str,
+        provider: str,
+        external_conversation_id: str,
+        thread_id: str,
+        external_topic_id: str | None = None,
+    ) -> None:
+        topic_id = external_topic_id or ""
+        async with self.session_factory() as session:
+            stmt = select(ChannelConversationRow).where(
+                ChannelConversationRow.connection_id == connection_id,
+                ChannelConversationRow.external_conversation_id == external_conversation_id,
+                ChannelConversationRow.external_topic_id == topic_id,
+            )
+            row = (await session.execute(stmt)).scalar_one_or_none()
+            if row is None:
+                row = ChannelConversationRow(
+                    id=self._new_id(),
+                    connection_id=connection_id,
+                    owner_user_id=owner_user_id,
+                    provider=provider,
+                    external_conversation_id=external_conversation_id,
+                    external_topic_id=topic_id,
+                    thread_id=thread_id,
+                )
+                session.add(row)
+            else:
+                row.thread_id = thread_id
+                row.owner_user_id = owner_user_id
+                row.provider = provider
+            await session.commit()
+
+    async def get_thread_id(
+        self,
+        connection_id: str,
+        external_conversation_id: str,
+        external_topic_id: str | None = None,
+    ) -> str | None:
+        async with self.session_factory() as session:
+            stmt = select(ChannelConversationRow.thread_id).where(
+                ChannelConversationRow.connection_id == connection_id,
+                ChannelConversationRow.external_conversation_id == external_conversation_id,
+                ChannelConversationRow.external_topic_id == (external_topic_id or ""),
+            )
+            return (await session.execute(stmt)).scalar_one_or_none()

backend/packages/harness/deerflow/persistence/models/__init__.py

@@ -14,10 +14,26 @@ its storage implementation lives in ``deerflow.runtime.events.store.db`` and
 there is no matching entity directory.
 """
 
+from deerflow.persistence.channel_connections.model import (
+    ChannelConnectionRow,
+    ChannelConversationRow,
+    ChannelCredentialRow,
+    ChannelOAuthStateRow,
+)
 from deerflow.persistence.feedback.model import FeedbackRow
 from deerflow.persistence.models.run_event import RunEventRow
 from deerflow.persistence.run.model import RunRow
 from deerflow.persistence.thread_meta.model import ThreadMetaRow
 from deerflow.persistence.user.model import UserRow
 
-__all__ = ["FeedbackRow", "RunEventRow", "RunRow", "ThreadMetaRow", "UserRow"]
+__all__ = [
+    "ChannelConnectionRow",
+    "ChannelConversationRow",
+    "ChannelCredentialRow",
+    "ChannelOAuthStateRow",
+    "FeedbackRow",
+    "RunEventRow",
+    "RunRow",
+    "ThreadMetaRow",
+    "UserRow",
+]

backend/packages/harness/deerflow/runtime/journal.py

@@ -164,7 +164,18 @@ class RunJournal(BaseCallbackHandler):
                 metadata={"caller": caller, **(metadata or {})},
             )
 
-    def on_chain_end(self, outputs: Any, *, run_id: UUID, **kwargs: Any) -> None:
+    def on_chain_end(
+        self,
+        outputs: Any,
+        *,
+        run_id: UUID,
+        parent_run_id: UUID | None = None,
+        **kwargs: Any,
+    ) -> None:
+        # Nested chain ends fire for internal graph nodes; only the root chain
+        # represents the user-visible run lifecycle.
+        if parent_run_id is not None:
+            return
         self._put(event_type="run.end", category="outputs", content=outputs, metadata={"status": "success"})
         self._flush_sync()
 

backend/packages/harness/deerflow/sandbox/local/local_sandbox_provider.py

@@ -147,7 +147,17 @@ class LocalSandboxProvider(SandboxProvider):
                             mount.container_path,
                         )
                         continue
-                    # Ensure the host path exists before adding mapping
+                    # Ensure the host path exists before adding mapping.
+                    #
+                    # ``host_path`` is resolved against the filesystem of the
+                    # process running this provider — for ``make dev`` that is
+                    # the host machine, but for ``make up`` it is the
+                    # ``deer-flow-gateway`` container, so any host path that
+                    # isn't bind-mounted into the gateway image will be missing
+                    # here. Skipping silently makes this a high-cost-to-debug
+                    # silent failure (sandbox skill / tool reads an empty dir
+                    # instead of the configured mount), so escalate to ERROR
+                    # and include actionable guidance. See #3244.
                     if host_path.exists():
                         mappings.append(
                             PathMapping(
@@ -157,10 +167,16 @@ class LocalSandboxProvider(SandboxProvider):
                             )
                         )
                     else:
-                        logger.warning(
+                        logger.error(
-                            "Mount host_path does not exist, skipping: %s -> %s",
+                            "sandbox.mounts entry %s -> %s ignored: host_path %s does not exist from the "
+                            "perspective of the gateway process. In Docker deployments (make up / docker-compose), "
+                            "this path must also be bind-mounted into the gateway container — add a matching "
+                            "volume entry under services.gateway.volumes in docker/docker-compose.yaml (and use "
+                            "the in-container path here), or run in local mode (make dev) where the gateway sees "
+                            "the host filesystem directly.",
                             mount.host_path,
                             mount.container_path,
+                            mount.host_path,
                         )
         except Exception as e:
             # Log but don't fail if config loading fails

backend/packages/harness/deerflow/skills/slash.py

@@ -0,0 +1,65 @@
+from __future__ import annotations
+
+import re
+from dataclasses import dataclass
+
+from deerflow.skills.types import Skill
+
+RESERVED_SLASH_SKILL_NAMES = frozenset({"bootstrap", "help", "memory", "models", "new", "status"})
+_SLASH_SKILL_RE = re.compile(r"^/([a-z0-9]+(?:-[a-z0-9]+)*)(?:\s+|$)")
+
+
+@dataclass(frozen=True, slots=True)
+class SlashSkillReference:
+    """Parsed slash-skill command with the skill name and remaining task text."""
+
+    name: str
+    remaining_text: str
+
+
+@dataclass(frozen=True, slots=True)
+class ResolvedSlashSkill:
+    """Slash-skill activation resolved against enabled runtime-visible skills."""
+
+    skill: Skill
+    remaining_text: str
+    container_file_path: str
+
+
+def parse_slash_skill_reference(text: str) -> SlashSkillReference | None:
+    """Parse strict `/skill-name task` syntax, ignoring reserved control commands."""
+    match = _SLASH_SKILL_RE.match(text)
+    if not match:
+        return None
+    name = match.group(1)
+    if name in RESERVED_SLASH_SKILL_NAMES:
+        return None
+    return SlashSkillReference(
+        name=name,
+        remaining_text=text[match.end() :].lstrip(),
+    )
+
+
+def resolve_slash_skill(
+    text: str,
+    skills: list[Skill],
+    *,
+    available_skills: set[str] | None = None,
+    container_base_path: str = "/mnt/skills",
+) -> ResolvedSlashSkill | None:
+    """Resolve text into an enabled, whitelisted skill activation if possible."""
+    reference = parse_slash_skill_reference(text)
+    if reference is None:
+        return None
+    if available_skills is not None and reference.name not in available_skills:
+        return None
+
+    skill = next((candidate for candidate in skills if candidate.name == reference.name and candidate.enabled), None)
+    if skill is None:
+        return None
+
+    return ResolvedSlashSkill(
+        skill=skill,
+        remaining_text=reference.remaining_text,
+        container_file_path=skill.get_container_file_path(container_base_path),
+    )

backend/packages/harness/deerflow/utils/messages.py

@@ -0,0 +1,31 @@
+from __future__ import annotations
+
+from collections.abc import Mapping
+from typing import Any
+
+ORIGINAL_USER_CONTENT_KEY = "original_user_content"
+
+
+def message_content_to_text(content: Any) -> str:
+    """Extract text from LangChain message content shapes."""
+    if isinstance(content, str):
+        return content
+    if isinstance(content, list):
+        parts: list[str] = []
+        for item in content:
+            if isinstance(item, str):
+                parts.append(item)
+            elif isinstance(item, dict):
+                text = item.get("text")
+                if isinstance(text, str):
+                    parts.append(text)
+        return "\n".join(part for part in parts if part)
+    return str(content)
+
+
+def get_original_user_content_text(content: Any, additional_kwargs: Mapping[str, Any] | None) -> str:
+    """Return pre-middleware user text when available, otherwise content text."""
+    original_content = (additional_kwargs or {}).get(ORIGINAL_USER_CONTENT_KEY)
+    if isinstance(original_content, str):
+        return original_content
+    return message_content_to_text(content)

backend/packages/harness/pyproject.toml

@@ -36,6 +36,7 @@ dependencies = [
     "sqlalchemy[asyncio]>=2.0,<3.0",
     "aiosqlite>=0.19",
     "alembic>=1.13",
+    "cryptography>=43.0.0",
 ]
 
 [project.optional-dependencies]

backend/scripts/build_fixture_from_jsonl.py

@@ -36,7 +36,8 @@ def main() -> int:
     for index, turn in enumerate(turns):
         data = turn["output"].get("data", {})
         tool_calls = [tc.get("name") for tc in (data.get("tool_calls") or [])]
-        print(f"  turn {index}: hash={turn['input_hash'][:12]} tool_calls={tool_calls} content={str(data.get('content'))[:50]!r}")
+        caller = turn.get("caller", "legacy")
+        print(f"  turn {index}: caller={caller} hash={turn['input_hash'][:12]} tool_calls={tool_calls} content={str(data.get('content'))[:50]!r}")
     return 0
 
 

backend/scripts/record_gateway.py

@@ -28,27 +28,45 @@ sys.path.insert(0, str(_BACKEND / "tests"))
 def _install_capture(out_path: Path) -> None:
     from langchain_core.callbacks import BaseCallbackHandler
     from langchain_core.messages import messages_to_dict
-    from replay_provider import hash_messages
+    from replay_provider import caller_identity, hash_messages, hash_replay_input
 
     import deerflow.models.factory as factory_mod
 
     class Capture(BaseCallbackHandler):
         def __init__(self) -> None:
-            self.inputs: dict[str, list] = {}
+            self.inputs: dict[str, tuple[list, str]] = {}
 
-        def on_chat_model_start(self, serialized, messages, *, run_id=None, **kwargs):  # noqa: ANN001
+        def on_chat_model_start(  # noqa: ANN001
-            self.inputs[str(run_id)] = messages[0] if messages else []
+            self,
+            serialized,
+            messages,
+            *,
+            run_id=None,
+            tags=None,
+            name=None,
+            **kwargs,
+        ):
+            self.inputs[str(run_id)] = (
+                messages[0] if messages else [],
+                caller_identity(name=name, tags=tags),
+            )
 
         def on_llm_end(self, response, *, run_id=None, **kwargs):  # noqa: ANN001
-            inp = self.inputs.pop(str(run_id), None)
+            captured = self.inputs.pop(str(run_id), None)
-            if inp is None:
+            if captured is None:
                 return
+            inp, caller = captured
             for batch in response.generations:
                 for gen in batch:
                     message = getattr(gen, "message", None)
                     if message is None:
                         continue
-                    record = {"input_hash": hash_messages(inp), "output": messages_to_dict([message])[0]}
+                    record = {
+                        "caller": caller,
+                        "conversation_hash": hash_messages(inp),
+                        "input_hash": hash_replay_input(inp, caller=caller),
+                        "output": messages_to_dict([message])[0],
+                    }
                     with open(out_path, "a", encoding="utf-8") as handle:
                         handle.write(json.dumps(record, ensure_ascii=False) + "\n")
                         handle.flush()

backend/sitecustomize.py

@@ -0,0 +1,26 @@
+"""Process-wide Python startup customizations for backend entrypoints.
+
+When ``backend/`` is on ``sys.path``, Python imports this module during
+interpreter startup. Keep changes here suitable for all gateway, script,
+migration, and test entrypoints that run in that environment.
+"""
+
+from __future__ import annotations
+
+import asyncio
+import sys
+
+
+def _configure_windows_event_loop_policy() -> None:
+    if sys.platform != "win32":
+        return
+
+    selector_policy = getattr(asyncio, "WindowsSelectorEventLoopPolicy", None)
+    if selector_policy is None:
+        return
+
+    if not isinstance(asyncio.get_event_loop_policy(), selector_policy):
+        asyncio.set_event_loop_policy(selector_policy())
+
+
+_configure_windows_event_loop_policy()

backend/tests/_replay_fixture.py

@@ -32,7 +32,8 @@ REPLAY_MODEL_BLOCK = """\
   - name: scenario-model
     display_name: Scenario Model
     use: replay_provider:ReplayChatModel
-    model: replay"""
+    model: replay
+    supports_thinking: true"""
 
 
 def real_model_block(model: str) -> str:

backend/tests/blocking_io/test_agents_router.py

@@ -0,0 +1,64 @@
+"""Regression anchors: the custom-agent router must not block the event loop.
+
+``app.gateway.routers.agents.create_agent_endpoint`` and ``delete_agent`` are
+async route handlers that resolve the agent directory (``Paths.base_dir`` calls
+``Path.resolve``), probe it (``Path.exists``), and create/remove it (``mkdir``,
+config/SOUL writes, ``shutil.rmtree``) — all blocking IO. Both offload that work
+via ``asyncio.to_thread``; if any of it regresses back onto the event loop, the
+strict Blockbuster gate raises ``BlockingError`` and these tests fail.
+
+Imports live at module scope so the one-time FastAPI app construction (which
+reads files while building OpenAPI schemas) happens at collection time, not on
+the event loop under test. Test-side path resolution is itself offloaded with
+``asyncio.to_thread`` (matching ``test_uploads_middleware``) so only the
+handlers' own filesystem access is exercised on the loop.
+"""
+
+from __future__ import annotations
+
+import asyncio
+from pathlib import Path
+
+import pytest
+
+from app.gateway.routers.agents import AgentCreateRequest, create_agent_endpoint, delete_agent
+from deerflow.config.agents_api_config import load_agents_api_config_from_dict
+from deerflow.config.paths import get_paths
+from deerflow.runtime.user_context import get_effective_user_id
+
+pytestmark = pytest.mark.asyncio
+
+
+async def test_create_agent_does_not_block_event_loop(tmp_path: Path, monkeypatch) -> None:
+    monkeypatch.setenv("DEER_FLOW_HOME", str(tmp_path))
+    monkeypatch.setattr("deerflow.config.paths._paths", None)
+    load_agents_api_config_from_dict({"enabled": True})
+    try:
+        response = await create_agent_endpoint(AgentCreateRequest(name="loop-make-agent", soul="You are a test agent."))
+        assert response is not None
+
+        user_id = get_effective_user_id()
+        # test-side check (resolution offloaded; not exercised on the loop)
+        agent_dir = await asyncio.to_thread(get_paths().user_agent_dir, user_id, "loop-make-agent")
+        assert await asyncio.to_thread((agent_dir / "config.yaml").exists)
+    finally:
+        load_agents_api_config_from_dict({})
+
+
+async def test_delete_agent_does_not_block_event_loop(tmp_path: Path, monkeypatch) -> None:
+    monkeypatch.setenv("DEER_FLOW_HOME", str(tmp_path))
+    monkeypatch.setattr("deerflow.config.paths._paths", None)
+    load_agents_api_config_from_dict({"enabled": True})
+    try:
+        user_id = get_effective_user_id()
+        user_id = get_effective_user_id()
+        # test-side seeding (resolution offloaded; not exercised on the loop)
+        agent_dir = await asyncio.to_thread(get_paths().user_agent_dir, user_id, "loop-test-agent")
+        await asyncio.to_thread(agent_dir.mkdir, parents=True, exist_ok=True)
+        await asyncio.to_thread((agent_dir / "config.yaml").write_text, "name: loop-test-agent\n", encoding="utf-8")
+
+        await delete_agent("loop-test-agent")
+
+        assert not await asyncio.to_thread(agent_dir.exists)
+    finally:
+        load_agents_api_config_from_dict({})

backend/tests/fixtures/replay/write_read_file.ultra.json

@@ -12,7 +12,9 @@
   },
   "turns": [
     {
-      "input_hash": "9c50eda6ab7e8593dabccbdeadc70a4a7bf778b2c0c3f275f1f96cf2c8ab58db",
+      "caller": "lead_agent",
+      "conversation_hash": "9c50eda6ab7e8593dabccbdeadc70a4a7bf778b2c0c3f275f1f96cf2c8ab58db",
+      "input_hash": "27aeb4c11bff2c3ebc182fe52a06556823c21928620a400c7f26be9733c31f3f",
       "output": {
         "type": "ai",
         "data": {
@@ -56,7 +58,9 @@
       }
     },
     {
-      "input_hash": "3598aeb87e221ca8f554e4d61ce6d5e8801754606fa5c95a89c38bd6cb623045",
+      "caller": "middleware:title",
+      "conversation_hash": "3598aeb87e221ca8f554e4d61ce6d5e8801754606fa5c95a89c38bd6cb623045",
+      "input_hash": "75101f9faa453b1a35deff920b1e3c1a9f0b013a7627fbbaa03436752776b953",
       "output": {
         "type": "ai",
         "data": {
@@ -89,7 +93,9 @@
       }
     },
     {
-      "input_hash": "6af134379b2a9efa01b4f63032f88211d5f38f459f8bed621eb6c65e8e05c1f9",
+      "caller": "lead_agent",
+      "conversation_hash": "6af134379b2a9efa01b4f63032f88211d5f38f459f8bed621eb6c65e8e05c1f9",
+      "input_hash": "f7468603a43d301fcc0167c2f7cd10e53137bfc584f1b3d776614b7a612ed7a6",
       "output": {
         "type": "ai",
         "data": {
@@ -132,7 +138,9 @@
       }
     },
     {
-      "input_hash": "04751c4f7b0107b78b5c97d417063883fd586f5ebcbc4acf79be6cb3c0cdaec1",
+      "caller": "lead_agent",
+      "conversation_hash": "04751c4f7b0107b78b5c97d417063883fd586f5ebcbc4acf79be6cb3c0cdaec1",
+      "input_hash": "218645dabc6926a1dbdf45dd20fba8a41e1e690cef78d7752566db3acf5a36ce",
       "output": {
         "type": "ai",
         "data": {
@@ -165,7 +173,9 @@
       }
     },
     {
-      "input_hash": "8b98ebdbb53e88f000556c4753adede8eaa076ff6fd7b8a1285bfd18aee8144d",
+      "caller": "suggest_agent",
+      "conversation_hash": "8b98ebdbb53e88f000556c4753adede8eaa076ff6fd7b8a1285bfd18aee8144d",
+      "input_hash": "dcd855d389d7179a1e4bc7074fa9ba7ce697570af8947225d6bacb538f14a0cb",
       "output": {
         "type": "ai",
         "data": {

backend/tests/replay_provider.py

@@ -2,14 +2,19 @@
 record/replay e2e (mirrors open-design's ``mocks/`` golden traces).
 
 A fixture is a JSON file capturing the *real* model calls of one scenario,
-keyed by a normalized hash of the **input** each call received::
+keyed by a normalized hash of the **caller + input** each call received::
 
     {
       "scenario": "write_read_file",
       "mode": "ultra",
       "model": "gpt-5.5",
       "turns": [
-        {"input_hash": "<sha256>", "input_preview": "...", "output": <message dict>},
+        {
+          "caller": "lead_agent",
+          "conversation_hash": "<sha256>",
+          "input_hash": "<sha256>",
+          "output": <message dict>,
+        },
         ...
       ]
     }
@@ -21,8 +26,11 @@ A real run makes model calls from several callers — the lead agent's own turns
 and their count/order is not something we want a replay to depend on. Matching by
 a normalized hash of the *input messages* means each call gets back exactly the
 output that was recorded for that input, regardless of order or which middleware
-issued it. That keeps the in-graph, deterministic title call part of the
+issued it. The caller name (``lead_agent``, ``middleware:title``,
-recording; memory/summarization, by contrast, are disabled in the replay config
+``suggest_agent``, ``subagent:*``, ...) is included so two different model
+callers with the same conversation text do not compete for the same replay
+bucket. That keeps the in-graph, deterministic title call part of the recording;
+memory/summarization, by contrast, are disabled in the replay config
 (``_replay_fixture.py``) because their background, debounced timing is not
 reproducible across runs.
 
@@ -67,7 +75,7 @@ from collections import deque
 from collections.abc import Iterator
 from typing import Any
 
-from langchain_core.callbacks import CallbackManagerForLLMRun
+from langchain_core.callbacks import BaseCallbackHandler, CallbackManagerForLLMRun
 from langchain_core.language_models.chat_models import BaseChatModel
 from langchain_core.messages import AIMessage, AIMessageChunk, BaseMessage, messages_from_dict
 from langchain_core.outputs import ChatGeneration, ChatGenerationChunk, ChatResult
@@ -75,6 +83,14 @@ from langchain_core.runnables import Runnable
 from pydantic import PrivateAttr
 
 _FIXTURE_ENV = "DEERFLOW_REPLAY_FIXTURE"
+_DEFAULT_CALLER = "lead_agent"
+_CALLER_TAG_PREFIXES = ("middleware:", "subagent:")
+_CALLER_NAME_ALIASES = {
+    # TitleMiddleware uses this run_name and tags the call as middleware:title.
+    # Some execution paths do not preserve the tag down to the model callback,
+    # so keep the run_name and tag in the same replay namespace.
+    "title_agent": "middleware:title",
+}
 
 # Process-wide record of replay misses. A miss raises inside the model, but the
 # gateway's LLMErrorHandlingMiddleware swallows it into a normal assistant error
@@ -94,6 +110,30 @@ def reset_replay_misses() -> None:
     _replay_misses.clear()
 
 
+def _normalize_caller(caller: str | None) -> str:
+    value = _normalize_text(str(caller or "").strip())
+    if not value:
+        return _DEFAULT_CALLER
+    return _CALLER_NAME_ALIASES.get(value, value)
+
+
+def _caller_from_tags(tags: list[str] | None) -> str | None:
+    for tag in tags or []:
+        if isinstance(tag, str) and (tag == _DEFAULT_CALLER or tag.startswith(_CALLER_TAG_PREFIXES)):
+            return tag
+    return None
+
+
+def caller_identity(*, name: str | None = None, tags: list[str] | None = None) -> str:
+    """Stable model-caller identity shared by record and replay.
+
+    Tags win because graph middleware and subagents already use them as the
+    explicit caller marker. ``run_name`` is exposed to callbacks as ``name`` and
+    covers route-level callers such as ``suggest_agent``.
+    """
+    return _normalize_caller(_caller_from_tags(tags) or name)
+
+
 # Volatile substrings that differ between a recording run and a replay run but
 # carry no semantic weight for matching. Normalized to stable placeholders
 # before hashing so the same logical input hashes identically across processes.
@@ -172,10 +212,30 @@ def _canonical_messages(messages: list[BaseMessage]) -> str:
 
 
 def hash_messages(messages: list[BaseMessage]) -> str:
-    """Stable hash of a model call's input. Shared by recorder and replayer."""
+    """Legacy stable hash of only a model call's conversation input."""
     return hashlib.sha256(_canonical_messages(messages).encode("utf-8")).hexdigest()
 
 
+def hash_replay_input(messages: list[BaseMessage], *, caller: str | None) -> str:
+    """Stable replay key for a caller-specific model input."""
+    return hash_input_key(hash_messages(messages), caller=caller)
+
+
+def hash_input_key(conversation_hash: str, *, caller: str | None) -> str:
+    """Namespace a conversation hash by caller identity.
+
+    Keeping this as ``hash(caller + legacy_conversation_hash)`` lets existing
+    fixtures migrate without a live-model re-record: their old ``input_hash`` is
+    exactly the conversation hash.
+    """
+    payload = json.dumps(
+        {"caller": _normalize_caller(caller), "conversation_hash": conversation_hash},
+        sort_keys=True,
+        ensure_ascii=False,
+    )
+    return hashlib.sha256(payload.encode("utf-8")).hexdigest()
+
+
 def _load_fixture(fixture_path: str) -> dict[str, deque[AIMessage]]:
     with open(fixture_path, encoding="utf-8") as handle:
         payload = json.load(handle)
@@ -199,24 +259,54 @@ class ReplayChatModel(BaseChatModel):
 
     _table: dict[str, deque] = PrivateAttr(default_factory=dict)
     _fixture_path: str = PrivateAttr(default="")
+    _run_callers: dict[str, str] = PrivateAttr(default_factory=dict)
 
     def __init__(self, **kwargs: Any) -> None:
         # Ignore provider noise the factory forwards from config (model, api_key,
         # base_url, ...). Fixture path comes from the ``fixture`` kwarg or env.
         fixture_path = kwargs.pop("fixture", None) or os.environ.get(_FIXTURE_ENV)
-        super().__init__()
+        callbacks = kwargs.pop("callbacks", None)
+        super().__init__(callbacks=callbacks)
         if not fixture_path:
             raise ValueError(f"ReplayChatModel needs a fixture path via the ``fixture`` kwarg or ${_FIXTURE_ENV}")
         self._fixture_path = fixture_path
         self._table = _load_fixture(fixture_path)
+        self.callbacks = [*(self.callbacks or []), _ReplayCallerCapture(self._run_callers)]
 
     @property
     def _llm_type(self) -> str:
         return "deerflow-replay"
 
-    def _match(self, messages: list[BaseMessage]) -> AIMessage:
+    def _caller_from_run_manager(self, run_manager: CallbackManagerForLLMRun | None) -> str:
-        key = hash_messages(messages)
+        if run_manager is None:
+            if len(self._run_callers) == 1:
+                # Some async LangGraph paths fire on_chat_model_start with the
+                # caller metadata but invoke the model implementation without a
+                # run_manager. When there is only one pending start event, it is
+                # the current call; use it so record/replay share the same
+                # caller key.
+                return self._run_callers.pop(next(iter(self._run_callers)))
+            return _DEFAULT_CALLER
+        run_id = str(getattr(run_manager, "run_id", ""))
+        caller = self._run_callers.pop(run_id, None)
+        if caller:
+            return caller
+        return caller_identity(
+            name=getattr(run_manager, "run_name", None) or getattr(run_manager, "name", None),
+            tags=getattr(run_manager, "tags", None),
+        )
+
+    def _match(self, messages: list[BaseMessage], run_manager: CallbackManagerForLLMRun | None = None) -> AIMessage:
+        caller = self._caller_from_run_manager(run_manager)
+        key = hash_replay_input(messages, caller=caller)
         bucket = self._table.get(key)
+        if not bucket:
+            # Backward compatibility for fixtures recorded before caller-aware
+            # keys. New recordings write caller-aware ``input_hash`` values.
+            legacy_key = hash_messages(messages)
+            bucket = self._table.get(legacy_key)
+            if bucket:
+                key = legacy_key
         if not bucket:
             _replay_misses.append(key)
             preview = _canonical_messages(messages)
@@ -224,6 +314,7 @@ class ReplayChatModel(BaseChatModel):
                 f"replay miss: no recorded output for input hash {key} in {self._fixture_path!r}. "
                 "The replayed run diverged from the recording (graph changed, a non-deterministic tool result "
                 "altered a downstream input, or a volatile field slipped past normalization). "
+                f"Caller: {caller!r}. "
                 f"Known hashes: {sorted(self._table)}. "
                 f"Normalized input (first 800 chars): {preview[:800]!r}"
             )
@@ -236,7 +327,7 @@ class ReplayChatModel(BaseChatModel):
         run_manager: CallbackManagerForLLMRun | None = None,
         **kwargs: Any,
     ) -> ChatResult:
-        return ChatResult(generations=[ChatGeneration(message=self._match(messages))])
+        return ChatResult(generations=[ChatGeneration(message=self._match(messages, run_manager))])
 
     def _stream(
         self,
@@ -245,9 +336,16 @@ class ReplayChatModel(BaseChatModel):
         run_manager: CallbackManagerForLLMRun | None = None,
         **kwargs: Any,
     ) -> Iterator[ChatGenerationChunk]:
-        turn = self._match(messages)
+        turn = self._match(messages, run_manager)
         text = turn.content if isinstance(turn.content, str) else ""
-        chunk = ChatGenerationChunk(message=AIMessageChunk(content=turn.content, tool_calls=turn.tool_calls, additional_kwargs=turn.additional_kwargs, id=turn.id))
+        chunk = ChatGenerationChunk(
+            message=AIMessageChunk(
+                content=turn.content,
+                tool_calls=turn.tool_calls,
+                additional_kwargs=turn.additional_kwargs,
+                id=turn.id,
+            )
+        )
         if run_manager is not None and text:
             run_manager.on_llm_new_token(text, chunk=chunk)
         yield chunk
@@ -256,5 +354,31 @@ class ReplayChatModel(BaseChatModel):
         return self
 
 
+class _ReplayCallerCapture(BaseCallbackHandler):
+    def __init__(self, run_callers: dict[str, str]) -> None:
+        self._run_callers = run_callers
+
+    def on_chat_model_start(
+        self,
+        serialized: dict,
+        messages: list[list[BaseMessage]],
+        *,
+        run_id: Any = None,
+        tags: list[str] | None = None,
+        name: str | None = None,
+        **kwargs: Any,
+    ) -> None:
+        if run_id is not None:
+            self._run_callers[str(run_id)] = caller_identity(name=name, tags=tags)
+
+
 # Re-export so the recorder shares the exact hashing logic.
-__all__ = ["ReplayChatModel", "hash_messages", "replay_misses", "reset_replay_misses"]
+__all__ = [
+    "ReplayChatModel",
+    "caller_identity",
+    "hash_input_key",
+    "hash_messages",
+    "hash_replay_input",
+    "replay_misses",
+    "reset_replay_misses",
+]

backend/tests/test_app_config_reload.py

@@ -140,6 +140,57 @@ def test_app_config_defaults_empty_database_to_sqlite(tmp_path, monkeypatch):
     assert config.database.sqlite_dir == ".deer-flow/data"
 
 
+def test_app_config_coerces_commented_out_list_sections(tmp_path, monkeypatch):
+    """Commenting out every entry under a list key makes PyYAML parse it as None.
+
+    Regression for the documented ``cp config.example.yaml config.yaml`` flow
+    (issue #1444): such a config must load with empty lists instead of raising
+    ``Input should be a valid list``.
+    """
+    config_path = tmp_path / "config.yaml"
+    extensions_path = tmp_path / "extensions_config.json"
+    _write_extensions_config(extensions_path)
+    config_path.write_text(
+        yaml.safe_dump(
+            {
+                "sandbox": {"use": "deerflow.sandbox.local:LocalSandboxProvider"},
+                "models": None,
+                "tools": None,
+                "tool_groups": None,
+            }
+        ),
+        encoding="utf-8",
+    )
+    monkeypatch.setenv("DEER_FLOW_EXTENSIONS_CONFIG_PATH", str(extensions_path))
+
+    config = AppConfig.from_file(str(config_path))
+
+    assert config.models == []
+    assert config.tools == []
+    assert config.tool_groups == []
+
+
+def test_app_config_warns_when_no_models_configured(tmp_path, monkeypatch, caplog):
+    config_path = tmp_path / "config.yaml"
+    extensions_path = tmp_path / "extensions_config.json"
+    _write_extensions_config(extensions_path)
+    config_path.write_text(
+        yaml.safe_dump(
+            {
+                "sandbox": {"use": "deerflow.sandbox.local:LocalSandboxProvider"},
+                "models": None,
+            }
+        ),
+        encoding="utf-8",
+    )
+    monkeypatch.setenv("DEER_FLOW_EXTENSIONS_CONFIG_PATH", str(extensions_path))
+
+    with caplog.at_level("WARNING", logger="deerflow.config.app_config"):
+        AppConfig.from_file(str(config_path))
+
+    assert "No models are configured" in caplog.text
+
+
 def test_get_app_config_reloads_when_file_changes(tmp_path, monkeypatch):
     config_path = tmp_path / "config.yaml"
     extensions_path = tmp_path / "extensions_config.json"

backend/tests/test_auth_middleware.py

@@ -4,6 +4,7 @@ import pytest
 from starlette.testclient import TestClient
 
 from app.gateway.auth_middleware import AuthMiddleware, _is_public
+from app.gateway.csrf_middleware import CSRFMiddleware
 
 # ── _is_public unit tests ─────────────────────────────────────────────────
 
@@ -38,6 +39,8 @@ def test_public_paths(path: str):
         "/api/threads/123/uploads",
         "/api/agents",
         "/api/channels",
+        "/api/channels/providers",
+        "/api/channels/slack/connect",
         "/api/runs/stream",
         "/api/threads/123/runs",
         "/api/v1/auth/me",
@@ -88,7 +91,9 @@ def test_unknown_api_path_is_protected():
 
 def _make_app():
     """Create a minimal FastAPI app with AuthMiddleware for testing."""
-    from fastapi import FastAPI
+    from fastapi import FastAPI, Request
+
+    from deerflow.runtime.user_context import get_effective_user_id
 
     app = FastAPI()
     app.add_middleware(AuthMiddleware)
@@ -98,8 +103,16 @@ def _make_app():
         return {"status": "ok"}
 
     @app.get("/api/v1/auth/me")
-    async def auth_me():
+    async def auth_me(request: Request):
-        return {"id": "1", "email": "test@test.com"}
+        from app.gateway.deps import get_current_user_from_request
+
+        user = await get_current_user_from_request(request)
+        return {
+            "id": str(user.id),
+            "email": user.email,
+            "system_role": user.system_role,
+            "needs_setup": user.needs_setup,
+        }
 
     @app.get("/api/v1/auth/setup-status")
     async def setup_status():
@@ -109,6 +122,29 @@ def _make_app():
     async def models_get():
         return {"models": []}
 
+    @app.get("/api/whoami")
+    async def whoami(request: Request):
+        user = request.state.user
+        return {
+            "id": str(user.id),
+            "email": getattr(user, "email", None),
+            "system_role": getattr(user, "system_role", None),
+            "context_user_id": get_effective_user_id(),
+        }
+
+    @app.get("/api/current-user-from-dep")
+    async def current_user_from_dep(request: Request):
+        from app.gateway.deps import get_current_user_from_request
+
+        user = await get_current_user_from_request(request)
+        state_user = request.state.user
+        return {
+            "id": str(user.id),
+            "state_id": str(state_user.id),
+            "auth_source": request.state.auth_source,
+            "context_user_id": get_effective_user_id(),
+        }
+
     @app.put("/api/mcp/config")
     async def mcp_put():
         return {"ok": True}
@@ -132,8 +168,24 @@ def _make_app():
     return app
 
 
+def _make_auth_csrf_app():
+    """Create a minimal app with production middleware ordering."""
+    from fastapi import FastAPI
+
+    app = FastAPI()
+    app.add_middleware(AuthMiddleware)
+    app.add_middleware(CSRFMiddleware)
+
+    @app.post("/api/threads/abc/runs/stream")
+    async def protected_mutation():
+        return {"ok": True}
+
+    return app
+
+
 @pytest.fixture
-def client():
+def client(monkeypatch):
+    monkeypatch.delenv("DEER_FLOW_AUTH_DISABLED", raising=False)
     return TestClient(_make_app())
 
 
@@ -161,6 +213,139 @@ def test_protected_path_no_cookie_returns_401(client):
     assert body["detail"]["code"] == "not_authenticated"
 
 
+def test_auth_disabled_allows_protected_path_without_cookie(monkeypatch):
+    monkeypatch.setenv("DEER_FLOW_AUTH_DISABLED", "1")
+    client = TestClient(_make_app())
+
+    res = client.get("/api/models")
+
+    assert res.status_code == 200
+    assert res.json() == {"models": []}
+
+
+def test_auth_disabled_stamps_e2e_admin_user_without_cookie(monkeypatch):
+    monkeypatch.setenv("DEER_FLOW_AUTH_DISABLED", "1")
+    client = TestClient(_make_app())
+
+    res = client.get("/api/whoami")
+
+    assert res.status_code == 200
+    assert res.json() == {
+        "id": "e2e-user",
+        "email": "e2e@test.local",
+        "system_role": "admin",
+        "context_user_id": "e2e-user",
+    }
+
+
+def test_auth_disabled_auth_me_reuses_middleware_user_without_cookie(monkeypatch):
+    monkeypatch.setenv("DEER_FLOW_AUTH_DISABLED", "1")
+    client = TestClient(_make_app())
+
+    res = client.get("/api/v1/auth/me")
+
+    assert res.status_code == 200
+    assert res.json() == {
+        "id": "e2e-user",
+        "email": "e2e@test.local",
+        "system_role": "admin",
+        "needs_setup": False,
+    }
+
+
+def test_auth_disabled_does_not_clobber_valid_session_cookie(monkeypatch):
+    from types import SimpleNamespace
+
+    async def fake_current_user(request):
+        return SimpleNamespace(
+            id="session-user",
+            email="session@test.local",
+            system_role="user",
+            needs_setup=False,
+        )
+
+    monkeypatch.setenv("DEER_FLOW_AUTH_DISABLED", "1")
+    monkeypatch.setattr("app.gateway.deps.get_current_user_from_request", fake_current_user)
+    client = TestClient(_make_app())
+
+    res = client.get("/api/whoami", cookies={"access_token": "valid-session"})
+
+    assert res.status_code == 200
+    assert res.json() == {
+        "id": "session-user",
+        "email": "session@test.local",
+        "system_role": "user",
+        "context_user_id": "session-user",
+    }
+
+
+def test_auth_disabled_does_not_clobber_internal_auth_identity(monkeypatch):
+    from app.gateway.internal_auth import create_internal_auth_headers
+    from deerflow.runtime.user_context import DEFAULT_USER_ID
+
+    monkeypatch.setenv("DEER_FLOW_AUTH_DISABLED", "1")
+    client = TestClient(_make_app())
+
+    res = client.get(
+        "/api/current-user-from-dep",
+        headers=create_internal_auth_headers(),
+    )
+
+    assert res.status_code == 200
+    assert res.json() == {
+        "id": DEFAULT_USER_ID,
+        "state_id": DEFAULT_USER_ID,
+        "auth_source": "internal",
+        "context_user_id": DEFAULT_USER_ID,
+    }
+
+
+def test_auth_disabled_skips_csrf_for_state_changing_requests(monkeypatch):
+    monkeypatch.setenv("DEER_FLOW_AUTH_DISABLED", "1")
+    client = TestClient(_make_auth_csrf_app())
+
+    res = client.post("/api/threads/abc/runs/stream")
+
+    assert res.status_code == 200
+    assert res.json() == {"ok": True}
+
+
+def test_auth_disabled_is_ignored_in_explicit_production_env(monkeypatch):
+    monkeypatch.setenv("DEER_FLOW_AUTH_DISABLED", "1")
+    monkeypatch.setenv("DEER_FLOW_ENV", "production")
+    client = TestClient(_make_app())
+
+    res = client.get("/api/models")
+
+    assert res.status_code == 401
+
+
+def test_auth_disabled_startup_warning_when_effective(monkeypatch, caplog):
+    from app.gateway.auth_disabled import warn_if_auth_disabled_enabled
+
+    monkeypatch.setenv("DEER_FLOW_AUTH_DISABLED", "1")
+    monkeypatch.delenv("DEER_FLOW_ENV", raising=False)
+    monkeypatch.delenv("ENVIRONMENT", raising=False)
+
+    with caplog.at_level("WARNING", logger="app.gateway.auth_disabled"):
+        warn_if_auth_disabled_enabled()
+
+    assert "authentication is bypassed" in caplog.text
+    assert "e2e-user" in caplog.text
+
+
+def test_auth_disabled_startup_warning_suppressed_in_explicit_production_env(monkeypatch, caplog):
+    from app.gateway.auth_disabled import warn_if_auth_disabled_enabled
+
+    monkeypatch.setenv("DEER_FLOW_AUTH_DISABLED", "1")
+    monkeypatch.setenv("ENVIRONMENT", "production")
+
+    with caplog.at_level("WARNING", logger="app.gateway.auth_disabled"):
+        warn_if_auth_disabled_enabled()
+
+    assert "authentication is bypassed" not in caplog.text
+
+
 def test_protected_path_with_junk_cookie_rejected(client):
     """Junk cookie → 401. Middleware strictly validates the JWT now
     (AUTH_TEST_PLAN test 7.5.8); it no longer silently passes bad

backend/tests/test_channel_connections_config.py

@@ -0,0 +1,40 @@
+"""Tests for user-facing IM channel connection configuration."""
+
+from deerflow.config.channel_connections_config import ChannelConnectionsConfig
+
+
+def test_channel_connections_disabled_by_default():
+    config = ChannelConnectionsConfig()
+
+    assert config.enabled is False
+    assert config.slack.enabled is False
+    assert config.telegram.enabled is False
+    assert config.discord.enabled is False
+
+
+def test_enabled_channel_connections_do_not_require_public_url_or_encryption_key():
+    config = ChannelConnectionsConfig.model_validate(
+        {
+            "enabled": True,
+            "telegram": {
+                "enabled": True,
+                "bot_username": "deerflow_bot",
+            },
+            "slack": {"enabled": True},
+            "discord": {"enabled": True},
+        }
+    )
+
+    assert config.enabled is True
+    assert config.provider_status("telegram") == {"enabled": True, "configured": True}
+    assert config.provider_status("slack") == {"enabled": True, "configured": True}
+    assert config.provider_status("discord") == {"enabled": True, "configured": True}
+
+
+def test_provider_status_reports_disabled_and_unknown_providers():
+    config = ChannelConnectionsConfig.model_validate({"enabled": True})
+
+    assert config.provider_status("slack") == {"enabled": False, "configured": False}
+    assert config.provider_status("telegram") == {"enabled": False, "configured": False}
+    assert config.provider_status("discord") == {"enabled": False, "configured": False}
+    assert config.provider_status("unknown") == {"enabled": False, "configured": False}

backend/tests/test_channel_connections_repository.py

@@ -0,0 +1,202 @@
+"""Tests for per-user IM channel connection persistence."""
+
+from __future__ import annotations
+
+from datetime import UTC, datetime, timedelta
+
+import pytest
+from sqlalchemy import select
+
+from deerflow.persistence.channel_connections import (
+    ChannelConnectionRepository,
+    ChannelConnectionRow,
+    ChannelCredentialCipher,
+    ChannelCredentialRow,
+)
+
+
+@pytest.fixture
+async def repo(tmp_path):
+    from deerflow.persistence.engine import close_engine, get_session_factory, init_engine
+
+    url = f"sqlite+aiosqlite:///{tmp_path / 'channels.db'}"
+    await init_engine("sqlite", url=url, sqlite_dir=str(tmp_path))
+    try:
+        yield ChannelConnectionRepository(
+            get_session_factory(),
+            cipher=ChannelCredentialCipher.from_key("test-encryption-key"),
+        )
+    finally:
+        await close_engine()
+
+
+class TestChannelConnectionRepository:
+    @pytest.mark.anyio
+    async def test_connections_are_listed_per_owner(self, repo):
+        alice = await repo.upsert_connection(
+            owner_user_id="alice",
+            provider="slack",
+            external_account_id="U-alice",
+            external_account_name="Alice",
+            workspace_id="T1",
+            workspace_name="Team One",
+            scopes=["chat:write"],
+        )
+        await repo.upsert_connection(
+            owner_user_id="bob",
+            provider="slack",
+            external_account_id="U-bob",
+            external_account_name="Bob",
+            workspace_id="T1",
+            workspace_name="Team One",
+            scopes=["chat:write"],
+        )
+
+        results = await repo.list_connections("alice")
+
+        assert [item["id"] for item in results] == [alice["id"]]
+        assert results[0]["owner_user_id"] == "alice"
+        assert results[0]["provider"] == "slack"
+        assert results[0]["scopes"] == ["chat:write"]
+        assert "encrypted_access_token" not in results[0]
+
+    @pytest.mark.anyio
+    async def test_upsert_connection_updates_existing_provider_identity(self, repo):
+        first = await repo.upsert_connection(
+            owner_user_id="alice",
+            provider="telegram",
+            external_account_id="42",
+            external_account_name="Alice",
+            workspace_id=None,
+            workspace_name=None,
+            status="pending",
+        )
+        second = await repo.upsert_connection(
+            owner_user_id="alice",
+            provider="telegram",
+            external_account_id="42",
+            external_account_name="Alice Telegram",
+            workspace_id=None,
+            workspace_name=None,
+            status="connected",
+        )
+
+        assert second["id"] == first["id"]
+        assert second["status"] == "connected"
+        assert second["external_account_name"] == "Alice Telegram"
+        assert len(await repo.list_connections("alice")) == 1
+
+    @pytest.mark.anyio
+    async def test_credentials_are_encrypted_at_rest_and_decrypted_by_repository(self, repo):
+        connection = await repo.upsert_connection(
+            owner_user_id="alice",
+            provider="slack",
+            external_account_id="U-alice",
+            workspace_id="T1",
+        )
+        expires_at = datetime.now(UTC) + timedelta(hours=1)
+
+        await repo.store_credentials(
+            connection["id"],
+            access_token="xoxb-secret-access-token",
+            refresh_token="secret-refresh-token",
+            token_type="Bearer",
+            expires_at=expires_at,
+            extra={"bot_user_id": "B123"},
+        )
+
+        async with repo.session_factory() as session:
+            row = (await session.execute(select(ChannelCredentialRow))).scalar_one()
+            assert row.encrypted_access_token is not None
+            assert "xoxb-secret-access-token" not in row.encrypted_access_token
+            assert "secret-refresh-token" not in (row.encrypted_refresh_token or "")
+            assert "B123" not in (row.encrypted_extra_json or "")
+
+        credentials = await repo.get_credentials(connection["id"])
+
+        assert credentials is not None
+        assert credentials["access_token"] == "xoxb-secret-access-token"
+        assert credentials["refresh_token"] == "secret-refresh-token"
+        assert credentials["token_type"] == "Bearer"
+        assert credentials["expires_at"] == expires_at
+        assert credentials["extra"] == {"bot_user_id": "B123"}
+
+    @pytest.mark.anyio
+    async def test_conversations_are_scoped_by_connection(self, repo):
+        alice = await repo.upsert_connection(
+            owner_user_id="alice",
+            provider="slack",
+            external_account_id="U-alice",
+            workspace_id="T1",
+        )
+        bob = await repo.upsert_connection(
+            owner_user_id="bob",
+            provider="slack",
+            external_account_id="U-bob",
+            workspace_id="T1",
+        )
+
+        await repo.set_thread_id(
+            connection_id=alice["id"],
+            owner_user_id="alice",
+            provider="slack",
+            external_conversation_id="C-shared",
+            external_topic_id="1710000000.000100",
+            thread_id="thread-alice",
+        )
+        await repo.set_thread_id(
+            connection_id=bob["id"],
+            owner_user_id="bob",
+            provider="slack",
+            external_conversation_id="C-shared",
+            external_topic_id="1710000000.000100",
+            thread_id="thread-bob",
+        )
+
+        assert await repo.get_thread_id(alice["id"], "C-shared", "1710000000.000100") == "thread-alice"
+        assert await repo.get_thread_id(bob["id"], "C-shared", "1710000000.000100") == "thread-bob"
+
+    @pytest.mark.anyio
+    async def test_disconnect_connection_revokes_owner_connection_and_removes_credentials(self, repo):
+        connection = await repo.upsert_connection(
+            owner_user_id="alice",
+            provider="telegram",
+            external_account_id="42",
+        )
+        await repo.store_credentials(connection["id"], access_token="secret-token")
+
+        disconnected = await repo.disconnect_connection(
+            connection_id=connection["id"],
+            owner_user_id="alice",
+        )
+
+        assert disconnected is True
+        async with repo.session_factory() as session:
+            connection_row = await session.get(ChannelConnectionRow, connection["id"])
+            credential_row = await session.get(ChannelCredentialRow, connection["id"])
+        assert connection_row is not None
+        assert connection_row.status == "revoked"
+        assert credential_row is None
+        assert (
+            await repo.find_connection_by_external_identity(
+                provider="telegram",
+                external_account_id="42",
+            )
+            is None
+        )
+
+    @pytest.mark.anyio
+    async def test_disconnect_connection_is_owner_scoped(self, repo):
+        connection = await repo.upsert_connection(
+            owner_user_id="alice",
+            provider="telegram",
+            external_account_id="42",
+        )
+
+        disconnected = await repo.disconnect_connection(
+            connection_id=connection["id"],
+            owner_user_id="bob",
+        )
+
+        assert disconnected is False
+        assert (await repo.list_connections("alice"))[0]["status"] == "connected"

backend/tests/test_channel_connections_router.py

@@ -0,0 +1,322 @@
+"""Router tests for browser-connectable IM channels."""
+
+from __future__ import annotations
+
+from uuid import UUID
+
+from _router_auth_helpers import make_authed_test_app
+from fastapi.testclient import TestClient
+
+from app.gateway.auth.models import User
+from app.gateway.routers import channel_connections
+from deerflow.config.channel_connections_config import ChannelConnectionsConfig
+
+
+def _user() -> User:
+    return User(
+        id=UUID("11111111-2222-3333-4444-555555555555"),
+        email="alice@example.com",
+        password_hash="x",
+        system_role="user",
+    )
+
+
+async def _make_repo(tmp_path):
+    from deerflow.persistence.channel_connections import ChannelConnectionRepository
+    from deerflow.persistence.engine import get_session_factory, init_engine
+
+    await init_engine("sqlite", url=f"sqlite+aiosqlite:///{tmp_path / 'router.db'}", sqlite_dir=str(tmp_path))
+    return ChannelConnectionRepository(get_session_factory())
+
+
+def _make_app(config: ChannelConnectionsConfig, repo, channels_config: dict | None = None):
+    app = make_authed_test_app(user_factory=_user)
+    app.state.channel_connections_config = config
+    app.state.channel_connection_repo = repo
+    app.state.channels_config = channels_config or {}
+    app.include_router(channel_connections.router)
+    return app
+
+
+def _enabled_connections_config() -> ChannelConnectionsConfig:
+    return ChannelConnectionsConfig.model_validate(
+        {
+            "enabled": True,
+            "telegram": {"enabled": True, "bot_username": "deerflow_bot"},
+            "slack": {"enabled": True},
+            "discord": {"enabled": True},
+        }
+    )
+
+
+def _channels_config() -> dict:
+    return {
+        "telegram": {"enabled": True, "bot_token": "telegram-token"},
+        "slack": {"enabled": True, "bot_token": "xoxb-operator", "app_token": "xapp-operator"},
+        "discord": {"enabled": True, "bot_token": "discord-bot"},
+    }
+
+
+def test_get_providers_uses_existing_channels_config(tmp_path):
+    import anyio
+
+    repo = anyio.run(_make_repo, tmp_path)
+    app = _make_app(_enabled_connections_config(), repo, _channels_config())
+
+    with TestClient(app) as client:
+        response = client.get("/api/channels/providers")
+
+    assert response.status_code == 200
+    body = response.json()
+    assert body["enabled"] is True
+    by_provider = {item["provider"]: item for item in body["providers"]}
+    assert by_provider["telegram"]["configured"] is True
+    assert by_provider["telegram"]["auth_mode"] == "deep_link"
+    assert by_provider["slack"]["configured"] is True
+    assert by_provider["slack"]["auth_mode"] == "binding_code"
+    assert by_provider["discord"]["configured"] is True
+    assert by_provider["discord"]["auth_mode"] == "binding_code"
+
+    anyio.run(repo.close)
+
+
+def test_get_providers_reports_unconfigured_when_runtime_channel_is_missing(tmp_path):
+    import anyio
+
+    repo = anyio.run(_make_repo, tmp_path)
+    app = _make_app(_enabled_connections_config(), repo, {"telegram": {"enabled": True, "bot_token": "telegram-token"}})
+
+    with TestClient(app) as client:
+        response = client.get("/api/channels/providers")
+
+    assert response.status_code == 200
+    by_provider = {item["provider"]: item for item in response.json()["providers"]}
+    assert by_provider["telegram"]["configured"] is True
+    assert by_provider["slack"]["configured"] is False
+    assert by_provider["slack"]["connectable"] is False
+    assert "channels.slack" in by_provider["slack"]["unavailable_reason"]
+    assert by_provider["discord"]["configured"] is False
+    assert "channels.discord" in by_provider["discord"]["unavailable_reason"]
+
+    anyio.run(repo.close)
+
+
+def test_get_providers_uses_newest_connection_status_per_provider(tmp_path):
+    import anyio
+
+    repo = anyio.run(_make_repo, tmp_path)
+
+    async def seed_connections():
+        await repo.upsert_connection(
+            owner_user_id=str(_user().id),
+            provider="slack",
+            external_account_id="U-old",
+            workspace_id="T-old",
+            status="revoked",
+        )
+        await anyio.sleep(0.01)
+        await repo.upsert_connection(
+            owner_user_id=str(_user().id),
+            provider="slack",
+            external_account_id="U-new",
+            workspace_id="T-new",
+            status="connected",
+        )
+
+    anyio.run(seed_connections)
+    app = _make_app(_enabled_connections_config(), repo, _channels_config())
+
+    with TestClient(app) as client:
+        response = client.get("/api/channels/providers")
+
+    assert response.status_code == 200
+    by_provider = {item["provider"]: item for item in response.json()["providers"]}
+    assert by_provider["slack"]["connection_status"] == "connected"
+
+    anyio.run(repo.close)
+
+
+def test_get_connections_returns_current_user_connections_only(tmp_path):
+    import anyio
+
+    repo = anyio.run(_make_repo, tmp_path)
+
+    async def seed_connections():
+        await repo.upsert_connection(
+            owner_user_id=str(_user().id),
+            provider="telegram",
+            external_account_id="42",
+            external_account_name="Alice",
+            status="connected",
+        )
+        await repo.upsert_connection(
+            owner_user_id="other-user",
+            provider="telegram",
+            external_account_id="99",
+            external_account_name="Bob",
+            status="connected",
+        )
+
+    anyio.run(seed_connections)
+    app = _make_app(_enabled_connections_config(), repo, _channels_config())
+
+    with TestClient(app) as client:
+        response = client.get("/api/channels/connections")
+
+    assert response.status_code == 200
+    body = response.json()
+    assert len(body["connections"]) == 1
+    assert body["connections"][0]["provider"] == "telegram"
+    assert body["connections"][0]["external_account_id"] == "42"
+
+    anyio.run(repo.close)
+
+
+def test_connect_telegram_returns_deep_link_and_persists_state(tmp_path):
+    import anyio
+
+    repo = anyio.run(_make_repo, tmp_path)
+    app = _make_app(_enabled_connections_config(), repo, _channels_config())
+
+    with TestClient(app) as client:
+        response = client.post("/api/channels/telegram/connect")
+
+    assert response.status_code == 200
+    body = response.json()
+    assert body["provider"] == "telegram"
+    assert body["mode"] == "deep_link"
+    assert body["url"].startswith("https://t.me/deerflow_bot?start=")
+    assert body["code"]
+    assert "/start" in body["instruction"]
+
+    async def count_states():
+        return await repo.count_oauth_states(owner_user_id=str(_user().id), provider="telegram")
+
+    assert anyio.run(count_states) == 1
+
+    anyio.run(repo.close)
+
+
+def test_connect_slack_returns_binding_command_and_persists_state(tmp_path):
+    import anyio
+
+    repo = anyio.run(_make_repo, tmp_path)
+    app = _make_app(_enabled_connections_config(), repo, _channels_config())
+
+    with TestClient(app) as client:
+        response = client.post("/api/channels/slack/connect")
+
+    assert response.status_code == 200
+    body = response.json()
+    assert body["provider"] == "slack"
+    assert body["mode"] == "binding_code"
+    assert body["url"] is None
+    assert len(body["code"]) >= 22
+    assert body["instruction"] == f"Send /connect {body['code']} to the DeerFlow Slack bot."
+
+    async def count_states():
+        return await repo.count_oauth_states(owner_user_id=str(_user().id), provider="slack")
+
+    assert anyio.run(count_states) == 1
+
+    anyio.run(repo.close)
+
+
+def test_connect_discord_returns_binding_command_and_persists_state(tmp_path):
+    import anyio
+
+    repo = anyio.run(_make_repo, tmp_path)
+    app = _make_app(_enabled_connections_config(), repo, _channels_config())
+
+    with TestClient(app) as client:
+        response = client.post("/api/channels/discord/connect")
+
+    assert response.status_code == 200
+    body = response.json()
+    assert body["provider"] == "discord"
+    assert body["mode"] == "binding_code"
+    assert body["url"] is None
+    assert body["code"]
+    assert body["instruction"] == f"Send /connect {body['code']} to the DeerFlow Discord bot."
+
+    async def count_states():
+        return await repo.count_oauth_states(owner_user_id=str(_user().id), provider="discord")
+
+    assert anyio.run(count_states) == 1
+
+    anyio.run(repo.close)
+
+
+def test_connect_unconfigured_runtime_channel_returns_400(tmp_path):
+    import anyio
+
+    repo = anyio.run(_make_repo, tmp_path)
+    app = _make_app(_enabled_connections_config(), repo, {})
+
+    with TestClient(app) as client:
+        response = client.post("/api/channels/slack/connect")
+
+    assert response.status_code == 400
+    assert "channels.slack" in response.json()["detail"]
+
+    anyio.run(repo.close)
+
+
+def test_disconnect_connection_revokes_current_user_connection(tmp_path):
+    import anyio
+
+    repo = anyio.run(_make_repo, tmp_path)
+
+    async def seed_connection():
+        connection = await repo.upsert_connection(
+            owner_user_id=str(_user().id),
+            provider="telegram",
+            external_account_id="42",
+            status="connected",
+        )
+        return connection["id"]
+
+    connection_id = anyio.run(seed_connection)
+    app = _make_app(_enabled_connections_config(), repo, _channels_config())
+
+    with TestClient(app) as client:
+        response = client.delete(f"/api/channels/connections/{connection_id}")
+
+    assert response.status_code == 204
+
+    async def get_connection_status():
+        return (await repo.list_connections(str(_user().id)))[0]["status"]
+
+    assert anyio.run(get_connection_status) == "revoked"
+
+    anyio.run(repo.close)
+
+
+def test_disconnect_connection_is_current_user_scoped(tmp_path):
+    import anyio
+
+    repo = anyio.run(_make_repo, tmp_path)
+
+    async def seed_connection():
+        connection = await repo.upsert_connection(
+            owner_user_id="other-user",
+            provider="telegram",
+            external_account_id="42",
+            status="connected",
+        )
+        return connection["id"]
+
+    connection_id = anyio.run(seed_connection)
+    app = _make_app(_enabled_connections_config(), repo, _channels_config())
+
+    with TestClient(app) as client:
+        response = client.delete(f"/api/channels/connections/{connection_id}")
+
+    assert response.status_code == 404
+
+    async def get_connection_status():
+        return (await repo.list_connections("other-user"))[0]["status"]
+
+    assert anyio.run(get_connection_status) == "connected"
+
+    anyio.run(repo.close)

backend/tests/test_checkpointer.py

@@ -747,7 +747,7 @@ class TestClientCheckpointerFallback:
             patch("deerflow.client.get_app_config", return_value=config_mock),
             patch("deerflow.client.create_agent", side_effect=fake_create_agent),
             patch("deerflow.client.create_chat_model", return_value=MagicMock()),
-            patch("deerflow.client._build_middlewares", return_value=[]),
+            patch("deerflow.client.build_middlewares", return_value=[]),
             patch("deerflow.client.apply_prompt_template", return_value=""),
             patch("deerflow.client.DeerFlowClient._get_tools", return_value=[]),
         ):
@@ -781,7 +781,7 @@ class TestClientCheckpointerFallback:
             patch("deerflow.client.get_app_config", return_value=config_mock),
             patch("deerflow.client.create_agent", side_effect=fake_create_agent),
             patch("deerflow.client.create_chat_model", return_value=MagicMock()),
-            patch("deerflow.client._build_middlewares", return_value=[]),
+            patch("deerflow.client.build_middlewares", return_value=[]),
             patch("deerflow.client.apply_prompt_template", return_value=""),
             patch("deerflow.client.DeerFlowClient._get_tools", return_value=[]),
         ):

backend/tests/test_client.py

@@ -910,7 +910,7 @@ class TestEnsureAgent:
         with (
             patch("deerflow.client.create_chat_model"),
             patch("deerflow.client.create_agent", return_value=mock_agent),
-            patch("deerflow.client._build_middlewares", return_value=[]) as mock_build_middlewares,
+            patch("deerflow.client.build_middlewares", return_value=[]) as mock_build_middlewares,
             patch("deerflow.client.apply_prompt_template", return_value="prompt") as mock_apply_prompt,
             patch.object(client, "_get_tools", return_value=[]),
             patch("deerflow.runtime.checkpointer.get_checkpointer", return_value=MagicMock()),
@@ -935,7 +935,7 @@ class TestEnsureAgent:
         with (
             patch("deerflow.client.create_chat_model"),
             patch("deerflow.client.create_agent", return_value=mock_agent) as mock_create_agent,
-            patch("deerflow.client._build_middlewares", return_value=[]),
+            patch("deerflow.client.build_middlewares", return_value=[]),
             patch("deerflow.client.apply_prompt_template", return_value="prompt"),
             patch.object(client, "_get_tools", return_value=[]),
             patch("deerflow.runtime.checkpointer.get_checkpointer", return_value=mock_checkpointer),
@@ -960,7 +960,7 @@ class TestEnsureAgent:
         with (
             patch("deerflow.client.create_chat_model"),
             patch("deerflow.client.create_agent", return_value=mock_agent) as mock_create_agent,
-            patch("deerflow.client._build_middlewares", side_effect=fake_build_middlewares),
+            patch("deerflow.client.build_middlewares", side_effect=fake_build_middlewares),
             patch("deerflow.client.apply_prompt_template", return_value="prompt"),
             patch.object(client, "_get_tools", return_value=[]),
             patch("deerflow.runtime.checkpointer.get_checkpointer", return_value=MagicMock()),
@@ -979,7 +979,7 @@ class TestEnsureAgent:
         with (
             patch("deerflow.client.create_chat_model"),
             patch("deerflow.client.create_agent", return_value=mock_agent) as mock_create_agent,
-            patch("deerflow.client._build_middlewares", return_value=[]),
+            patch("deerflow.client.build_middlewares", return_value=[]),
             patch("deerflow.client.apply_prompt_template", return_value="prompt"),
             patch.object(client, "_get_tools", return_value=[]),
             patch("deerflow.runtime.checkpointer.get_checkpointer", return_value=None),
@@ -1957,7 +1957,7 @@ class TestScenarioAgentRecreation:
         with (
             patch("deerflow.client.create_chat_model"),
             patch("deerflow.client.create_agent", side_effect=fake_create_agent),
-            patch("deerflow.client._build_middlewares", return_value=[]),
+            patch("deerflow.client.build_middlewares", return_value=[]),
             patch("deerflow.client.apply_prompt_template", return_value="prompt"),
             patch.object(client, "_get_tools", return_value=[]),
             patch("deerflow.runtime.checkpointer.get_checkpointer", return_value=MagicMock()),
@@ -1985,7 +1985,7 @@ class TestScenarioAgentRecreation:
         with (
             patch("deerflow.client.create_chat_model"),
             patch("deerflow.client.create_agent", side_effect=fake_create_agent),
-            patch("deerflow.client._build_middlewares", return_value=[]),
+            patch("deerflow.client.build_middlewares", return_value=[]),
             patch("deerflow.client.apply_prompt_template", return_value="prompt"),
             patch.object(client, "_get_tools", return_value=[]),
             patch("deerflow.runtime.checkpointer.get_checkpointer", return_value=MagicMock()),
@@ -2010,7 +2010,7 @@ class TestScenarioAgentRecreation:
         with (
             patch("deerflow.client.create_chat_model"),
             patch("deerflow.client.create_agent", side_effect=fake_create_agent),
-            patch("deerflow.client._build_middlewares", return_value=[]),
+            patch("deerflow.client.build_middlewares", return_value=[]),
             patch("deerflow.client.apply_prompt_template", return_value="prompt"),
             patch.object(client, "_get_tools", return_value=[]),
             patch("deerflow.runtime.checkpointer.get_checkpointer", return_value=MagicMock()),

backend/tests/test_client_e2e.py

@@ -144,14 +144,14 @@ def e2e_env(tmp_path, monkeypatch):
     #    non-determinism and cost to E2E tests (title generation is already
     #    disabled via TitleConfig above, but the middleware still participates
     #    in the chain and can interfere with event ordering).
-    from deerflow.agents.lead_agent.agent import _build_middlewares as _original_build_middlewares
+    from deerflow.agents.lead_agent.agent import build_middlewares as _original_build_middlewares
     from deerflow.agents.middlewares.title_middleware import TitleMiddleware
 
     def _sync_safe_build_middlewares(*args, **kwargs):
         mws = _original_build_middlewares(*args, **kwargs)
         return [m for m in mws if not isinstance(m, TitleMiddleware)]
 
-    monkeypatch.setattr("deerflow.client._build_middlewares", _sync_safe_build_middlewares)
+    monkeypatch.setattr("deerflow.client.build_middlewares", _sync_safe_build_middlewares)
 
     return {"tmp_path": tmp_path}
 

backend/tests/test_compose_default_workers.py

@@ -0,0 +1,45 @@
+"""Regression test for the Docker Compose default Gateway worker count.
+
+The Gateway holds run state (RunManager and the stream bridge) in process, so
+the default deployment must run a single Uvicorn worker. Running more than one
+worker without a shared cross-worker stream bridge breaks run cancellation, SSE
+reconnects, request de-duplication, and IM channels (nginx has no sticky
+sessions, so requests scatter across workers that each keep their own run
+state). This test pins the safe default so it cannot silently regress to a
+multi-worker default, while still allowing operators to override it once a
+shared stream bridge exists.
+"""
+
+from __future__ import annotations
+
+import re
+from pathlib import Path
+
+import yaml
+
+REPO_ROOT = Path(__file__).resolve().parents[2]
+COMPOSE_PATH = REPO_ROOT / "docker" / "docker-compose.yaml"
+
+
+def _gateway_command() -> str:
+    """Return the gateway service command as a single string."""
+    compose = yaml.safe_load(COMPOSE_PATH.read_text(encoding="utf-8"))
+    command = compose["services"]["gateway"]["command"]
+    # ``command`` may load as a scalar string or a list depending on YAML style.
+    if isinstance(command, list):
+        command = " ".join(str(part) for part in command)
+    return command
+
+
+def test_gateway_defaults_to_single_worker():
+    """With GATEWAY_WORKERS unset, the worker count must default to 1."""
+    command = _gateway_command()
+    match = re.search(r"GATEWAY_WORKERS:-(\d+)", command)
+    assert match is not None, f"gateway command must set a GATEWAY_WORKERS default; got: {command}"
+    assert match.group(1) == "1", f"default Gateway worker count must be 1, got {match.group(1)}"
+
+
+def test_gateway_worker_count_remains_overridable():
+    """The worker count must stay configurable, not hard-coded to 1."""
+    command = _gateway_command()
+    assert "${GATEWAY_WORKERS:-1}" in command, f"worker count must use ${{GATEWAY_WORKERS:-1}} so operators can override it; got: {command}"

backend/tests/test_csrf_middleware.py

@@ -233,3 +233,15 @@ def test_non_auth_mutation_rejects_mismatched_double_submit_token():
 
     assert response.status_code == 403
     assert response.json()["detail"] == "CSRF token mismatch."
+
+
+def test_channel_posts_require_double_submit_csrf():
+    client = TestClient(_make_app(), base_url="https://deerflow.example")
+
+    response = client.post(
+        "/api/channels/slack/connect",
+        headers={"Origin": "https://deerflow.example"},
+    )
+
+    assert response.status_code == 403
+    assert response.json()["detail"] == "CSRF token missing. Include X-CSRF-Token header."

backend/tests/test_dev_entrypoint.py

@@ -44,7 +44,8 @@ def test_entrypoint_excludes_runtime_state_from_uvicorn_reload():
     content = ENTRYPOINT.read_text(encoding="utf-8")
 
     assert ': "${DEER_FLOW_HOME:=/app/backend/.deer-flow}"' in content
-    assert 'mkdir -p "$DEER_FLOW_HOME" /app/backend/.deer-flow' in content
+    # sandbox must be created too, not just .deer-flow (#3459 / #3454).
+    assert 'mkdir -p "$DEER_FLOW_HOME" /app/backend/.deer-flow /app/backend/sandbox' in content
     assert "--reload-include='*.yaml .env'" not in content
     assert "--reload-include='*.yaml'" in content
     assert "--reload-include='.env'" in content

backend/tests/test_discord_channel.py

@@ -2,9 +2,13 @@
 
 from __future__ import annotations
 
+from types import SimpleNamespace
+
+import pytest
+
 from app.channels.discord import DiscordChannel
 from app.channels.manager import CHANNEL_CAPABILITIES
-from app.channels.message_bus import MessageBus
+from app.channels.message_bus import InboundMessageType, MessageBus
 from app.channels.service import _CHANNEL_REGISTRY
 
 
@@ -21,3 +25,64 @@ def test_discord_channel_init() -> None:
     channel = DiscordChannel(bus=bus, config={"bot_token": "token"})
 
     assert channel.name == "discord"
+
+
+def _make_discord_message(text: str):
+    return SimpleNamespace(
+        id=111,
+        content=text,
+        author=SimpleNamespace(id=123, bot=False, display_name="alice"),
+        guild=SimpleNamespace(id=321),
+        channel=SimpleNamespace(id=456),
+        add_reaction=lambda _emoji: None,
+    )
+
+
+@pytest.mark.asyncio
+async def test_discord_bot_mention_slash_skill_routes_as_chat() -> None:
+    bus = MessageBus()
+    channel = DiscordChannel(bus=bus, config={"bot_token": "token"})
+    captured = []
+    channel._running = True
+    channel._client = SimpleNamespace(user=SimpleNamespace(id=999, mention="<@999>"))
+    channel._discord_module = SimpleNamespace(Thread=type("FakeThread", (), {}))
+    channel._publish = captured.append
+
+    async def noop(*_args, **_kwargs):
+        return None
+
+    channel._start_typing = noop
+    channel._add_reaction = noop
+
+    await channel._on_message(_make_discord_message("<@999> /data-analysis analyze uploads/foo.csv"))
+
+    assert len(captured) == 1
+    inbound = captured[0]
+    assert inbound.text == "/data-analysis analyze uploads/foo.csv"
+    assert inbound.msg_type == InboundMessageType.CHAT
+    assert inbound.topic_id == "456"
+
+
+@pytest.mark.asyncio
+async def test_discord_bot_mention_known_command_routes_as_command() -> None:
+    bus = MessageBus()
+    channel = DiscordChannel(bus=bus, config={"bot_token": "token"})
+    captured = []
+    channel._running = True
+    channel._client = SimpleNamespace(user=SimpleNamespace(id=999, mention="<@999>"))
+    channel._discord_module = SimpleNamespace(Thread=type("FakeThread", (), {}))
+    channel._publish = captured.append
+
+    async def noop(*_args, **_kwargs):
+        return None
+
+    channel._start_typing = noop
+    channel._add_reaction = noop
+
+    await channel._on_message(_make_discord_message("<@999> /help"))
+
+    assert len(captured) == 1
+    inbound = captured[0]
+    assert inbound.text == "/help"
+    assert inbound.msg_type == InboundMessageType.COMMAND
+    assert inbound.topic_id == "456"

backend/tests/test_discord_channel_connections.py

@@ -0,0 +1,88 @@
+"""Discord connection routing tests."""
+
+from __future__ import annotations
+
+from datetime import UTC, datetime, timedelta
+from unittest.mock import AsyncMock, MagicMock
+
+import pytest
+
+from app.channels.discord import DiscordChannel
+from app.channels.message_bus import InboundMessage, MessageBus
+
+
+@pytest.fixture
+async def repo(tmp_path):
+    from deerflow.persistence.channel_connections import ChannelConnectionRepository, ChannelCredentialCipher
+    from deerflow.persistence.engine import close_engine, get_session_factory, init_engine
+
+    await init_engine("sqlite", url=f"sqlite+aiosqlite:///{tmp_path / 'discord.db'}", sqlite_dir=str(tmp_path))
+    try:
+        yield ChannelConnectionRepository(
+            get_session_factory(),
+            cipher=ChannelCredentialCipher.from_key("discord-secret"),
+        )
+    finally:
+        await close_engine()
+
+
+@pytest.mark.anyio
+async def test_discord_inbound_attaches_owner_identity_from_user_level_connection(repo):
+    connection = await repo.upsert_connection(
+        owner_user_id="alice",
+        provider="discord",
+        external_account_id="987",
+        external_account_name="Alice",
+        status="connected",
+    )
+    channel = DiscordChannel(
+        bus=MessageBus(),
+        config={"bot_token": "discord-bot", "connection_repo": repo},
+    )
+    inbound = InboundMessage(
+        channel_name="discord",
+        chat_id="C123",
+        user_id="987",
+        text="hello",
+    )
+
+    attached = await channel._attach_connection_identity(inbound, guild_id="G123")
+
+    assert attached.connection_id == connection["id"]
+    assert attached.owner_user_id == "alice"
+    assert attached.workspace_id is None
+
+
+@pytest.mark.anyio
+async def test_discord_connect_command_binds_gateway_identity(repo):
+    state = "discord-bind-code"
+    await repo.create_oauth_state(
+        owner_user_id="deerflow-user-1",
+        provider="discord",
+        state=state,
+        expires_at=datetime.now(UTC) + timedelta(minutes=5),
+    )
+    channel = DiscordChannel(
+        bus=MessageBus(),
+        config={"bot_token": "discord-bot", "connection_repo": repo},
+    )
+    message = MagicMock()
+    message.author.id = 987
+    message.author.display_name = "Alice"
+    message.guild.id = 123
+    message.guild.name = "Deer Guild"
+    message.channel.id = 456
+    message.channel.send = AsyncMock()
+
+    handled = await channel._bind_connection_from_connect_code(message, state)
+
+    connections = await repo.list_connections("deerflow-user-1")
+    assert handled is True
+    assert len(connections) == 1
+    assert connections[0]["provider"] == "discord"
+    assert connections[0]["external_account_id"] == "987"
+    assert connections[0]["external_account_name"] == "Alice"
+    assert connections[0]["workspace_id"] == "123"
+    assert connections[0]["workspace_name"] == "Deer Guild"
+    assert connections[0]["metadata"]["channel_id"] == "456"
+    message.channel.send.assert_awaited_once()

backend/tests/test_gateway_runtime_cleanup.py

@@ -49,7 +49,9 @@ def test_local_dev_gateway_reload_excludes_runtime_state_with_absolute_dirs():
     assert 'export DEER_FLOW_PROJECT_ROOT="$REPO_ROOT"' in serve_sh
     assert 'BACKEND_RUNTIME_HOME="$REPO_ROOT/backend/.deer-flow"' in serve_sh
     assert 'export DEER_FLOW_HOME="$BACKEND_RUNTIME_HOME"' in serve_sh
-    assert 'mkdir -p "$DEER_FLOW_HOME" "$BACKEND_RUNTIME_HOME"' in serve_sh
+    # Every absolute reload-exclude must be pre-created, including backend/sandbox
+    # (#3459 / #3454) — see test_uvicorn_reload_exclude.py for the mechanism.
+    assert 'mkdir -p "$DEER_FLOW_HOME" "$BACKEND_RUNTIME_HOME" "$REPO_ROOT/backend/sandbox"' in serve_sh
     assert "--reload-exclude='$DEER_FLOW_HOME'" in serve_sh
     assert "--reload-exclude='$BACKEND_RUNTIME_HOME'" in serve_sh
     assert "--reload-exclude='sandbox/'" not in serve_sh

backend/tests/test_jina_client.py

@@ -8,7 +8,12 @@ import pytest
 
 import deerflow.community.jina_ai.jina_client as jina_client_module
 from deerflow.community.jina_ai.jina_client import JinaClient
-from deerflow.community.jina_ai.tools import web_fetch_tool
+from deerflow.community.jina_ai.tools import (
+    _coerce_bool,
+    _coerce_proxy,
+    _coerce_timeout,
+    web_fetch_tool,
+)
 
 
 @pytest.fixture
@@ -117,6 +122,59 @@ async def test_crawl_passes_headers(jina_client, monkeypatch):
     assert captured_headers["X-Timeout"] == "30"
 
 
+@pytest.mark.anyio
+async def test_crawl_passes_proxy_to_httpx_client(jina_client, monkeypatch):
+    """Explicit proxy config should be passed to httpx.AsyncClient."""
+    captured_client_kwargs = {}
+
+    class MockAsyncClient:
+        def __init__(self, **kwargs):
+            captured_client_kwargs.update(kwargs)
+
+        async def __aenter__(self):
+            return self
+
+        async def __aexit__(self, exc_type, exc, tb):
+            return None
+
+        async def post(self, url, **kwargs):
+            return httpx.Response(200, text="ok", request=httpx.Request("POST", url))
+
+    monkeypatch.setattr(httpx, "AsyncClient", MockAsyncClient)
+
+    result = await jina_client.crawl("https://example.com", proxy="http://127.0.0.1:7890")
+
+    assert result == "ok"
+    assert captured_client_kwargs["proxy"] == "http://127.0.0.1:7890"
+    assert captured_client_kwargs["trust_env"] is True
+
+
+@pytest.mark.anyio
+async def test_crawl_can_disable_trust_env(jina_client, monkeypatch):
+    """Callers can disable environment proxy lookup for deterministic networking."""
+    captured_client_kwargs = {}
+
+    class MockAsyncClient:
+        def __init__(self, **kwargs):
+            captured_client_kwargs.update(kwargs)
+
+        async def __aenter__(self):
+            return self
+
+        async def __aexit__(self, exc_type, exc, tb):
+            return None
+
+        async def post(self, url, **kwargs):
+            return httpx.Response(200, text="ok", request=httpx.Request("POST", url))
+
+    monkeypatch.setattr(httpx, "AsyncClient", MockAsyncClient)
+
+    result = await jina_client.crawl("https://example.com", trust_env=False)
+
+    assert result == "ok"
+    assert captured_client_kwargs == {"trust_env": False}
+
+
 @pytest.mark.anyio
 async def test_crawl_includes_api_key_when_set(jina_client, monkeypatch):
     """Test that Authorization header is set when JINA_API_KEY is available."""
@@ -199,6 +257,60 @@ async def test_web_fetch_tool_returns_markdown_on_success(monkeypatch):
     assert not result.startswith("Error:")
 
 
+@pytest.mark.anyio
+async def test_web_fetch_tool_forwards_proxy_and_trust_env(monkeypatch):
+    """web_fetch tool config should be forwarded to JinaClient.crawl."""
+    captured_crawl_kwargs = {}
+
+    async def mock_crawl(self, url, **kwargs):
+        captured_crawl_kwargs.update(kwargs)
+        return "<html><body><p>Hello world</p></body></html>"
+
+    mock_config = MagicMock()
+    mock_tool_config = MagicMock()
+    mock_tool_config.model_extra = {
+        "timeout": "20",
+        "proxy": "http://host.docker.internal:7890",
+        "trust_env": "false",
+    }
+    mock_config.get_tool_config.return_value = mock_tool_config
+    monkeypatch.setattr("deerflow.community.jina_ai.tools.get_app_config", lambda: mock_config)
+    monkeypatch.setattr(JinaClient, "crawl", mock_crawl)
+
+    result = await web_fetch_tool.ainvoke("https://example.com")
+
+    assert "Hello world" in result
+    assert captured_crawl_kwargs == {
+        "return_format": "html",
+        "timeout": 20,
+        "proxy": "http://host.docker.internal:7890",
+        "trust_env": False,
+    }
+
+
+@pytest.mark.anyio
+async def test_web_fetch_tool_ignores_empty_proxy(monkeypatch):
+    """Empty proxy values from unresolved env vars should not be passed to httpx."""
+    captured_crawl_kwargs = {}
+
+    async def mock_crawl(self, url, **kwargs):
+        captured_crawl_kwargs.update(kwargs)
+        return "<html><body><p>Hello world</p></body></html>"
+
+    mock_config = MagicMock()
+    mock_tool_config = MagicMock()
+    mock_tool_config.model_extra = {"proxy": "   ", "trust_env": True}
+    mock_config.get_tool_config.return_value = mock_tool_config
+    monkeypatch.setattr("deerflow.community.jina_ai.tools.get_app_config", lambda: mock_config)
+    monkeypatch.setattr(JinaClient, "crawl", mock_crawl)
+
+    result = await web_fetch_tool.ainvoke("https://example.com")
+
+    assert "Hello world" in result
+    assert captured_crawl_kwargs["proxy"] is None
+    assert captured_crawl_kwargs["trust_env"] is True
+
+
 @pytest.mark.anyio
 async def test_web_fetch_tool_offloads_extraction_to_thread(monkeypatch):
     """Test that readability extraction is offloaded via asyncio.to_thread to avoid blocking the event loop."""
@@ -224,3 +336,60 @@ async def test_web_fetch_tool_offloads_extraction_to_thread(monkeypatch):
     result = await web_fetch_tool.ainvoke("https://example.com")
     assert to_thread_called, "extract_article must be called via asyncio.to_thread to avoid blocking the event loop"
     assert "threaded" in result
+
+
+@pytest.mark.parametrize(
+    ("value", "default", "expected"),
+    [
+        (True, False, True),
+        (False, True, False),
+        ("true", False, True),
+        ("YES", False, True),
+        (" on ", False, True),
+        ("1", False, True),
+        ("false", True, False),
+        ("No", True, False),
+        ("off", True, False),
+        ("0", True, False),
+        ("maybe", True, True),
+        ("maybe", False, False),
+        (None, True, True),
+        (123, False, False),
+    ],
+)
+def test_coerce_bool(value, default, expected):
+    """_coerce_bool normalizes booleans, known strings, and falls back to the default."""
+    assert _coerce_bool(value, default) is expected
+
+
+@pytest.mark.parametrize(
+    ("value", "default", "expected"),
+    [
+        (30, 10, 30),
+        ("45", 10, 45),
+        ("not-a-number", 10, 10),
+        (True, 10, 10),
+        (False, 10, 10),
+        (None, 10, 10),
+        (1.5, 10, 10),
+    ],
+)
+def test_coerce_timeout(value, default, expected):
+    """_coerce_timeout accepts ints and numeric strings, rejecting bools and junk."""
+    assert _coerce_timeout(value, default) == expected
+
+
+@pytest.mark.parametrize(
+    ("value", "expected"),
+    [
+        ("http://127.0.0.1:7890", "http://127.0.0.1:7890"),
+        ("  http://proxy:8080  ", "http://proxy:8080"),
+        ("", None),
+        ("   ", None),
+        (None, None),
+        (123, None),
+    ],
+)
+def test_coerce_proxy(value, expected):
+    """_coerce_proxy trims strings and treats empty/non-string values as None."""
+    assert _coerce_proxy(value) == expected

backend/tests/test_langgraph_auth.py

@@ -21,6 +21,7 @@ from langgraph_sdk import Auth
 from app.gateway.auth.config import AuthConfig, set_auth_config
 from app.gateway.auth.jwt import create_access_token, decode_token
 from app.gateway.auth.models import User
+from app.gateway.auth_disabled import AUTH_DISABLED_USER_ID
 from app.gateway.langgraph_auth import add_owner_filter, authenticate
 
 # ── Helpers ───────────────────────────────────────────────────────────────
@@ -59,6 +60,14 @@ def test_no_cookie_raises_401():
     assert "Not authenticated" in str(exc.value.detail)
 
 
+def test_auth_disabled_skips_csrf_and_authenticates_e2e_user(monkeypatch):
+    monkeypatch.setenv("DEER_FLOW_AUTH_DISABLED", "1")
+
+    identity = asyncio.run(authenticate(_req(method="POST")))
+
+    assert identity == AUTH_DISABLED_USER_ID
+
+
 def test_invalid_jwt_raises_401():
     with pytest.raises(Auth.exceptions.HTTPException) as exc:
         asyncio.run(authenticate(_req({"access_token": "garbage"})))

backend/tests/test_lead_agent_model_resolution.py

@@ -56,7 +56,7 @@ def test_make_lead_agent_attaches_tracing_callbacks_at_graph_root(monkeypatch):
 
     monkeypatch.setattr(lead_agent_module, "get_app_config", lambda: app_config)
     monkeypatch.setattr(tools_module, "get_available_tools", lambda **kwargs: [])
-    monkeypatch.setattr(lead_agent_module, "_build_middlewares", lambda config, model_name, agent_name=None, **kwargs: [])
+    monkeypatch.setattr(lead_agent_module, "build_middlewares", lambda config, model_name, agent_name=None, **kwargs: [])
 
     sentinel_handler = object()
     monkeypatch.setattr(lead_agent_module, "build_tracing_callbacks", lambda: [sentinel_handler])
@@ -94,7 +94,7 @@ def test_internal_make_lead_agent_uses_explicit_app_config(monkeypatch):
 
     monkeypatch.setattr(lead_agent_module, "get_app_config", _raise_get_app_config)
     monkeypatch.setattr(tools_module, "get_available_tools", lambda **kwargs: [])
-    monkeypatch.setattr(lead_agent_module, "_build_middlewares", lambda config, model_name, agent_name=None, **kwargs: [])
+    monkeypatch.setattr(lead_agent_module, "build_middlewares", lambda config, model_name, agent_name=None, **kwargs: [])
 
     captured: dict[str, object] = {}
 
@@ -128,7 +128,7 @@ def test_make_lead_agent_uses_runtime_app_config_from_context_without_global_rea
 
     monkeypatch.setattr(lead_agent_module, "get_app_config", _raise_get_app_config)
     monkeypatch.setattr(tools_module, "get_available_tools", lambda **kwargs: [])
-    monkeypatch.setattr(lead_agent_module, "_build_middlewares", lambda config, model_name, agent_name=None, **kwargs: [])
+    monkeypatch.setattr(lead_agent_module, "build_middlewares", lambda config, model_name, agent_name=None, **kwargs: [])
 
     captured: dict[str, object] = {}
 
@@ -207,7 +207,7 @@ def test_make_lead_agent_disables_thinking_when_model_does_not_support_it(monkey
 
     monkeypatch.setattr(lead_agent_module, "get_app_config", lambda: app_config)
     monkeypatch.setattr(tools_module, "get_available_tools", lambda **kwargs: [])
-    monkeypatch.setattr(lead_agent_module, "_build_middlewares", lambda config, model_name, agent_name=None, **kwargs: [])
+    monkeypatch.setattr(lead_agent_module, "build_middlewares", lambda config, model_name, agent_name=None, **kwargs: [])
 
     captured: dict[str, object] = {}
 
@@ -251,7 +251,7 @@ def test_make_lead_agent_reads_runtime_options_from_context(monkeypatch):
     get_available_tools = MagicMock(return_value=[])
     monkeypatch.setattr(lead_agent_module, "get_app_config", lambda: app_config)
     monkeypatch.setattr(tools_module, "get_available_tools", get_available_tools)
-    monkeypatch.setattr(lead_agent_module, "_build_middlewares", lambda config, model_name, agent_name=None, **kwargs: [])
+    monkeypatch.setattr(lead_agent_module, "build_middlewares", lambda config, model_name, agent_name=None, **kwargs: [])
 
     captured: dict[str, object] = {}
 
@@ -328,7 +328,7 @@ def test_build_middlewares_uses_resolved_model_name_for_vision(monkeypatch):
     monkeypatch.setattr(lead_agent_module, "_create_summarization_middleware", lambda **kwargs: None)
     monkeypatch.setattr(lead_agent_module, "_create_todo_list_middleware", lambda is_plan_mode: None)
 
-    middlewares = lead_agent_module._build_middlewares(
+    middlewares = lead_agent_module.build_middlewares(
         {"configurable": {"model_name": "stale-model", "is_plan_mode": False, "subagent_enabled": False}},
         model_name="vision-model",
         custom_middlewares=[MagicMock()],
@@ -374,7 +374,7 @@ def test_build_middlewares_passes_explicit_app_config_to_shared_factory(monkeypa
         lambda agent_name=None, *, memory_config: captured.setdefault("memory_config", memory_config) or "memory-middleware",
     )
 
-    middlewares = lead_agent_module._build_middlewares(
+    middlewares = lead_agent_module.build_middlewares(
         {"configurable": {"is_plan_mode": False, "subagent_enabled": False}},
         model_name="safe-model",
         app_config=app_config,
@@ -407,7 +407,7 @@ def test_build_middlewares_uses_loop_detection_config(monkeypatch):
     monkeypatch.setattr(lead_agent_module, "_create_summarization_middleware", lambda *, app_config=None: None)
     monkeypatch.setattr(lead_agent_module, "_create_todo_list_middleware", lambda is_plan_mode: None)
 
-    middlewares = lead_agent_module._build_middlewares(
+    middlewares = lead_agent_module.build_middlewares(
         {"configurable": {"is_plan_mode": False, "subagent_enabled": False}},
         model_name="safe-model",
         app_config=app_config,
@@ -433,7 +433,7 @@ def test_build_middlewares_omits_loop_detection_when_disabled(monkeypatch):
     monkeypatch.setattr(lead_agent_module, "_create_summarization_middleware", lambda *, app_config=None: None)
     monkeypatch.setattr(lead_agent_module, "_create_todo_list_middleware", lambda is_plan_mode: None)
 
-    middlewares = lead_agent_module._build_middlewares(
+    middlewares = lead_agent_module.build_middlewares(
         {"configurable": {"is_plan_mode": False, "subagent_enabled": False}},
         model_name="safe-model",
         app_config=app_config,

backend/tests/test_lead_agent_skills.py

@@ -60,6 +60,17 @@ def test_get_skills_prompt_section_returns_all_when_available_skills_is_none(mon
     assert "skill2" in result
 
 
+def test_get_skills_prompt_section_includes_slash_activation_guidance(monkeypatch):
+    skills = [_make_skill("data-analysis")]
+    monkeypatch.setattr("deerflow.agents.lead_agent.prompt._get_enabled_skills", lambda: skills)
+
+    result = get_skills_prompt_section(available_skills={"data-analysis"})
+
+    assert "Explicit Slash Skill Activation" in result
+    assert "The runtime injects the activated skill content" in result
+    assert "do not call `read_file` for that SKILL.md again" in result
+
+
 def test_get_skills_prompt_section_includes_self_evolution_rules(monkeypatch):
     skills = [_make_skill("skill1")]
     monkeypatch.setattr("deerflow.agents.lead_agent.prompt._get_enabled_skills", lambda: skills)
@@ -139,7 +150,7 @@ def test_make_lead_agent_empty_skills_passed_correctly(monkeypatch):
     monkeypatch.setattr(lead_agent_module, "create_chat_model", lambda **kwargs: "model")
     monkeypatch.setattr("deerflow.tools.get_available_tools", lambda **kwargs: [])
     monkeypatch.setattr(lead_agent_module, "_load_enabled_skills_for_tool_policy", lambda available_skills, *, app_config: [])
-    monkeypatch.setattr(lead_agent_module, "_build_middlewares", lambda *args, **kwargs: [])
+    monkeypatch.setattr(lead_agent_module, "build_middlewares", lambda *args, **kwargs: [])
     monkeypatch.setattr(lead_agent_module, "create_agent", lambda **kwargs: kwargs)
 
     class MockModelConfig:
@@ -180,7 +191,7 @@ def test_make_lead_agent_filters_tools_from_available_skills(monkeypatch):
 
     monkeypatch.setattr(lead_agent_module, "_resolve_model_name", lambda x=None, **kwargs: "default-model")
     monkeypatch.setattr(lead_agent_module, "create_chat_model", lambda **kwargs: "model")
-    monkeypatch.setattr(lead_agent_module, "_build_middlewares", lambda *args, **kwargs: [])
+    monkeypatch.setattr(lead_agent_module, "build_middlewares", lambda *args, **kwargs: [])
     monkeypatch.setattr(lead_agent_module, "apply_prompt_template", lambda **kwargs: "mock_prompt")
     monkeypatch.setattr(lead_agent_module, "create_agent", lambda **kwargs: kwargs)
     monkeypatch.setattr(lead_agent_module, "load_agent_config", lambda x: AgentConfig(name="test", skills=["restricted", "legacy"]))
@@ -203,7 +214,7 @@ def test_make_lead_agent_all_legacy_skills_preserve_all_tools(monkeypatch):
 
     monkeypatch.setattr(lead_agent_module, "_resolve_model_name", lambda x=None, **kwargs: "default-model")
     monkeypatch.setattr(lead_agent_module, "create_chat_model", lambda **kwargs: "model")
-    monkeypatch.setattr(lead_agent_module, "_build_middlewares", lambda *args, **kwargs: [])
+    monkeypatch.setattr(lead_agent_module, "build_middlewares", lambda *args, **kwargs: [])
     monkeypatch.setattr(lead_agent_module, "apply_prompt_template", lambda **kwargs: "mock_prompt")
     monkeypatch.setattr(lead_agent_module, "create_agent", lambda **kwargs: kwargs)
     monkeypatch.setattr(lead_agent_module, "load_agent_config", lambda x: AgentConfig(name="test", skills=None))
@@ -227,7 +238,7 @@ def test_make_lead_agent_enforces_allowed_tools_when_skill_cache_is_cold(monkeyp
 
     monkeypatch.setattr(lead_agent_module, "_resolve_model_name", lambda x=None, **kwargs: "default-model")
     monkeypatch.setattr(lead_agent_module, "create_chat_model", lambda **kwargs: "model")
-    monkeypatch.setattr(lead_agent_module, "_build_middlewares", lambda *args, **kwargs: [])
+    monkeypatch.setattr(lead_agent_module, "build_middlewares", lambda *args, **kwargs: [])
     monkeypatch.setattr(lead_agent_module, "apply_prompt_template", lambda **kwargs: "mock_prompt")
     monkeypatch.setattr(lead_agent_module, "create_agent", lambda **kwargs: kwargs)
     monkeypatch.setattr(lead_agent_module, "load_agent_config", lambda x: AgentConfig(name="test", skills=["restricted"]))

backend/tests/test_local_sandbox_provider_mounts.py

@@ -612,6 +612,54 @@ class TestLocalSandboxProviderMounts:
 
         assert [m.container_path for m in provider._path_mappings] == ["/mnt/skills"]
 
+    def test_setup_path_mappings_logs_actionable_error_for_missing_host_path(self, tmp_path, caplog):
+        """Regression for #3244.
+
+        When ``sandbox.mounts[].host_path`` is absent from the gateway process's
+        filesystem (the typical symptom in Docker production mode: host_path is a
+        host machine path that is not bind-mounted into the gateway container),
+        the mount is still skipped — but the failure must be a hard-to-miss ERROR
+        log with explicit, actionable guidance about Docker bind mounts, not the
+        old DEBUG/WARNING that buried the silent failure.
+        """
+        skills_dir = tmp_path / "skills"
+        skills_dir.mkdir()
+        missing_host_path = tmp_path / "does-not-exist"
+
+        from deerflow.config.sandbox_config import SandboxConfig, VolumeMountConfig
+
+        sandbox_config = SandboxConfig(
+            use="deerflow.sandbox.local:LocalSandboxProvider",
+            mounts=[
+                VolumeMountConfig(host_path=str(missing_host_path), container_path="/mnt/knowledge", read_only=True),
+            ],
+        )
+        config = SimpleNamespace(
+            skills=SimpleNamespace(container_path="/mnt/skills", get_skills_path=lambda: skills_dir, use="deerflow.skills.storage.local_skill_storage:LocalSkillStorage"),
+            sandbox=sandbox_config,
+        )
+
+        with caplog.at_level("ERROR", logger="deerflow.sandbox.local.local_sandbox_provider"):
+            with patch("deerflow.config.get_app_config", return_value=config):
+                provider = LocalSandboxProvider()
+
+        # Silent-skip behaviour is preserved (no breaking change for existing deployments).
+        assert [m.container_path for m in provider._path_mappings] == ["/mnt/skills"]
+
+        # The failure must be observable at ERROR level and reference the offending paths.
+        error_records = [r for r in caplog.records if r.levelname == "ERROR"]
+        assert error_records, "expected an ERROR log when host_path is missing"
+        message = "\n".join(r.getMessage() for r in error_records)
+        assert str(missing_host_path) in message
+        assert "/mnt/knowledge" in message
+
+        # And it must include actionable Docker guidance so users don't lose hours
+        # to a silent empty-mount failure in production.
+        lowered = message.lower()
+        assert "docker" in lowered
+        assert "gateway" in lowered
+        assert "docker-compose" in lowered
+
     def test_write_file_resolves_container_paths_in_content(self, tmp_path):
         """write_file should replace container paths in file content with local paths."""
         data_dir = tmp_path / "data"

backend/tests/test_patched_stepfun.py

@@ -0,0 +1,305 @@
+"""Tests for deerflow.models.patched_stepfun.PatchedChatStepFun."""
+
+from __future__ import annotations
+
+from unittest.mock import MagicMock, patch
+
+from langchain_core.messages import AIMessage, AIMessageChunk, HumanMessage
+
+
+def _make_model(**kwargs):
+    from deerflow.models.patched_stepfun import PatchedChatStepFun
+
+    return PatchedChatStepFun(
+        model="step-3.7-flash",
+        api_key="test-key",
+        base_url="https://api.stepfun.com/v1",
+        **kwargs,
+    )
+
+
+# ---------------------------------------------------------------------------
+# Basic properties
+# ---------------------------------------------------------------------------
+
+
+def test_is_lc_serializable_returns_true():
+    from deerflow.models.patched_stepfun import PatchedChatStepFun
+
+    assert PatchedChatStepFun.is_lc_serializable() is True
+
+
+def test_lc_secrets_contains_stepfun_api_key_mapping():
+    model = _make_model()
+    assert model.lc_secrets["api_key"] == "STEPFUN_API_KEY"
+    assert model.lc_secrets["openai_api_key"] == "STEPFUN_API_KEY"
+
+
+# ---------------------------------------------------------------------------
+# _extract_reasoning helper
+# ---------------------------------------------------------------------------
+
+
+def test_extract_reasoning_from_dict_with_reasoning():
+    from deerflow.models.patched_stepfun import _extract_reasoning
+
+    assert _extract_reasoning({"reasoning": "thinking..."}) == "thinking..."
+
+
+def test_extract_reasoning_from_dict_with_reasoning_content():
+    from deerflow.models.patched_stepfun import _extract_reasoning
+
+    assert _extract_reasoning({"reasoning_content": "thinking..."}) == "thinking..."
+
+
+def test_extract_reasoning_prefers_reasoning_content_over_reasoning():
+    from deerflow.models.patched_stepfun import _extract_reasoning
+
+    result = _extract_reasoning({"reasoning_content": "deepseek", "reasoning": "native"})
+    assert result == "deepseek"
+
+
+def test_extract_reasoning_missing_returns_sentinel():
+    from deerflow.models.patched_stepfun import _MISSING, _extract_reasoning
+
+    assert _extract_reasoning({}) is _MISSING
+    assert _extract_reasoning({"reasoning": None}) is _MISSING
+
+
+# ---------------------------------------------------------------------------
+# Request payload replay (_get_request_payload)
+# ---------------------------------------------------------------------------
+
+
+def test_reasoning_content_injected_into_assistant_tool_call_message():
+    model = _make_model()
+
+    human = HumanMessage(content="Check Beijing weather.")
+    ai = AIMessage(
+        content="",
+        additional_kwargs={"reasoning_content": "I need to call the weather tool."},
+    )
+    payload_message = {
+        "role": "assistant",
+        "content": "",
+        "tool_calls": [
+            {
+                "id": "call_weather",
+                "type": "function",
+                "function": {"name": "get_weather", "arguments": '{"location":"Beijing"}'},
+            }
+        ],
+    }
+    base_payload = {
+        "messages": [
+            {"role": "user", "content": "Check Beijing weather."},
+            payload_message,
+        ]
+    }
+
+    with patch.object(type(model).__bases__[0], "_get_request_payload", return_value=base_payload):
+        with patch.object(model, "_convert_input") as mock_convert:
+            mock_convert.return_value = MagicMock(to_messages=lambda: [human, ai])
+            payload = model._get_request_payload([human, ai])
+
+    assert payload["messages"][1]["reasoning_content"] == "I need to call the weather tool."
+
+
+def test_reasoning_content_is_noop_when_missing():
+    model = _make_model()
+
+    human = HumanMessage(content="hello")
+    ai = AIMessage(content="hi", additional_kwargs={})
+    base_payload = {
+        "messages": [
+            {"role": "user", "content": "hello"},
+            {"role": "assistant", "content": "hi"},
+        ]
+    }
+
+    with patch.object(type(model).__bases__[0], "_get_request_payload", return_value=base_payload):
+        with patch.object(model, "_convert_input") as mock_convert:
+            mock_convert.return_value = MagicMock(to_messages=lambda: [human, ai])
+            payload = model._get_request_payload([human, ai])
+
+    assert "reasoning_content" not in payload["messages"][1]
+
+
+# ---------------------------------------------------------------------------
+# Streaming reasoning capture (_convert_chunk_to_generation_chunk)
+# ---------------------------------------------------------------------------
+
+
+def test_convert_chunk_captures_reasoning_field():
+    """StepFun default format: delta.reasoning."""
+    model = _make_model()
+
+    chunk = model._convert_chunk_to_generation_chunk(
+        {"choices": [{"delta": {"role": "assistant", "reasoning": "I need "}}]},
+        AIMessageChunk,
+        {},
+    )
+
+    assert chunk is not None
+    assert chunk.message.additional_kwargs["reasoning_content"] == "I need "
+
+
+def test_convert_chunk_captures_reasoning_content_field():
+    """StepFun deepseek-style format: delta.reasoning_content."""
+    model = _make_model()
+
+    chunk = model._convert_chunk_to_generation_chunk(
+        {"choices": [{"delta": {"role": "assistant", "reasoning_content": "I need "}}]},
+        AIMessageChunk,
+        {},
+    )
+
+    assert chunk is not None
+    assert chunk.message.additional_kwargs["reasoning_content"] == "I need "
+
+
+def test_convert_chunk_streams_reasoning_then_content():
+    """Full streaming flow: reasoning deltas followed by content."""
+    model = _make_model()
+
+    first = model._convert_chunk_to_generation_chunk(
+        {"choices": [{"delta": {"role": "assistant", "reasoning": "I need "}}]},
+        AIMessageChunk,
+        {},
+    )
+    second = model._convert_chunk_to_generation_chunk(
+        {"choices": [{"delta": {"reasoning": "a tool."}}]},
+        AIMessageChunk,
+        {},
+    )
+    answer = model._convert_chunk_to_generation_chunk(
+        {"choices": [{"delta": {"content": "Done."}, "finish_reason": "stop"}], "model": "step-3.7-flash"},
+        AIMessageChunk,
+        {},
+    )
+
+    assert first is not None
+    assert second is not None
+    assert answer is not None
+
+    combined = first.message + second.message + answer.message
+    assert combined.additional_kwargs["reasoning_content"] == "I need a tool."
+    assert combined.content == "Done."
+
+
+def test_convert_chunk_noop_when_no_reasoning():
+    model = _make_model()
+
+    chunk = model._convert_chunk_to_generation_chunk(
+        {"choices": [{"delta": {"content": "Hello."}, "finish_reason": "stop"}], "model": "step-3.7-flash"},
+        AIMessageChunk,
+        {},
+    )
+
+    assert chunk is not None
+    assert "reasoning_content" not in chunk.message.additional_kwargs
+
+
+# ---------------------------------------------------------------------------
+# Non-streaming reasoning capture (_create_chat_result)
+# ---------------------------------------------------------------------------
+
+
+def test_create_chat_result_extracts_reasoning_field():
+    """StepFun default format: message.reasoning."""
+    model = _make_model()
+    response = {
+        "choices": [
+            {
+                "message": {
+                    "role": "assistant",
+                    "content": "The weather is sunny.",
+                    "reasoning": "The tool returned sunny weather.",
+                },
+                "finish_reason": "stop",
+            }
+        ],
+        "model": "step-3.7-flash",
+    }
+
+    result = model._create_chat_result(response)
+    message = result.generations[0].message
+
+    assert message.content == "The weather is sunny."
+    assert message.additional_kwargs["reasoning_content"] == "The tool returned sunny weather."
+
+
+def test_create_chat_result_extracts_reasoning_content_field():
+    """StepFun deepseek-style format: message.reasoning_content."""
+    model = _make_model()
+    response = {
+        "choices": [
+            {
+                "message": {
+                    "role": "assistant",
+                    "content": "The weather is sunny.",
+                    "reasoning_content": "The tool returned sunny weather.",
+                },
+                "finish_reason": "stop",
+            }
+        ],
+        "model": "step-3.7-flash",
+    }
+
+    result = model._create_chat_result(response)
+    message = result.generations[0].message
+
+    assert message.content == "The weather is sunny."
+    assert message.additional_kwargs["reasoning_content"] == "The tool returned sunny weather."
+
+
+def test_create_chat_result_reads_reasoning_from_sdk_object():
+    """When the response is a Pydantic model, reasoning is an attribute."""
+    model = _make_model()
+
+    class FakeMessage:
+        reasoning = "Reasoning stored on the SDK message object."
+        reasoning_content = None
+        model_extra = None
+
+    class FakeChoice:
+        message = FakeMessage()
+
+    class FakeResponse:
+        choices = [FakeChoice()]
+
+        def model_dump(self, **kwargs):
+            return {
+                "choices": [
+                    {
+                        "message": {
+                            "role": "assistant",
+                            "content": "Answer.",
+                        },
+                        "finish_reason": "stop",
+                    }
+                ],
+                "model": "step-3.7-flash",
+            }
+
+    result = model._create_chat_result(FakeResponse())
+    assert result.generations[0].message.additional_kwargs["reasoning_content"] == "Reasoning stored on the SDK message object."
+
+
+def test_create_chat_result_noop_when_no_reasoning():
+    model = _make_model()
+    response = {
+        "choices": [
+            {
+                "message": {
+                    "role": "assistant",
+                    "content": "Hello!",
+                },
+                "finish_reason": "stop",
+            }
+        ],
+        "model": "step-3.7-flash",
+    }
+
+    result = model._create_chat_result(response)
+    assert "reasoning_content" not in result.generations[0].message.additional_kwargs

backend/tests/test_replay_provider.py

@@ -0,0 +1,116 @@
+from __future__ import annotations
+
+import json
+from pathlib import Path
+
+from langchain_core.messages import AIMessage, HumanMessage, messages_to_dict
+from replay_provider import ReplayChatModel, caller_identity, hash_messages, hash_replay_input
+
+
+def _write_fixture(path: Path, turns: list[dict]) -> None:
+    path.write_text(
+        json.dumps(
+            {
+                "scenario": "unit",
+                "mode": "unit",
+                "model": "replay",
+                "prompt": "unit",
+                "context": {},
+                "turns": turns,
+            }
+        ),
+        encoding="utf-8",
+    )
+
+
+def test_replay_key_includes_caller_identity(tmp_path: Path):
+    messages = [HumanMessage(content="same conversation")]
+    lead_output = AIMessage(content="lead")
+    suggest_output = AIMessage(content="suggest")
+    fixture_path = tmp_path / "fixture.json"
+
+    _write_fixture(
+        fixture_path,
+        [
+            {
+                "caller": "lead_agent",
+                "conversation_hash": hash_messages(messages),
+                "input_hash": hash_replay_input(messages, caller="lead_agent"),
+                "output": messages_to_dict([lead_output])[0],
+            },
+            {
+                "caller": "suggest_agent",
+                "conversation_hash": hash_messages(messages),
+                "input_hash": hash_replay_input(messages, caller="suggest_agent"),
+                "output": messages_to_dict([suggest_output])[0],
+            },
+        ],
+    )
+
+    model = ReplayChatModel(fixture=str(fixture_path))
+
+    assert model.invoke(messages, config={"run_name": "suggest_agent"}).content == "suggest"
+    assert model.invoke(messages, config={"run_name": "lead_agent"}).content == "lead"
+
+
+def test_replay_supports_legacy_conversation_only_fixture(tmp_path: Path):
+    messages = [HumanMessage(content="legacy conversation")]
+    fixture_path = tmp_path / "legacy.json"
+
+    _write_fixture(
+        fixture_path,
+        [
+            {
+                "input_hash": hash_messages(messages),
+                "output": messages_to_dict([AIMessage(content="legacy")])[0],
+            }
+        ],
+    )
+
+    model = ReplayChatModel(fixture=str(fixture_path))
+
+    assert model.invoke(messages, config={"run_name": "suggest_agent"}).content == "legacy"
+
+
+def test_title_run_name_uses_middleware_caller_namespace(tmp_path: Path):
+    messages = [HumanMessage(content="title prompt")]
+    fixture_path = tmp_path / "fixture.json"
+
+    _write_fixture(
+        fixture_path,
+        [
+            {
+                "caller": "middleware:title",
+                "conversation_hash": hash_messages(messages),
+                "input_hash": hash_replay_input(messages, caller="middleware:title"),
+                "output": messages_to_dict([AIMessage(content="generated title")])[0],
+            }
+        ],
+    )
+
+    model = ReplayChatModel(fixture=str(fixture_path))
+
+    assert caller_identity(name="title_agent") == "middleware:title"
+    assert model.invoke(messages, config={"run_name": "title_agent"}).content == "generated title"
+
+
+def test_replay_uses_single_pending_capture_when_run_manager_is_missing(tmp_path: Path):
+    messages = [HumanMessage(content="title prompt")]
+    fixture_path = tmp_path / "fixture.json"
+
+    _write_fixture(
+        fixture_path,
+        [
+            {
+                "caller": "middleware:title",
+                "conversation_hash": hash_messages(messages),
+                "input_hash": hash_replay_input(messages, caller="middleware:title"),
+                "output": messages_to_dict([AIMessage(content="generated title")])[0],
+            }
+        ],
+    )
+
+    model = ReplayChatModel(fixture=str(fixture_path))
+    model._run_callers["captured-run"] = caller_identity(name="title_agent", tags=["middleware:title"])
+
+    assert model._match(messages, run_manager=None).content == "generated title"

backend/tests/test_run_journal.py

@@ -179,15 +179,16 @@ class TestLifecycleCallbacks:
         assert "run.end" in types
 
     @pytest.mark.anyio
-    async def test_nested_chain_no_run_start(self, journal_setup):
+    async def test_nested_chain_no_run_lifecycle_events(self, journal_setup):
-        """Nested chains (parent_run_id set) should NOT produce run.start."""
+        """Nested chains (parent_run_id set) should NOT produce root run lifecycle events."""
         j, store = journal_setup
         parent_id = uuid4()
         j.on_chain_start({}, {}, run_id=uuid4(), parent_run_id=parent_id)
-        j.on_chain_end({}, run_id=uuid4())
+        j.on_chain_end({}, run_id=uuid4(), parent_run_id=parent_id)
         await j.flush()
         events = await store.list_events("t1", "r1")
         assert not any(e["event_type"] == "run.start" for e in events)
+        assert not any(e["event_type"] == "run.end" for e in events)
 
 
 class TestToolCallbacks:

backend/tests/test_slack_channel_connections.py

@@ -0,0 +1,154 @@
+"""Slack connection tests for user-owned channel bindings."""
+
+from __future__ import annotations
+
+import sys
+from datetime import UTC, datetime, timedelta
+from types import ModuleType
+from unittest.mock import AsyncMock, MagicMock
+
+from app.channels.message_bus import MessageBus, OutboundMessage
+
+
+async def _make_repo(tmp_path):
+    from deerflow.persistence.channel_connections import ChannelConnectionRepository, ChannelCredentialCipher
+    from deerflow.persistence.engine import get_session_factory, init_engine
+
+    await init_engine("sqlite", url=f"sqlite+aiosqlite:///{tmp_path / 'slack.db'}", sqlite_dir=str(tmp_path))
+    return ChannelConnectionRepository(
+        get_session_factory(),
+        cipher=ChannelCredentialCipher.from_key("slack-secret"),
+    )
+
+
+def test_slack_connect_command_binds_socket_mode_identity(tmp_path):
+    import anyio
+
+    from app.channels.slack import SlackChannel
+
+    async def go():
+        repo = await _make_repo(tmp_path)
+        state = "slack-bind-code"
+        await repo.create_oauth_state(
+            owner_user_id="deerflow-user-1",
+            provider="slack",
+            state=state,
+            expires_at=datetime.now(UTC) + timedelta(minutes=5),
+        )
+        channel = SlackChannel(
+            bus=MessageBus(),
+            config={"bot_token": "xoxb-operator", "app_token": "xapp-operator", "connection_repo": repo},
+        )
+        channel._web_client = MagicMock()
+
+        handled = await channel._bind_connection_from_connect_code(
+            event={
+                "user": "U123",
+                "channel": "C123",
+                "ts": "1710000000.000100",
+            },
+            team_id="T123",
+            code=state,
+        )
+
+        connections = await repo.list_connections("deerflow-user-1")
+        assert handled is True
+        assert len(connections) == 1
+        assert connections[0]["provider"] == "slack"
+        assert connections[0]["external_account_id"] == "U123"
+        assert connections[0]["workspace_id"] == "T123"
+        assert connections[0]["metadata"]["channel_id"] == "C123"
+        channel._web_client.chat_postMessage.assert_called_once()
+        await repo.close()
+
+    anyio.run(go)
+
+
+def test_slack_send_uses_connection_bot_token_when_connection_id_is_present():
+    import anyio
+
+    from app.channels.slack import SlackChannel
+
+    async def go():
+        repo = AsyncMock()
+        repo.get_credentials.return_value = {"access_token": "xoxb-connection-token"}
+        web_client = MagicMock()
+        web_client_factory = MagicMock(return_value=web_client)
+        channel = SlackChannel(
+            bus=MessageBus(),
+            config={
+                "connection_repo": repo,
+                "web_client_factory": web_client_factory,
+            },
+        )
+
+        msg = OutboundMessage(
+            channel_name="slack",
+            chat_id="C123",
+            thread_id="thread-1",
+            text="hello",
+            connection_id="connection-1",
+        )
+        await channel.send(msg)
+
+        repo.get_credentials.assert_awaited_once_with("connection-1")
+        web_client_factory.assert_called_once_with(token="xoxb-connection-token")
+        web_client.chat_postMessage.assert_called_once()
+
+    anyio.run(go)
+
+
+def test_slack_http_events_mode_initializes_operator_web_client(monkeypatch):
+    import anyio
+
+    from app.channels.slack import SlackChannel
+
+    class FakeWebClient:
+        def __init__(self, token: str) -> None:
+            self.token = token
+            self.messages: list[dict] = []
+
+        def auth_test(self):
+            return {"user_id": "B-http"}
+
+        def chat_postMessage(self, **kwargs):
+            self.messages.append(kwargs)
+
+    slack_sdk = ModuleType("slack_sdk")
+    slack_sdk.WebClient = FakeWebClient
+    socket_mode = ModuleType("slack_sdk.socket_mode")
+    socket_mode.SocketModeClient = object
+    response = ModuleType("slack_sdk.socket_mode.response")
+    response.SocketModeResponse = object
+    monkeypatch.setitem(sys.modules, "slack_sdk", slack_sdk)
+    monkeypatch.setitem(sys.modules, "slack_sdk.socket_mode", socket_mode)
+    monkeypatch.setitem(sys.modules, "slack_sdk.socket_mode.response", response)
+
+    async def go():
+        channel = SlackChannel(
+            bus=MessageBus(),
+            config={
+                "bot_token": "xoxb-operator",
+                "event_delivery": "http",
+                "connection_repo": MagicMock(),
+            },
+        )
+
+        await channel.start()
+        assert channel._running is True
+        assert channel._web_client is not None
+        assert channel._web_client.token == "xoxb-operator"
+        assert channel._bot_user_id == "B-http"
+
+        channel._post_connection_reply("C123", "Slack connected to DeerFlow.", "1710000000.000100")
+
+        assert channel._web_client.messages == [
+            {
+                "channel": "C123",
+                "text": "Slack connected to DeerFlow.",
+                "thread_ts": "1710000000.000100",
+            }
+        ]
+        await channel.stop()
+
+    anyio.run(go)

backend/tests/test_slash_skills.py

@@ -0,0 +1,557 @@
+import asyncio
+import hashlib
+from pathlib import Path
+from types import SimpleNamespace
+
+from langchain.agents.middleware.types import ModelRequest
+from langchain_core.messages import AIMessage, HumanMessage
+
+from app.channels.commands import KNOWN_CHANNEL_COMMANDS
+from deerflow.agents.middlewares import skill_activation_middleware as middleware_module
+from deerflow.agents.middlewares.skill_activation_middleware import SkillActivationMiddleware, is_slash_skill_activation_reminder
+from deerflow.skills.slash import RESERVED_SLASH_SKILL_NAMES, parse_slash_skill_reference, resolve_slash_skill
+from deerflow.skills.types import Skill, SkillCategory
+from deerflow.utils.messages import ORIGINAL_USER_CONTENT_KEY
+
+
+def _make_skill(tmp_path: Path, name: str, content: str = "skill body") -> Skill:
+    skill_dir = tmp_path / name
+    skill_dir.mkdir()
+    skill_file = skill_dir / "SKILL.md"
+    skill_file.write_text(content, encoding="utf-8")
+    return Skill(
+        name=name,
+        description=f"Description for {name}",
+        license="MIT",
+        skill_dir=skill_dir,
+        skill_file=skill_file,
+        relative_path=Path(name),
+        category=SkillCategory.CUSTOM,
+        enabled=True,
+    )
+
+
+def _make_storage(tmp_path: Path, skills: list[Skill]):
+    return SimpleNamespace(
+        load_skills=lambda *, enabled_only: [skill for skill in skills if skill.enabled] if enabled_only else skills,
+        get_container_root=lambda: "/mnt/skills",
+        get_skills_root_path=lambda: tmp_path,
+    )
+
+
+def _make_model_request(messages: list[HumanMessage], *, runtime=None) -> ModelRequest:
+    return ModelRequest(
+        model=object(),
+        messages=messages,
+        state={"messages": list(messages)},
+        runtime=runtime,
+    )
+
+
+def test_parse_slash_skill_reference_extracts_name_and_remaining_text():
+    parsed = parse_slash_skill_reference("/data-analysis analyze uploads/foo.csv")
+
+    assert parsed is not None
+    assert parsed.name == "data-analysis"
+    assert parsed.remaining_text == "analyze uploads/foo.csv"
+
+
+def test_parse_slash_skill_reference_accepts_skill_name_without_task():
+    parsed = parse_slash_skill_reference("/data-analysis")
+
+    assert parsed is not None
+    assert parsed.name == "data-analysis"
+    assert parsed.remaining_text == ""
+
+
+def test_parse_slash_skill_reference_rejects_invalid_names():
+    assert parse_slash_skill_reference("/DataAnalysis run") is None
+    assert parse_slash_skill_reference("/data_analysis run") is None
+    assert parse_slash_skill_reference("please use /data-analysis") is None
+    assert parse_slash_skill_reference("  /data-analysis run") is None
+    assert parse_slash_skill_reference("/data-analysis分析这个文档") is None
+
+
+def test_resolve_slash_skill_ignores_reserved_control_commands(tmp_path):
+    for command in ["bootstrap", "help", "memory", "models", "new", "status"]:
+        skill = _make_skill(tmp_path, command)
+
+        assert resolve_slash_skill(f"/{command} create an agent", [skill]) is None
+
+
+def test_reserved_slash_skill_names_match_channel_commands():
+    assert RESERVED_SLASH_SKILL_NAMES == {command.removeprefix("/") for command in KNOWN_CHANNEL_COMMANDS}
+
+
+def test_resolve_slash_skill_respects_available_skill_whitelist(tmp_path):
+    skill = _make_skill(tmp_path, "data-analysis")
+
+    assert resolve_slash_skill("/data-analysis run", [skill], available_skills=set()) is None
+
+    resolved = resolve_slash_skill("/data-analysis run", [skill], available_skills={"data-analysis"})
+    assert resolved is not None
+    assert resolved.skill.name == "data-analysis"
+    assert resolved.remaining_text == "run"
+    assert resolved.container_file_path == "/mnt/skills/custom/data-analysis/SKILL.md"
+
+
+def test_resolve_slash_skill_rejects_disabled_skills(tmp_path):
+    skill = _make_skill(tmp_path, "data-analysis")
+    skill.enabled = False
+
+    assert resolve_slash_skill("/data-analysis run", [skill]) is None
+
+
+def test_skill_activation_middleware_injects_hidden_human_context_for_model_call(monkeypatch, tmp_path):
+    skill = _make_skill(tmp_path, "data-analysis", content="# Data Analysis\nUse pandas.")
+    monkeypatch.setattr(middleware_module, "get_or_new_skill_storage", lambda **kwargs: _make_storage(tmp_path, [skill]))
+
+    middleware = SkillActivationMiddleware()
+    original = HumanMessage(content="/data-analysis analyze uploads/foo.csv", id="msg-1")
+    request = _make_model_request([original])
+    captured = {}
+
+    def handler(model_request: ModelRequest):
+        captured["messages"] = model_request.messages
+        return AIMessage(content="ok")
+
+    result = middleware.wrap_model_call(request, handler)
+
+    assert isinstance(result, AIMessage)
+    assert result.content == "ok"
+    activation_msg, user_msg = captured["messages"]
+    assert is_slash_skill_activation_reminder(activation_msg)
+    assert activation_msg.additional_kwargs["hide_from_ui"] is True
+    assert "Use pandas." in activation_msg.content
+    assert "<user_request>\nanalyze uploads/foo.csv\n</user_request>" in activation_msg.content
+    assert user_msg.content == original.content
+    assert request.state["messages"] == [original]
+
+
+def test_skill_activation_middleware_does_not_duplicate_existing_activation(monkeypatch, tmp_path):
+    skill = _make_skill(tmp_path, "data-analysis", content="# Data Analysis\nUse pandas.")
+    monkeypatch.setattr(middleware_module, "get_or_new_skill_storage", lambda **kwargs: _make_storage(tmp_path, [skill]))
+
+    middleware = SkillActivationMiddleware()
+    original = HumanMessage(content="/data-analysis analyze uploads/foo.csv", id="msg-1")
+    first_capture = {}
+
+    def first_handler(model_request: ModelRequest):
+        first_capture["messages"] = model_request.messages
+        return AIMessage(content="ok")
+
+    first_result = middleware.wrap_model_call(_make_model_request([original]), first_handler)
+
+    assert isinstance(first_result, AIMessage)
+    activation_msg, user_msg = first_capture["messages"]
+    assert is_slash_skill_activation_reminder(activation_msg)
+
+    second_capture = {}
+
+    def second_handler(model_request: ModelRequest):
+        second_capture["messages"] = model_request.messages
+        return AIMessage(content="ok")
+
+    second_result = middleware.wrap_model_call(_make_model_request([activation_msg, user_msg]), second_handler)
+
+    assert isinstance(second_result, AIMessage)
+    assert second_capture["messages"] == [activation_msg, user_msg]
+    assert sum(is_slash_skill_activation_reminder(message) for message in second_capture["messages"]) == 1
+
+
+def test_skill_activation_middleware_does_not_duplicate_activation_separated_by_hidden_context(monkeypatch, tmp_path):
+    skill = _make_skill(tmp_path, "data-analysis", content="# Data Analysis\nUse pandas.")
+    monkeypatch.setattr(middleware_module, "get_or_new_skill_storage", lambda **kwargs: _make_storage(tmp_path, [skill]))
+
+    middleware = SkillActivationMiddleware()
+    original = HumanMessage(content="/data-analysis analyze uploads/foo.csv", id="msg-1")
+    first_capture = {}
+
+    def first_handler(model_request: ModelRequest):
+        first_capture["messages"] = model_request.messages
+        return AIMessage(content="ok")
+
+    middleware.wrap_model_call(_make_model_request([original]), first_handler)
+    activation_msg, user_msg = first_capture["messages"]
+    hidden_context = HumanMessage(content="dynamic context", additional_kwargs={"hide_from_ui": True})
+    second_capture = {}
+
+    def second_handler(model_request: ModelRequest):
+        second_capture["messages"] = model_request.messages
+        return AIMessage(content="ok")
+
+    second_result = middleware.wrap_model_call(_make_model_request([activation_msg, hidden_context, user_msg]), second_handler)
+
+    assert isinstance(second_result, AIMessage)
+    assert second_capture["messages"] == [activation_msg, hidden_context, user_msg]
+    assert sum(is_slash_skill_activation_reminder(message) for message in second_capture["messages"]) == 1
+
+
+def test_skill_activation_middleware_dedupes_immediately_previous_activation_without_target_id(monkeypatch, tmp_path):
+    skill = _make_skill(tmp_path, "data-analysis", content="# Data Analysis\nUse pandas.")
+    monkeypatch.setattr(middleware_module, "get_or_new_skill_storage", lambda **kwargs: _make_storage(tmp_path, [skill]))
+
+    middleware = SkillActivationMiddleware()
+    legacy_activation_msg = SkillActivationMiddleware._make_activation_message(
+        HumanMessage(content="/data-analysis analyze uploads/foo.csv"),
+        "existing activation context",
+    )
+    target = HumanMessage(content="/data-analysis analyze uploads/foo.csv", id="msg-1")
+    captured = {}
+
+    def handler(model_request: ModelRequest):
+        captured["messages"] = model_request.messages
+        return AIMessage(content="ok")
+
+    result = middleware.wrap_model_call(_make_model_request([legacy_activation_msg, target]), handler)
+
+    assert isinstance(result, AIMessage)
+    assert captured["messages"] == [legacy_activation_msg, target]
+    assert sum(is_slash_skill_activation_reminder(message) for message in captured["messages"]) == 1
+
+
+def test_skill_activation_middleware_async_injects_hidden_human_context_for_model_call(monkeypatch, tmp_path):
+    skill = _make_skill(tmp_path, "data-analysis", content="# Data Analysis\nUse pandas.")
+    monkeypatch.setattr(middleware_module, "get_or_new_skill_storage", lambda **kwargs: _make_storage(tmp_path, [skill]))
+
+    middleware = SkillActivationMiddleware()
+    original = HumanMessage(content="/data-analysis analyze uploads/foo.csv", id="msg-1")
+    request = _make_model_request([original])
+    captured = {}
+
+    async def handler(model_request: ModelRequest):
+        captured["messages"] = model_request.messages
+        return AIMessage(content="ok")
+
+    result = asyncio.run(middleware.awrap_model_call(request, handler))
+
+    assert isinstance(result, AIMessage)
+    assert result.content == "ok"
+    activation_msg, user_msg = captured["messages"]
+    assert is_slash_skill_activation_reminder(activation_msg)
+    assert activation_msg.additional_kwargs["hide_from_ui"] is True
+    assert "Use pandas." in activation_msg.content
+    assert "<user_request>\nanalyze uploads/foo.csv\n</user_request>" in activation_msg.content
+    assert user_msg.content == original.content
+    assert request.state["messages"] == [original]
+
+
+def test_skill_activation_middleware_uses_fallback_when_task_text_is_empty(monkeypatch, tmp_path):
+    skill = _make_skill(tmp_path, "data-analysis", content="# Data Analysis\nUse pandas.")
+    monkeypatch.setattr(middleware_module, "get_or_new_skill_storage", lambda **kwargs: _make_storage(tmp_path, [skill]))
+
+    middleware = SkillActivationMiddleware()
+    original = HumanMessage(content="/data-analysis", id="msg-1")
+    captured = {}
+
+    def handler(model_request: ModelRequest):
+        captured["messages"] = model_request.messages
+        return AIMessage(content="ok")
+
+    result = middleware.wrap_model_call(_make_model_request([original]), handler)
+
+    assert isinstance(result, AIMessage)
+    activation_msg = captured["messages"][0]
+    assert "No additional task text was provided after the slash skill command." in activation_msg.content
+
+
+def test_skill_activation_middleware_uses_original_user_content_when_uploads_are_injected(monkeypatch, tmp_path):
+    skill = _make_skill(tmp_path, "data-analysis", content="# Data Analysis\nUse pandas.")
+    monkeypatch.setattr(middleware_module, "get_or_new_skill_storage", lambda **kwargs: _make_storage(tmp_path, [skill]))
+
+    middleware = SkillActivationMiddleware()
+    original = HumanMessage(
+        content="<uploaded_files>\n- report.pdf\n</uploaded_files>\n\n/data-analysis 分析这个文档",
+        id="msg-1",
+        additional_kwargs={ORIGINAL_USER_CONTENT_KEY: "/data-analysis 分析这个文档"},
+    )
+    captured = {}
+
+    def handler(model_request: ModelRequest):
+        captured["messages"] = model_request.messages
+        return AIMessage(content="ok")
+
+    result = middleware.wrap_model_call(_make_model_request([original]), handler)
+
+    assert isinstance(result, AIMessage)
+    assert result.content == "ok"
+    activation_msg, user_msg = captured["messages"]
+    assert is_slash_skill_activation_reminder(activation_msg)
+    assert "Use pandas." in activation_msg.content
+    assert "<user_request>\n分析这个文档\n</user_request>" in activation_msg.content
+    assert user_msg.content == original.content
+    assert user_msg.additional_kwargs[ORIGINAL_USER_CONTENT_KEY] == "/data-analysis 分析这个文档"
+
+
+def test_skill_activation_middleware_activates_from_list_content(monkeypatch, tmp_path):
+    skill = _make_skill(tmp_path, "data-analysis", content="# Data Analysis\nUse pandas.")
+    monkeypatch.setattr(middleware_module, "get_or_new_skill_storage", lambda **kwargs: _make_storage(tmp_path, [skill]))
+
+    middleware = SkillActivationMiddleware()
+    original = HumanMessage(content=[{"type": "text", "text": "/data-analysis analyze uploads/foo.csv"}], id="msg-1")
+    captured = {}
+
+    def handler(model_request: ModelRequest):
+        captured["messages"] = model_request.messages
+        return AIMessage(content="ok")
+
+    result = middleware.wrap_model_call(_make_model_request([original]), handler)
+
+    assert isinstance(result, AIMessage)
+    activation_msg, user_msg = captured["messages"]
+    assert is_slash_skill_activation_reminder(activation_msg)
+    assert "<user_request>\nanalyze uploads/foo.csv\n</user_request>" in activation_msg.content
+    assert user_msg.content == original.content
+
+
+def test_skill_activation_middleware_records_activation_audit_event(monkeypatch, tmp_path):
+    skill = _make_skill(tmp_path, "data-analysis", content="# Data Analysis\nUse pandas.")
+    monkeypatch.setattr(middleware_module, "get_or_new_skill_storage", lambda **kwargs: _make_storage(tmp_path, [skill]))
+
+    recorded = []
+    journal = SimpleNamespace(record_middleware=lambda *args, **kwargs: recorded.append((args, kwargs)))
+    runtime = SimpleNamespace(context={"__run_journal": journal})
+    middleware = SkillActivationMiddleware()
+    original = HumanMessage(content="/data-analysis analyze uploads/foo.csv", id="msg-1")
+
+    def handler(model_request: ModelRequest):
+        return AIMessage(content="ok")
+
+    result = middleware.wrap_model_call(_make_model_request([original], runtime=runtime), handler)
+
+    assert isinstance(result, AIMessage)
+    assert len(recorded) == 1
+    args, kwargs = recorded[0]
+    assert args == ("skill_activation",)
+    assert kwargs["name"] == "SkillActivationMiddleware"
+    assert kwargs["hook"] == "wrap_model_call"
+    assert kwargs["action"] == "activate"
+    assert kwargs["changes"] == {
+        "skill_name": "data-analysis",
+        "category": "custom",
+        "path": "/mnt/skills/custom/data-analysis/SKILL.md",
+        "content_hash": hashlib.sha256(b"# Data Analysis\nUse pandas.").hexdigest(),
+    }
+
+
+def test_skill_activation_middleware_async_records_activation_audit_event(monkeypatch, tmp_path):
+    skill = _make_skill(tmp_path, "data-analysis", content="# Data Analysis\nUse pandas.")
+    monkeypatch.setattr(middleware_module, "get_or_new_skill_storage", lambda **kwargs: _make_storage(tmp_path, [skill]))
+
+    recorded = []
+    journal = SimpleNamespace(record_middleware=lambda *args, **kwargs: recorded.append((args, kwargs)))
+    runtime = SimpleNamespace(context={"__run_journal": journal})
+    middleware = SkillActivationMiddleware()
+    original = HumanMessage(content="/data-analysis analyze uploads/foo.csv", id="msg-1")
+
+    async def handler(model_request: ModelRequest):
+        return AIMessage(content="ok")
+
+    result = asyncio.run(middleware.awrap_model_call(_make_model_request([original], runtime=runtime), handler))
+
+    assert isinstance(result, AIMessage)
+    assert len(recorded) == 1
+    args, kwargs = recorded[0]
+    assert args == ("skill_activation",)
+    assert kwargs["hook"] == "awrap_model_call"
+    assert kwargs["changes"]["skill_name"] == "data-analysis"
+    assert kwargs["changes"]["content_hash"] == hashlib.sha256(b"# Data Analysis\nUse pandas.").hexdigest()
+
+
+def test_skill_activation_middleware_ignores_activation_audit_errors(monkeypatch, tmp_path):
+    skill = _make_skill(tmp_path, "data-analysis", content="# Data Analysis\nUse pandas.")
+    monkeypatch.setattr(middleware_module, "get_or_new_skill_storage", lambda **kwargs: _make_storage(tmp_path, [skill]))
+
+    journal = SimpleNamespace(record_middleware=lambda *args, **kwargs: (_ for _ in ()).throw(RuntimeError("db down")))
+    runtime = SimpleNamespace(context={"__run_journal": journal})
+    middleware = SkillActivationMiddleware()
+    original = HumanMessage(content="/data-analysis analyze uploads/foo.csv", id="msg-1")
+
+    def handler(model_request: ModelRequest):
+        return AIMessage(content="ok")
+
+    result = middleware.wrap_model_call(_make_model_request([original], runtime=runtime), handler)
+
+    assert isinstance(result, AIMessage)
+    assert result.content == "ok"
+
+
+def test_skill_activation_middleware_activates_only_latest_real_user_message(monkeypatch, tmp_path):
+    skill = _make_skill(tmp_path, "data-analysis", content="# Data Analysis\nUse pandas.")
+    monkeypatch.setattr(middleware_module, "get_or_new_skill_storage", lambda **kwargs: _make_storage(tmp_path, [skill]))
+
+    middleware = SkillActivationMiddleware()
+    old_slash = HumanMessage(content="/data-analysis old request", id="msg-1")
+    latest_user = HumanMessage(content="continue normally", id="msg-2")
+    request = _make_model_request([old_slash, AIMessage(content="done"), latest_user])
+    captured = {}
+
+    def handler(model_request: ModelRequest):
+        captured["messages"] = model_request.messages
+        return AIMessage(content="ok")
+
+    result = middleware.wrap_model_call(request, handler)
+
+    assert isinstance(result, AIMessage)
+    assert captured["messages"] == request.messages
+    assert not any(is_slash_skill_activation_reminder(message) for message in captured["messages"])
+
+
+def test_skill_activation_middleware_ignores_hidden_and_summary_user_messages(monkeypatch, tmp_path):
+    skill = _make_skill(tmp_path, "data-analysis", content="# Data Analysis\nUse pandas.")
+    monkeypatch.setattr(middleware_module, "get_or_new_skill_storage", lambda **kwargs: _make_storage(tmp_path, [skill]))
+
+    middleware = SkillActivationMiddleware()
+    real_user = HumanMessage(content="continue normally", id="msg-1")
+    hidden_slash = HumanMessage(content="/data-analysis hidden request", id="msg-2", additional_kwargs={"hide_from_ui": True})
+    summary_slash = HumanMessage(content="/data-analysis summary request", id="msg-3", name="summary")
+    request = _make_model_request([real_user, hidden_slash, summary_slash])
+    captured = {}
+
+    def handler(model_request: ModelRequest):
+        captured["messages"] = model_request.messages
+        return AIMessage(content="ok")
+
+    result = middleware.wrap_model_call(request, handler)
+
+    assert isinstance(result, AIMessage)
+    assert captured["messages"] == request.messages
+    assert not any(is_slash_skill_activation_reminder(message) for message in captured["messages"])
+
+
+def test_skill_activation_middleware_returns_clear_error_for_disallowed_skill(monkeypatch, tmp_path):
+    skill = _make_skill(tmp_path, "data-analysis")
+    monkeypatch.setattr(middleware_module, "get_or_new_skill_storage", lambda **kwargs: _make_storage(tmp_path, [skill]))
+
+    middleware = SkillActivationMiddleware(available_skills={"frontend-design"})
+    original = HumanMessage(content="/data-analysis run")
+
+    def handler(model_request: ModelRequest):
+        raise AssertionError("handler should not be called for invalid slash skills")
+
+    result = middleware.wrap_model_call(_make_model_request([original]), handler)
+
+    assert isinstance(result, AIMessage)
+    assert "not available for this agent" in result.content
+
+
+def test_skill_activation_middleware_returns_clear_error_for_missing_skill(monkeypatch, tmp_path):
+    monkeypatch.setattr(middleware_module, "get_or_new_skill_storage", lambda **kwargs: _make_storage(tmp_path, []))
+
+    middleware = SkillActivationMiddleware()
+    original = HumanMessage(content="/data-analysis run")
+
+    def handler(model_request: ModelRequest):
+        raise AssertionError("handler should not be called for missing slash skills")
+
+    result = middleware.wrap_model_call(_make_model_request([original]), handler)
+
+    assert isinstance(result, AIMessage)
+    assert "not installed" in result.content
+
+
+def test_skill_activation_middleware_returns_clear_error_for_disabled_skill(monkeypatch, tmp_path):
+    skill = _make_skill(tmp_path, "data-analysis")
+    skill.enabled = False
+    monkeypatch.setattr(middleware_module, "get_or_new_skill_storage", lambda **kwargs: _make_storage(tmp_path, [skill]))
+
+    middleware = SkillActivationMiddleware()
+    original = HumanMessage(content="/data-analysis run")
+
+    def handler(model_request: ModelRequest):
+        raise AssertionError("handler should not be called for disabled slash skills")
+
+    result = middleware.wrap_model_call(_make_model_request([original]), handler)
+
+    assert isinstance(result, AIMessage)
+    assert "installed but disabled" in result.content
+
+
+def test_skill_activation_middleware_escapes_activation_content(monkeypatch, tmp_path):
+    skill = _make_skill(
+        tmp_path,
+        "data-analysis",
+        content="# Data Analysis\nUse <xml> & avoid </skill> collisions.\n----- END SKILL.md -----",
+    )
+    monkeypatch.setattr(middleware_module, "get_or_new_skill_storage", lambda **kwargs: _make_storage(tmp_path, [skill]))
+
+    middleware = SkillActivationMiddleware()
+    original = HumanMessage(content="/data-analysis analyze </user_request>")
+    captured = {}
+
+    def handler(model_request: ModelRequest):
+        captured["messages"] = model_request.messages
+        return AIMessage(content="ok")
+
+    result = middleware.wrap_model_call(_make_model_request([original]), handler)
+
+    assert isinstance(result, AIMessage)
+    activation_msg = captured["messages"][0]
+    assert '<skill_content encoding="xml-escaped">' in activation_msg.content
+    assert "analyze &lt;/user_request&gt;" in activation_msg.content
+    assert "Use &lt;xml&gt; &amp; avoid &lt;/skill&gt; collisions." in activation_msg.content
+    assert "----- BEGIN SKILL.md -----" not in activation_msg.content
+
+
+def test_skill_activation_middleware_rejects_skill_file_outside_skills_root(monkeypatch, tmp_path):
+    skills_root = tmp_path / "skills"
+    skill_dir = skills_root / "custom" / "data-analysis"
+    skill_dir.mkdir(parents=True)
+    outside_dir = tmp_path / "outside"
+    outside_dir.mkdir()
+    outside_file = outside_dir / "SKILL.md"
+    outside_file.write_text("# Leaked\nDo not read me.", encoding="utf-8")
+    (skill_dir / "SKILL.md").symlink_to(outside_file)
+    skill = Skill(
+        name="data-analysis",
+        description="Description for data-analysis",
+        license="MIT",
+        skill_dir=skill_dir,
+        skill_file=skill_dir / "SKILL.md",
+        relative_path=Path("data-analysis"),
+        category=SkillCategory.CUSTOM,
+        enabled=True,
+    )
+    monkeypatch.setattr(middleware_module, "get_or_new_skill_storage", lambda **kwargs: _make_storage(skills_root, [skill]))
+
+    middleware = SkillActivationMiddleware()
+
+    def handler(model_request: ModelRequest):
+        raise AssertionError("handler should not be called when SKILL.md fails safety checks")
+
+    result = middleware.wrap_model_call(_make_model_request([HumanMessage(content="/data-analysis run")]), handler)
+
+    assert isinstance(result, AIMessage)
+    assert "could not be loaded safely" in result.content
+
+
+def test_skill_activation_middleware_reports_missing_skill_file_safely(monkeypatch, tmp_path):
+    skill = _make_skill(tmp_path, "data-analysis")
+    skill.skill_file.unlink()
+    monkeypatch.setattr(middleware_module, "get_or_new_skill_storage", lambda **kwargs: _make_storage(tmp_path, [skill]))
+
+    middleware = SkillActivationMiddleware()
+
+    def handler(model_request: ModelRequest):
+        raise AssertionError("handler should not be called when SKILL.md is missing")
+
+    result = middleware.wrap_model_call(_make_model_request([HumanMessage(content="/data-analysis run")]), handler)
+
+    assert isinstance(result, AIMessage)
+    assert "could not be loaded safely" in result.content
+
+
+def test_skill_activation_middleware_reports_invalid_utf8_skill_file_safely(monkeypatch, tmp_path):
+    skill = _make_skill(tmp_path, "data-analysis")
+    skill.skill_file.write_bytes(b"\xff\xfe\x00")
+    monkeypatch.setattr(middleware_module, "get_or_new_skill_storage", lambda **kwargs: _make_storage(tmp_path, [skill]))
+
+    middleware = SkillActivationMiddleware()
+
+    def handler(model_request: ModelRequest):
+        raise AssertionError("handler should not be called when SKILL.md is not valid UTF-8")
+
+    result = middleware.wrap_model_call(_make_model_request([HumanMessage(content="/data-analysis run")]), handler)
+
+    assert isinstance(result, AIMessage)
+    assert "could not be loaded safely" in result.content

backend/tests/test_telegram_channel_connections.py

@@ -0,0 +1,100 @@
+"""Tests for Telegram deep-link channel connections."""
+
+from __future__ import annotations
+
+from datetime import UTC, datetime, timedelta
+from pathlib import Path
+from unittest.mock import AsyncMock, MagicMock
+
+import pytest
+
+from app.channels.message_bus import MessageBus
+from app.channels.telegram import TelegramChannel
+
+
+@pytest.fixture
+async def repo(tmp_path: Path):
+    from deerflow.persistence.channel_connections import ChannelConnectionRepository, ChannelCredentialCipher
+    from deerflow.persistence.engine import close_engine, get_session_factory, init_engine
+
+    await init_engine("sqlite", url=f"sqlite+aiosqlite:///{tmp_path / 'telegram.db'}", sqlite_dir=str(tmp_path))
+    try:
+        yield ChannelConnectionRepository(
+            get_session_factory(),
+            cipher=ChannelCredentialCipher.from_key("telegram-secret"),
+        )
+    finally:
+        await close_engine()
+
+
+def _telegram_update(*, text: str = "/start", user_id: int = 42, chat_id: int = 100, chat_type: str = "private"):
+    update = MagicMock()
+    update.effective_user.id = user_id
+    update.effective_user.username = "alice"
+    update.effective_user.full_name = "Alice Example"
+    update.effective_chat.id = chat_id
+    update.effective_chat.type = chat_type
+    update.message.text = text
+    update.message.message_id = 55
+    update.message.reply_to_message = None
+    update.message.reply_text = AsyncMock()
+    return update
+
+
+@pytest.mark.anyio
+async def test_start_with_deep_link_state_binds_telegram_chat(repo):
+    state = "telegram-bind-state"
+    await repo.create_oauth_state(
+        owner_user_id="deerflow-user-1",
+        provider="telegram",
+        state=state,
+        expires_at=datetime.now(UTC) + timedelta(minutes=5),
+    )
+    channel = TelegramChannel(
+        bus=MessageBus(),
+        config={"bot_token": "test-token", "connection_repo": repo},
+    )
+    update = _telegram_update(text=f"/start {state}")
+    context = MagicMock()
+    context.args = [state]
+
+    await channel._cmd_start(update, context)
+
+    connections = await repo.list_connections("deerflow-user-1")
+    assert len(connections) == 1
+    assert connections[0]["provider"] == "telegram"
+    assert connections[0]["external_account_id"] == "42"
+    assert connections[0]["external_account_name"] == "Alice Example"
+    assert connections[0]["workspace_id"] == "100"
+    assert connections[0]["metadata"]["chat_type"] == "private"
+    update.message.reply_text.assert_awaited_once()
+    assert "connected" in update.message.reply_text.await_args.args[0].lower()
+
+
+@pytest.mark.anyio
+async def test_bound_telegram_message_publishes_connection_identity(repo):
+    connection = await repo.upsert_connection(
+        owner_user_id="deerflow-user-1",
+        provider="telegram",
+        external_account_id="42",
+        external_account_name="Alice Example",
+        workspace_id="100",
+        metadata={"chat_type": "private"},
+    )
+    bus = MessageBus()
+    channel = TelegramChannel(
+        bus=bus,
+        config={"bot_token": "test-token", "connection_repo": repo},
+    )
+    channel._main_loop = __import__("asyncio").get_event_loop()
+    channel._send_running_reply = AsyncMock()
+
+    await channel._on_text(_telegram_update(text="hello"), None)
+    inbound = await bus.get_inbound()
+
+    assert inbound.connection_id == connection["id"]
+    assert inbound.owner_user_id == "deerflow-user-1"
+    assert inbound.workspace_id == "100"
+    assert inbound.user_id == "42"
+    assert inbound.chat_id == "100"
+    assert inbound.text == "hello"

backend/tests/test_uploads_middleware_core_logic.py

@@ -14,6 +14,7 @@ from langchain_core.messages import AIMessage, HumanMessage
 
 from deerflow.agents.middlewares.uploads_middleware import UploadsMiddleware
 from deerflow.config.paths import Paths
+from deerflow.utils.messages import ORIGINAL_USER_CONTENT_KEY
 
 THREAD_ID = "thread-abc123"
 
@@ -263,6 +264,22 @@ class TestBeforeAgent:
         assert "<uploaded_files>" in combined_text
         assert "analyse this" in combined_text
 
+    def test_list_content_preserves_original_slash_skill_text(self, tmp_path):
+        mw = _middleware(tmp_path)
+        uploads_dir = _uploads_dir(tmp_path)
+        (uploads_dir / "data.csv").write_bytes(b"a,b")
+
+        msg = _human(
+            [{"type": "text", "text": "/data-analysis analyze data.csv"}],
+            files=[{"filename": "data.csv", "size": 3, "path": "/mnt/user-data/uploads/data.csv"}],
+        )
+        result = mw.before_agent(self._state(msg), _runtime())
+
+        assert result is not None
+        updated_msg = result["messages"][-1]
+        assert isinstance(updated_msg.content, list)
+        assert updated_msg.additional_kwargs[ORIGINAL_USER_CONTENT_KEY] == "/data-analysis analyze data.csv"
+
     def test_preserves_additional_kwargs_on_updated_message(self, tmp_path):
         mw = _middleware(tmp_path)
         uploads_dir = _uploads_dir(tmp_path)
@@ -278,6 +295,37 @@ class TestBeforeAgent:
         assert updated_kwargs.get("files") == files_meta
         assert updated_kwargs.get("element") == "task"
 
+    def test_preserves_original_user_content_before_upload_context(self, tmp_path):
+        mw = _middleware(tmp_path)
+        uploads_dir = _uploads_dir(tmp_path)
+        (uploads_dir / "report.pdf").write_bytes(b"pdf")
+
+        msg = _human(
+            "/data-analysis 分析这个文档",
+            files=[{"filename": "report.pdf", "size": 3, "path": "/mnt/user-data/uploads/report.pdf"}],
+        )
+        result = mw.before_agent(self._state(msg), _runtime())
+
+        assert result is not None
+        updated_msg = result["messages"][-1]
+        assert updated_msg.content.startswith("<uploaded_files>")
+        assert updated_msg.additional_kwargs[ORIGINAL_USER_CONTENT_KEY] == "/data-analysis 分析这个文档"
+
+    def test_preserves_existing_original_user_content_marker(self, tmp_path):
+        mw = _middleware(tmp_path)
+        uploads_dir = _uploads_dir(tmp_path)
+        (uploads_dir / "report.pdf").write_bytes(b"pdf")
+
+        msg = _human(
+            "<uploaded_files>\nold\n</uploaded_files>\n\n/data-analysis run",
+            files=[{"filename": "report.pdf", "size": 3, "path": "/mnt/user-data/uploads/report.pdf"}],
+            **{ORIGINAL_USER_CONTENT_KEY: "/data-analysis run"},
+        )
+        result = mw.before_agent(self._state(msg), _runtime())
+
+        assert result is not None
+        assert result["messages"][-1].additional_kwargs[ORIGINAL_USER_CONTENT_KEY] == "/data-analysis run"
+
     def test_uploaded_files_returned_in_state_update(self, tmp_path):
         mw = _middleware(tmp_path)
         uploads_dir = _uploads_dir(tmp_path)

backend/tests/test_uvicorn_reload_exclude.py

@@ -0,0 +1,185 @@
+"""Regression for #3459 / #3454 — dev gateway reload-exclude must not crash.
+
+#3426 switched the dev gateway's ``--reload-exclude`` patterns from relative
+(``sandbox/``) to absolute (``$REPO_ROOT/backend/sandbox``). uvicorn only
+excludes such a path directly when it already exists as a directory; otherwise
+it falls back to ``Path.cwd().glob(pattern)``, and on **Python 3.12**
+``pathlib.Path.glob()`` raises ``NotImplementedError: Non-relative patterns are
+unsupported`` for an absolute pattern. ``serve.sh`` created the ``.deer-flow``
+excludes but not ``backend/sandbox``, so a fresh checkout crashed ``make dev``
+on startup.
+
+Two layers of coverage:
+
+* ``test_*_resolve_*`` exercises uvicorn's real ``resolve_reload_patterns`` to
+  pin the failure mode and the fix's mechanism.
+* ``test_launcher_precreates_every_absolute_reload_exclude`` enforces the actual
+  invariant on both launchers: every absolute exclude dir is ``mkdir -p``'d
+  before uvicorn starts. This encodes the root cause, so any future absolute
+  exclude that forgets its ``mkdir`` fails here.
+"""
+
+from __future__ import annotations
+
+import re
+import shlex
+import subprocess
+import sys
+from pathlib import Path
+
+import pytest
+from uvicorn.config import resolve_reload_patterns
+
+REPO_ROOT = Path(__file__).resolve().parents[2]
+
+LAUNCHERS = {
+    "scripts/serve.sh": REPO_ROOT / "scripts" / "serve.sh",
+    "docker/dev-entrypoint.sh": REPO_ROOT / "docker" / "dev-entrypoint.sh",
+}
+
+# Shell terminators / redirects that end a simple command's argument list.
+_CMD_BOUNDARY = re.compile(r"[;&|<>]")
+
+
+def _logical_lines(script: str) -> list[str]:
+    """Fold ``\\``-continuations and drop comment lines, yielding logical lines.
+
+    A ``mkdir`` or ``--reload-exclude`` list split across lines with a trailing
+    backslash becomes one line here, so an argument on a continuation line can't
+    be silently dropped by per-line scanning.
+    """
+    folded = script.replace("\\\n", " ")
+    return [line for line in folded.splitlines() if not line.lstrip().startswith("#")]
+
+
+def _shlex(fragment: str) -> list[str]:
+    """Tokenize a shell fragment (quotes stripped, ``$VAR`` kept literal,
+    trailing ``# comment`` honored); tolerate pathological quoting."""
+    try:
+        return shlex.split(fragment, comments=True)
+    except ValueError:
+        return fragment.split()
+
+
+# ``--reload-exclude`` followed by ``=`` or whitespace, then a value that is a
+# single-quoted group, a double-quoted group, or a bare token. The quoted
+# alternatives match a *balanced* pair first, so serve.sh's surrounding
+# ``GATEWAY_EXTRA_FLAGS="..."`` closing quote is never swallowed into the value.
+_RELOAD_EXCLUDE = re.compile(r"""--reload-exclude[=\s]+('[^']*'|"[^"]*"|[^\s'"]+)""")
+
+
+def _reload_exclude_values(script: str) -> list[str]:
+    """Every ``--reload-exclude`` value, with surrounding quotes removed.
+
+    Handles both CLI forms (``--reload-exclude=<value>`` and the space form
+    ``--reload-exclude <value>``) and both shell quotings the launchers use:
+
+    * ``docker/dev-entrypoint.sh`` puts each flag on its own line.
+    * ``scripts/serve.sh`` packs every flag into a single double-quoted
+      ``GATEWAY_EXTRA_FLAGS="... --reload-exclude='$X' ..."`` assignment. A
+      whole-line ``shlex`` would collapse that assignment into one token and
+      find no flags (this is what regressed serve.sh in CI); matching balanced
+      inner quotes here keeps the assignment's closing ``"`` out of the value,
+      so every exclude — including the last ``$BACKEND_RUNTIME_HOME`` — is seen.
+    """
+    values: list[str] = []
+    for line in _logical_lines(script):
+        for raw in _RELOAD_EXCLUDE.findall(line):
+            values.append(raw.strip("\"'"))
+    return values
+
+
+def _mkdir_dirs(script: str) -> set[str]:
+    """Exact set of directories created by every ``mkdir`` command.
+
+    Tokenizes each ``mkdir`` argument list rather than substring-matching, so
+    ``/app/backend/sandbox`` is not falsely considered created by, say,
+    ``mkdir -p /app/backend/sandbox-other``.
+    """
+    dirs: set[str] = set()
+    for line in _logical_lines(script):
+        match = re.search(r"\bmkdir\b(.*)", line)
+        if not match:
+            continue
+        args = _CMD_BOUNDARY.split(match.group(1), maxsplit=1)[0]
+        for token in _shlex(args):
+            if token.startswith("-"):  # skip flags such as -p
+                continue
+            dirs.add(token)
+    return dirs
+
+
+@pytest.mark.skipif(
+    sys.version_info >= (3, 13),
+    reason="pathlib accepts absolute glob patterns on 3.13+, so the crash is 3.12-only",
+)
+def test_resolve_reload_patterns_crashes_on_missing_absolute_dir(tmp_path):
+    """The exact #3454 failure: absolute exclude + missing dir on Python 3.12."""
+    missing = tmp_path / "sandbox"  # absolute path that does not exist yet
+    assert not missing.exists()
+    with pytest.raises(NotImplementedError):
+        resolve_reload_patterns([str(missing)], [])
+
+
+def test_resolve_reload_patterns_is_safe_once_dir_exists(tmp_path):
+    """The fix's mechanism: a pre-created dir takes uvicorn's is_dir() path."""
+    sandbox = tmp_path / "sandbox"
+    sandbox.mkdir()
+    _patterns, directories = resolve_reload_patterns([str(sandbox)], [])
+    resolved = {d.resolve() for d in directories}
+    assert sandbox.resolve() in resolved
+
+
+@pytest.mark.parametrize("name", list(LAUNCHERS))
+def test_launcher_precreates_every_absolute_reload_exclude(name):
+    """Every absolute ``--reload-exclude`` dir must be created by ``mkdir`` first.
+
+    Relative glob patterns (``*.pyc``, ``__pycache__``) are safe and skipped;
+    anything anchored at ``/`` or a shell variable is an absolute path that
+    uvicorn would glob — and crash on — unless it already exists. Membership is
+    an exact match against the parsed ``mkdir`` argument set (not a substring
+    test), so a path-prefix can't produce a false pass.
+    """
+    script = LAUNCHERS[name].read_text(encoding="utf-8")
+    created = _mkdir_dirs(script)
+
+    absolute_excludes = [v for v in _reload_exclude_values(script) if v.startswith(("/", "$"))]
+    assert absolute_excludes, f"{name}: expected at least one absolute reload-exclude"
+
+    for value in absolute_excludes:
+        assert value in created, f"{name}: absolute reload-exclude {value!r} is never created via mkdir (created dirs: {sorted(created)})"
+
+
+@pytest.mark.parametrize("name", list(LAUNCHERS))
+def test_sandbox_mkdir_precedes_uvicorn_launch(name):
+    """The sandbox mkdir must come before the uvicorn launch, not just exist.
+
+    ``_mkdir_dirs`` only proves the mkdir is present somewhere; this pins script
+    order so a future edit can't move (or guard) the mkdir below the launch and
+    silently reintroduce the #3454 crash on a fresh checkout. ``uv run uvicorn``
+    matches the launch but not serve.sh's ``stop_all`` kill line.
+    """
+    lines = LAUNCHERS[name].read_text(encoding="utf-8").splitlines()
+    launch_idx = next((i for i, ln in enumerate(lines) if "uv run uvicorn" in ln), None)
+    mkdir_idx = next((i for i, ln in enumerate(lines) if re.search(r"\bmkdir\b", ln) and "sandbox" in ln), None)
+
+    assert launch_idx is not None, f"{name}: could not locate the 'uv run uvicorn' launch line"
+    assert mkdir_idx is not None, f"{name}: could not locate the sandbox mkdir line"
+    assert mkdir_idx < launch_idx, f"{name}: sandbox mkdir (line {mkdir_idx + 1}) must precede uvicorn launch (line {launch_idx + 1})"
+
+
+def test_precreated_sandbox_artifacts_are_gitignored():
+    """backend/sandbox is runtime state — its contents must stay out of git so
+    sandbox artifacts can't be accidentally committed (matches the reload-exclude
+    intent). A content path is existence-independent, unlike the bare dir path.
+
+    Guards against the inaccurate "gitignored" claim by making it verifiable.
+    """
+    probe = "backend/sandbox/__artifact_probe__"
+    result = subprocess.run(
+        ["git", "-C", str(REPO_ROOT), "check-ignore", "-q", probe],
+        capture_output=True,
+    )
+    if result.returncode == 128:  # not a git checkout (e.g. packaged install)
+        pytest.skip("not inside a git working tree")
+    assert result.returncode == 0, "backend/sandbox/* should be gitignored (see backend/.gitignore '/sandbox/')"

backend/uv.lock

@@ -820,6 +820,7 @@ dependencies = [
     { name = "agent-sandbox" },
     { name = "aiosqlite" },
     { name = "alembic" },
+    { name = "cryptography" },
     { name = "ddgs" },
     { name = "dotenv" },
     { name = "duckdb" },
@@ -871,6 +872,7 @@ requires-dist = [
     { name = "aiosqlite", specifier = ">=0.19" },
     { name = "alembic", specifier = ">=1.13" },
     { name = "asyncpg", marker = "extra == 'postgres'", specifier = ">=0.29" },
+    { name = "cryptography", specifier = ">=43.0.0" },
     { name = "ddgs", specifier = ">=9.10.0" },
     { name = "dotenv", specifier = ">=0.9.9" },
     { name = "duckdb", specifier = ">=1.4.4" },

config.example.yaml

@@ -15,7 +15,7 @@
 # ============================================================================
 # Bump this number when the config schema changes.
 # Run `make config-upgrade` to merge new fields into your local config.yaml.
-config_version: 11
+config_version: 12
 
 # ============================================================================
 # Logging
@@ -274,6 +274,32 @@ models:
   #       thinking:
   #         type: disabled
 
+  # Example: StepFun (阶跃星辰) reasoning models
+  # StepFun provides OpenAI-compatible API with reasoning models.
+  # With reasoning_format: deepseek-style, the API returns reasoning_content
+  # (same field as DeepSeek), which must be replayed on historical assistant
+  # messages in multi-turn tool-call conversations.
+  # Use PatchedChatStepFun instead of plain ChatOpenAI.
+  # Docs: https://platform.stepfun.com/docs/api-reference/chat-completions
+  # - name: step-3.7-flash
+  #   display_name: Step 3.7 Flash
+  #   use: deerflow.models.patched_stepfun:PatchedChatStepFun
+  #   model: step-3.7-flash
+  #   api_key: $STEPFUN_API_KEY
+  #   base_url: https://api.stepfun.com/v1
+  #   request_timeout: 600.0
+  #   max_retries: 2
+  #   max_tokens: 4096
+  #   supports_thinking: true
+  #   supports_reasoning_effort: true
+  #   supports_vision: true
+  #   when_thinking_enabled:
+  #     extra_body:
+  #       reasoning_format: deepseek-style
+  #   when_thinking_disabled:
+  #     extra_body:
+  #       reasoning_format: deepseek-style
+
   # Example: MiniMax (OpenAI-compatible) - International Edition
   # MiniMax provides high-performance models with 512K context window and 128K max output
   # Docs: https://platform.minimax.io/docs/api-reference/text-openai-api
@@ -537,6 +563,10 @@ tools:
     group: web
     use: deerflow.community.jina_ai.tools:web_fetch_tool
     timeout: 10
+    # Optional proxy for restricted networks / Docker / WSL.
+    # Use host.docker.internal instead of 127.0.0.1 when the proxy runs on the host.
+    # proxy: $HTTPS_PROXY
+    # trust_env: true
 
   # Web fetch tool (uses InfoQuest)
   # - name: web_fetch
@@ -738,8 +768,12 @@ sandbox:
   allow_host_bash: false
   # Optional: Mount additional host directories into the sandbox.
   # Each mount maps a host path to a virtual container path accessible by the agent.
+  # Note: with LocalSandboxProvider under `make up` (docker-compose), host_path is
+  # checked from inside the deer-flow-gateway container — you must also bind-mount
+  # the same directory into services.gateway.volumes in docker/docker-compose.yaml
+  # for this mount to take effect (see issue #3244).
   # mounts:
-  #   - host_path: /home/user/my-project    # Absolute path on the host machine
+  #   - host_path: /home/user/my-project    # Absolute path; see note above for Docker mode
   #     container_path: /mnt/my-project     # Virtual path inside the sandbox
   #     read_only: true                     # Whether the mount is read-only (default: false)
 
@@ -1097,6 +1131,33 @@ run_events:
   max_trace_content: 10240
   track_token_usage: true
 
+# ============================================================================
+# User-Owned IM Channel Connections
+# ============================================================================
+# Lets logged-in users connect their own Telegram, Slack, and Discord accounts
+# from the DeerFlow frontend while reusing the existing `channels` runtime
+# configuration below.
+#
+# Security notes:
+# - No public IP, OAuth callback URL, or provider webhook is required.
+# - Provider bot/app credentials stay under `channels.*`.
+# - `channel_connections` stores per-user bindings and one-time connect codes.
+# - Telegram uses a deep link when `bot_username` is configured.
+# - Slack and Discord use `/connect <code>` through the already-running bot.
+#
+# channel_connections:
+#   enabled: false
+#
+#   telegram:
+#     enabled: false
+#     bot_username: $TELEGRAM_BOT_USERNAME
+#
+#   slack:
+#     enabled: false
+#
+#   discord:
+#     enabled: false
+
 # ============================================================================
 # IM Channels Configuration
 # ============================================================================

docker/dev-entrypoint.sh

@@ -64,12 +64,14 @@ if [ -n "$EXTRAS_FLAGS" ]; then
     echo "[startup] uv extras:$EXTRAS_FLAGS"
 fi
 
-# Keep runtime-owned files out of uvicorn's reload watcher. The directory must
+# Keep runtime-owned files out of uvicorn's reload watcher. Each excluded path
-# exist before uvicorn starts so watchfiles treats it as an excluded directory,
+# must exist before uvicorn starts so watchfiles treats it as an excluded
-# not as a plain glob pattern.
+# directory, not as a plain glob pattern — on Python 3.12, globbing an absolute
+# pattern raises NotImplementedError and crashes startup (#3459 / #3454). That
+# means `sandbox` must be created here too, not just `.deer-flow`.
 : "${DEER_FLOW_HOME:=/app/backend/.deer-flow}"
 export DEER_FLOW_HOME
-mkdir -p "$DEER_FLOW_HOME" /app/backend/.deer-flow
+mkdir -p "$DEER_FLOW_HOME" /app/backend/.deer-flow /app/backend/sandbox
 
 # ── Sync dependencies (with self-heal) ──────────────────────────────────────
 

docker/docker-compose-dev.yaml

@@ -172,6 +172,10 @@ services:
       - DEER_FLOW_HOST_BASE_DIR=${DEER_FLOW_ROOT}/backend/.deer-flow
       - DEER_FLOW_HOST_SKILLS_PATH=${DEER_FLOW_ROOT}/skills
       - DEER_FLOW_SANDBOX_HOST=host.docker.internal
+      # Proxy values (HTTP_PROXY/HTTPS_PROXY/ALL_PROXY) are inherited from ../.env via env_file.
+      # Only NO_PROXY is declared here so internal service hostnames are always exempt from the proxy.
+      - NO_PROXY=${NO_PROXY:-}${NO_PROXY:+,}localhost,127.0.0.1,::1,gateway,frontend,nginx,provisioner,host.docker.internal
+      - no_proxy=${no_proxy:-}${no_proxy:+,}localhost,127.0.0.1,::1,gateway,frontend,nginx,provisioner,host.docker.internal
     env_file:
       - ../.env
     extra_hosts:

docker/docker-compose.yaml

@@ -72,7 +72,13 @@ services:
         UV_INDEX_URL: ${UV_INDEX_URL:-https://pypi.org/simple}
         UV_EXTRAS: ${UV_EXTRAS:-}
     container_name: deer-flow-gateway
-    command: sh -c "cd backend && PYTHONPATH=. uv run uvicorn app.gateway.app:app --host 0.0.0.0 --port 8001 --workers ${GATEWAY_WORKERS:-4}"
+    # Gateway hosts the agent runtime with in-process RunManager + StreamBridge
+    # singletons -- run state lives in this worker's memory. Default to a single
+    # worker: with >1 worker and no nginx sticky sessions, run cancel, SSE
+    # reconnect, request dedup, and per-worker IM channel services all break
+    # across workers until a shared (e.g. redis) stream bridge lands, which is
+    # not yet implemented. Override GATEWAY_WORKERS only once that is in place.
+    command: sh -c "cd backend && PYTHONPATH=. uv run uvicorn app.gateway.app:app --host 0.0.0.0 --port 8001 --workers ${GATEWAY_WORKERS:-1}"
     volumes:
       - ${DEER_FLOW_CONFIG_PATH}:/app/backend/config.yaml:ro
       - ${DEER_FLOW_EXTENSIONS_CONFIG_PATH}:/app/backend/extensions_config.json:ro
@@ -107,6 +113,10 @@ services:
       - DEER_FLOW_HOST_BASE_DIR=${DEER_FLOW_HOME}
       - DEER_FLOW_HOST_SKILLS_PATH=${DEER_FLOW_REPO_ROOT}/skills
       - DEER_FLOW_SANDBOX_HOST=host.docker.internal
+      # Proxy values (HTTP_PROXY/HTTPS_PROXY/ALL_PROXY) are inherited from ../.env via env_file.
+      # Only NO_PROXY is declared here so internal service hostnames are always exempt from the proxy.
+      - NO_PROXY=${NO_PROXY:-}${NO_PROXY:+,}localhost,127.0.0.1,::1,gateway,frontend,nginx,provisioner,host.docker.internal
+      - no_proxy=${no_proxy:-}${no_proxy:+,}localhost,127.0.0.1,::1,gateway,frontend,nginx,provisioner,host.docker.internal
     env_file:
       - ../.env
     extra_hosts:

frontend/eslint.config.js

@@ -9,6 +9,8 @@ export default tseslint.config(
   {
     ignores: [
       ".next",
+      "playwright-report",
+      "test-results",
       "src/components/ui/**",
       "src/components/ai-elements/**",
       "*.js",

frontend/playwright.real-backend.config.ts

@@ -7,8 +7,9 @@ import { defineConfig, devices } from "@playwright/test";
  * so the mock-based suite is untouched.
  *
  * Two webServers are started: the replay gateway (:8011) and the frontend
- * (:3000, pointed at the gateway). Auth uses a throwaway test account the spec
+ * (:3000, pointed at the gateway). Auth-disabled mode is enabled on both
- * registers at runtime — no secrets.
+ * servers so the no-cookie e2e contract is covered; specs that need session
+ * cookies still register a throwaway test account at runtime.
  */
 export default defineConfig({
   testDir: "./tests/e2e-real-backend",
@@ -38,7 +39,10 @@ export default defineConfig({
       // Mount the test-only run/message seeder used by multi-run-order.spec.ts
       // (#3352). The endpoint exists only on this replay gateway, never in the
       // production app.
-      env: { DEERFLOW_ENABLE_TEST_SEED: "1" },
+      env: {
+        DEERFLOW_ENABLE_TEST_SEED: "1",
+        DEER_FLOW_AUTH_DISABLED: "1",
+      },
     },
     {
       command: "pnpm build && pnpm start",

frontend/src/components/ai-elements/message.tsx

@@ -18,7 +18,8 @@ import {
 } from "lucide-react";
 import type { ComponentProps, HTMLAttributes, ReactElement } from "react";
 import { createContext, memo, useContext, useEffect, useState } from "react";
-import { Streamdown } from "streamdown";
+
+import { ClipboardSafeStreamdown } from "./streamdown";
 
 export type MessageProps = HTMLAttributes<HTMLDivElement> & {
   from: UIMessage["role"];
@@ -302,11 +303,13 @@ export const MessageBranchPage = ({
   );
 };
 
-export type MessageResponseProps = ComponentProps<typeof Streamdown>;
+export type MessageResponseProps = ComponentProps<
+  typeof ClipboardSafeStreamdown
+>;
 
 export const MessageResponse = memo(
   ({ className, ...props }: MessageResponseProps) => (
-    <Streamdown
+    <ClipboardSafeStreamdown
       className={cn(
         "size-full [&>*:first-child]:mt-0 [&>*:last-child]:mb-0",
         className,

frontend/src/components/ai-elements/prompt-input.tsx

@@ -881,6 +881,7 @@ export type PromptInputTextareaProps = ComponentProps<
 
 export const PromptInputTextarea = ({
   onChange,
+  onKeyDown,
   className,
   placeholder = "What would you like to know?",
   ...props
@@ -891,6 +892,10 @@ export const PromptInputTextarea = ({
   const [isComposing, setIsComposing] = useState(false);
 
   const handleKeyDown: KeyboardEventHandler<HTMLTextAreaElement> = (e) => {
+    onKeyDown?.(e);
+    if (e.defaultPrevented) {
+      return;
+    }
     if (e.key === "Enter") {
       if (isIMEComposing(e, isComposing)) {
         return;

frontend/src/components/ai-elements/reasoning.tsx

@@ -10,9 +10,9 @@ import { cn } from "@/lib/utils";
 import { BrainIcon, ChevronDownIcon } from "lucide-react";
 import type { ComponentProps, ReactNode } from "react";
 import { createContext, memo, useContext, useEffect, useState } from "react";
-import { Streamdown } from "streamdown";
 import { reasoningPlugins } from "@/core/streamdown/plugins";
 import { Shimmer } from "./shimmer";
+import { ClipboardSafeStreamdown } from "./streamdown";
 
 type ReasoningContextValue = {
   isStreaming: boolean;
@@ -178,7 +178,9 @@ export const ReasoningContent = memo(
       )}
       {...props}
     >
-      <Streamdown {...reasoningPlugins}>{children}</Streamdown>
+      <ClipboardSafeStreamdown {...reasoningPlugins}>
+        {children}
+      </ClipboardSafeStreamdown>
     </CollapsibleContent>
   ),
 );

frontend/src/components/ai-elements/streamdown.tsx

@@ -0,0 +1,17 @@
+"use client";
+
+import { type ComponentProps } from "react";
+import { Streamdown } from "streamdown";
+
+import { installClipboardFallback } from "@/core/clipboard";
+
+export type ClipboardSafeStreamdownProps = ComponentProps<typeof Streamdown>;
+
+// Only patch browser globals in client context; skip during SSR
+if (typeof document !== "undefined") {
+  installClipboardFallback();
+}
+
+export function ClipboardSafeStreamdown(props: ClipboardSafeStreamdownProps) {
+  return <Streamdown {...props} />;
+}

frontend/src/components/workspace/artifacts/artifact-file-detail.tsx

@@ -10,7 +10,6 @@ import {
 } from "lucide-react";
 import { useCallback, useEffect, useMemo, useRef, useState } from "react";
 import { toast } from "sonner";
-import { Streamdown } from "streamdown";
 
 import {
   Artifact,
@@ -20,6 +19,7 @@ import {
   ArtifactHeader,
   ArtifactTitle,
 } from "@/components/ai-elements/artifact";
+import { ClipboardSafeStreamdown } from "@/components/ai-elements/streamdown";
 import { Select, SelectItem } from "@/components/ui/select";
 import {
   SelectContent,
@@ -400,13 +400,13 @@ export function ArtifactFilePreview({
   if (language === "markdown") {
     return (
       <div className="size-full px-4">
-        <Streamdown
+        <ClipboardSafeStreamdown
           className="size-full"
           {...streamdownPlugins}
           components={{ a: ArtifactLink }}
         >
           {content ?? ""}
-        </Streamdown>
+        </ClipboardSafeStreamdown>
       </div>
     );
   }