refactor(tools): move session tools to the agent_tools registry (#4454)

Moves create_session, list_sessions, send_to_session and manage_session out of
ai_interaction.py into src/agent_tools/session_tools.py (the do_ prefix
dropped) and registers them in TOOL_HANDLERS, so dispatch flows through the
registry instead of the dispatch_ai_tool elif in tool_execution.py. Same
pattern as the model-interaction move.

The bodies move verbatim; each fetches the runtime-set session manager via a
get_session_manager() shim, and reuses _resolve_model / AI_CHAT_TIMEOUT from
ai_interaction. manage_session's internal 'list' alias is repointed from the
old do_list_sessions to the moved list_sessions. stream_ai_tool (dead, no
callers) and do_pipeline stay put. dispatch_ai_tool loses its four now-unused
branches.

Tests: test_session_tools_registry covers registration, owner threading, the
manage_session->list_sessions delegation, graceful no-manager handling, and
registry dispatch. Verified end-to-end against a live SessionManager.

.dockerignore

@@ -15,6 +15,10 @@ build/
 # at runtime — never baked into the image. Mirrored in .gitignore.
 secrets.env
 secrets.env.*
+secrets.env~
+.secrets.env.swp
+.secrets.env.swo
+**/#secrets.env#
 !secrets.env.example
 /data/
 /logs/

.github/workflows/ci.yml

@@ -19,7 +19,7 @@ jobs:
     name: Python syntax (compileall)
     runs-on: ubuntu-latest
     steps:
-      - uses: actions/checkout@df4cb1c069e1874edd31b4311f1884172cec0e10  # v6.0.3
+      - uses: actions/checkout@9c091bb21b7c1c1d1991bb908d89e4e9dddfe3e0  # v7.0.0
         with:
           persist-credentials: false
       - uses: actions/setup-python@a309ff8b426b58ec0e2a45f0f869d46889d02405  # v6.2.0
@@ -32,7 +32,7 @@ jobs:
     name: JS syntax (node --check)
     runs-on: ubuntu-latest
     steps:
-      - uses: actions/checkout@df4cb1c069e1874edd31b4311f1884172cec0e10  # v6.0.3
+      - uses: actions/checkout@9c091bb21b7c1c1d1991bb908d89e4e9dddfe3e0  # v7.0.0
         with:
           persist-credentials: false
       - uses: actions/setup-node@48b55a011bda9f5d6aeb4c2d9c7362e8dae4041e  # v6.4.0
@@ -54,7 +54,7 @@ jobs:
     # ROADMAP "fresh install smoke tests" item; make this required once green.
     continue-on-error: true
     steps:
-      - uses: actions/checkout@df4cb1c069e1874edd31b4311f1884172cec0e10  # v6.0.3
+      - uses: actions/checkout@9c091bb21b7c1c1d1991bb908d89e4e9dddfe3e0  # v7.0.0
         with:
           fetch-depth: 0
           persist-credentials: false

.github/workflows/container-scan.yml

@@ -37,7 +37,7 @@ jobs:
       contents: read
     steps:
       - name: Checkout repository
-        uses: actions/checkout@df4cb1c069e1874edd31b4311f1884172cec0e10  # v6.0.3
+        uses: actions/checkout@9c091bb21b7c1c1d1991bb908d89e4e9dddfe3e0  # v7.0.0
         with:
           persist-credentials: false
 

.github/workflows/container-trivy.yml

@@ -52,7 +52,7 @@ jobs:
       contents: read
     steps:
       - name: Checkout repository
-        uses: actions/checkout@df4cb1c069e1874edd31b4311f1884172cec0e10  # v6.0.3
+        uses: actions/checkout@9c091bb21b7c1c1d1991bb908d89e4e9dddfe3e0  # v7.0.0
         with:
           persist-credentials: false
 
@@ -93,7 +93,7 @@ jobs:
       security-events: write  # upload SARIF to the Security tab
     steps:
       - name: Checkout repository
-        uses: actions/checkout@df4cb1c069e1874edd31b4311f1884172cec0e10  # v6.0.3
+        uses: actions/checkout@9c091bb21b7c1c1d1991bb908d89e4e9dddfe3e0  # v7.0.0
         with:
           persist-credentials: false
 

.github/workflows/dependency-review.yml

@@ -36,7 +36,7 @@ jobs:
       contents: read
     steps:
       - name: Checkout repository
-        uses: actions/checkout@df4cb1c069e1874edd31b4311f1884172cec0e10  # v6.0.3
+        uses: actions/checkout@9c091bb21b7c1c1d1991bb908d89e4e9dddfe3e0  # v7.0.0
         with:
           persist-credentials: false
 
@@ -55,7 +55,7 @@ jobs:
       contents: read
     steps:
       - name: Checkout repository
-        uses: actions/checkout@df4cb1c069e1874edd31b4311f1884172cec0e10  # v6.0.3
+        uses: actions/checkout@9c091bb21b7c1c1d1991bb908d89e4e9dddfe3e0  # v7.0.0
         with:
           persist-credentials: false
 

.github/workflows/docker-publish.yml

@@ -45,7 +45,7 @@ jobs:
             arch: arm64
             runner: ubuntu-24.04-arm
     steps:
-      - uses: actions/checkout@df4cb1c069e1874edd31b4311f1884172cec0e10  # v6.0.3
+      - uses: actions/checkout@9c091bb21b7c1c1d1991bb908d89e4e9dddfe3e0  # v7.0.0
         with:
           persist-credentials: false
       - name: Set up Buildx
@@ -86,7 +86,7 @@ jobs:
       contents: read
       packages: write
     steps:
-      - uses: actions/checkout@df4cb1c069e1874edd31b4311f1884172cec0e10  # v6.0.3
+      - uses: actions/checkout@9c091bb21b7c1c1d1991bb908d89e4e9dddfe3e0  # v7.0.0
         with:
           persist-credentials: false
       - name: Read APP_VERSION + short sha

.github/workflows/issue-description-check.yml

@@ -14,7 +14,7 @@ jobs:
     # Skip bots (Dependabot, release-drafter, etc.)
     if: ${{ github.event.issue.user.type != 'Bot' }}
     steps:
-      - uses: actions/checkout@df4cb1c069e1874edd31b4311f1884172cec0e10  # v6.0.3
+      - uses: actions/checkout@9c091bb21b7c1c1d1991bb908d89e4e9dddfe3e0  # v7.0.0
         with:
           sparse-checkout: .github/scripts
           persist-credentials: false

.github/workflows/pr-description-check.yml

@@ -23,7 +23,7 @@ jobs:
     # Skip bots: they open PRs programmatically and have their own process.
     if: github.event.pull_request.user.type != 'Bot'
     steps:
-      - uses: actions/checkout@df4cb1c069e1874edd31b4311f1884172cec0e10  # v6.0.3
+      - uses: actions/checkout@9c091bb21b7c1c1d1991bb908d89e4e9dddfe3e0  # v7.0.0
         with:
           ref: ${{ github.base_ref }}
           sparse-checkout: .github/scripts

.github/workflows/secret-scan.yml

@@ -35,7 +35,7 @@ jobs:
       contents: read
     steps:
       - name: Checkout repository
-        uses: actions/checkout@df4cb1c069e1874edd31b4311f1884172cec0e10  # v6.0.3
+        uses: actions/checkout@9c091bb21b7c1c1d1991bb908d89e4e9dddfe3e0  # v7.0.0
         with:
           # Full history so a secret committed in an earlier commit (and later
           # deleted) is still caught -- deletion does not remove it from Git.

.github/workflows/workflow-security.yml

@@ -36,7 +36,7 @@ jobs:
       contents: read
     steps:
       - name: Checkout repository
-        uses: actions/checkout@df4cb1c069e1874edd31b4311f1884172cec0e10  # v6.0.3
+        uses: actions/checkout@9c091bb21b7c1c1d1991bb908d89e4e9dddfe3e0  # v7.0.0
         with:
           persist-credentials: false
 
@@ -61,7 +61,7 @@ jobs:
       contents: read
     steps:
       - name: Checkout repository
-        uses: actions/checkout@df4cb1c069e1874edd31b4311f1884172cec0e10  # v6.0.3
+        uses: actions/checkout@9c091bb21b7c1c1d1991bb908d89e4e9dddfe3e0  # v7.0.0
         with:
           persist-credentials: false
 

CONTRIBUTING.md

@@ -37,7 +37,7 @@ Manual development uses Python 3.11+:
 python3 -m venv venv
 source venv/bin/activate
 pip install -r requirements.txt
-python -m uvicorn app:app --host 0.0.0.0 --port 7000
+python -m uvicorn app:app --host 127.0.0.1 --port 7000
 ```
 
 Windows is not actively tested. Docker on Linux or a Linux/macOS manual install is the safer path for now.

Odysseus.spec

@@ -0,0 +1,45 @@
+# -*- mode: python ; coding: utf-8 -*-
+
+
+a = Analysis(
+    ['launcher.py'],
+    pathex=[],
+    binaries=[],
+    datas=[('static', 'static'), ('scripts', 'scripts'), ('mcp_servers', 'mcp_servers'), ('services/hwfit/data', 'services/hwfit/data'), ('config', 'config'), ('.env.example', '.env.example')],
+    hiddenimports=[],
+    hookspath=[],
+    hooksconfig={},
+    runtime_hooks=[],
+    excludes=[],
+    noarchive=False,
+    optimize=0,
+)
+pyz = PYZ(a.pure)
+
+exe = EXE(
+    pyz,
+    a.scripts,
+    [],
+    exclude_binaries=True,
+    name='Odysseus',
+    debug=False,
+    bootloader_ignore_signals=False,
+    strip=False,
+    upx=True,
+    console=False,
+    disable_windowed_traceback=False,
+    argv_emulation=False,
+    target_arch=None,
+    codesign_identity=None,
+    entitlements_file=None,
+    icon=['static\\icon.ico'],
+)
+coll = COLLECT(
+    exe,
+    a.binaries,
+    a.datas,
+    strip=False,
+    upx=True,
+    upx_exclude=[],
+    name='Odysseus',
+)

app.py

@@ -1,6 +1,7 @@
 # app.py — slim orchestrator
 import mimetypes
 import os
+import sys
 
 
 def register_static_mime_types() -> None:
@@ -38,7 +39,7 @@ load_dotenv(encoding="utf-8-sig")
 import asyncio
 import logging
 import secrets
-from datetime import datetime
+from datetime import datetime, timezone
 from typing import Dict
 
 from contextlib import asynccontextmanager
@@ -113,12 +114,13 @@ app = FastAPI(
 )
 
 # ========= CORS =========
+CORS_ALLOW_METHODS = ["GET", "POST", "PUT", "PATCH", "DELETE"]
 allowed_origins = os.getenv("ALLOWED_ORIGINS", "http://localhost,http://127.0.0.1").split(",")
 app.add_middleware(
     CORSMiddleware,
     allow_origins=allowed_origins,
     allow_credentials=True,
-    allow_methods=["GET", "POST", "PUT", "DELETE"],
+    allow_methods=CORS_ALLOW_METHODS,
     allow_headers=[
         "Accept",
         "Authorization",
@@ -316,7 +318,7 @@ if AUTH_ENABLED:
             # (no admin cookie available in that context). Restricted to
             # loopback clients + matching token to keep it locked down.
             try:
-                from core.middleware import INTERNAL_TOOL_HEADER, INTERNAL_TOOL_TOKEN as _ITT
+                from core.middleware import INTERNAL_TOOL_HEADER, INTERNAL_TOOL_TOKEN as _ITT, INTERNAL_TOOL_USER
                 _hdr = request.headers.get(INTERNAL_TOOL_HEADER)
                 if _hdr and secrets.compare_digest(_hdr, _ITT) and _is_trusted_loopback(request):
                     # Impersonation: when the agent's loopback call sets
@@ -328,7 +330,7 @@ if AUTH_ENABLED:
                     if _impersonate and _impersonate in getattr(_auth_mgr, "users", {}):
                         request.state.current_user = _impersonate
                     else:
-                        request.state.current_user = "internal-tool"
+                        request.state.current_user = INTERNAL_TOOL_USER
                     request.state.api_token = False
                     return await call_next(request)
             except Exception as _e:
@@ -437,7 +439,7 @@ class _RevalidatingStatic(StaticFiles):
         return resp
 
 
-app.mount("/static", _RevalidatingStatic(directory="static"), name="static")
+app.mount("/static", _RevalidatingStatic(directory=STATIC_DIR), name="static")
 
 # ========= GENERATED IMAGES =========
 @app.get("/api/generated-image/{filename}")
@@ -527,6 +529,7 @@ memory_vector     = components.get("memory_vector")
 upload_handler    = components["upload_handler"]
 app.state.upload_handler = upload_handler
 personal_docs_mgr = components["personal_docs_manager"]
+app.state.personal_docs_manager = personal_docs_mgr
 api_key_manager   = components["api_key_manager"]
 preset_manager    = components["preset_manager"]
 chat_processor    = components["chat_processor"]
@@ -860,7 +863,7 @@ async def get_version():
 
 @app.get("/api/health")
 async def health_check() -> Dict[str, str]:
-    return {"status": "healthy", "timestamp": datetime.utcnow().isoformat()}
+    return {"status": "healthy", "timestamp": datetime.now(timezone.utc).isoformat()}
 
 @app.get("/api/ready")
 async def readiness_check() -> JSONResponse:
@@ -1170,3 +1173,12 @@ async def _shutdown_event():
     except Exception as e:
         logger.warning(f"MCP shutdown error: {e}")
     logger.info("Application shutdown complete")
+
+
+if __name__ == "__main__":
+    import uvicorn
+
+    bind_host = os.getenv("APP_BIND", "127.0.0.1")
+    bind_port = int(os.getenv("APP_PORT", "7000"))
+
+    uvicorn.run(app, host=bind_host, port=bind_port, log_level="info")

build-windows-portable.ps1

@@ -0,0 +1,72 @@
+#Requires -Version 5.1
+<#
+  Build a portable Windows distribution for Odysseus.
+
+  Output layout:
+    dist\Odysseus\Odysseus.exe
+    dist\Odysseus\static\...
+    dist\Odysseus\scripts\...
+    dist\Odysseus\mcp_servers\...
+    dist\Odysseus\services\hwfit\data\...
+
+  The app then keeps using its normal filesystem layout when frozen.
+
+  Usage:
+    powershell -ExecutionPolicy Bypass -File .\build-windows-portable.ps1
+#>
+
+$ErrorActionPreference = "Stop"
+Set-Location -Path $PSScriptRoot
+
+function Write-Step($msg) { Write-Host ""; Write-Host ("==> " + $msg) -ForegroundColor Cyan }
+function Fail($msg) {
+    Write-Host ""
+    Write-Host ("ERROR: " + $msg) -ForegroundColor Red
+    exit 1
+}
+
+Write-Step "Checking for Python"
+$pyExe = $null
+if (Test-Path ".\.venv\Scripts\python.exe") {
+    $pyExe = (Resolve-Path ".\.venv\Scripts\python.exe").Path
+} else {
+    foreach ($c in @("py", "python")) {
+        $cmd = Get-Command $c -ErrorAction SilentlyContinue
+        if ($cmd) { $pyExe = $cmd.Source; break }
+    }
+    if ($pyExe -like "*WindowsApps*python.exe") {
+        $pyCmd = Get-Command py -ErrorAction SilentlyContinue
+        if ($pyCmd) {
+            $pyExe = $pyCmd.Source
+        }
+    }
+}
+if (-not $pyExe) {
+    Fail "Python not found on PATH. Install Python 3.11+ first."
+}
+Write-Host ("Using Python: " + $pyExe)
+
+Write-Step "Installing build dependencies"
+& $pyExe -m pip install --upgrade pip --quiet
+& $pyExe -m pip install -r requirements.txt pyinstaller pystray Pillow
+if ($LASTEXITCODE -ne 0) { Fail "Dependency install failed." }
+
+Write-Step "Building portable exe bundle"
+Remove-Item -Recurse -Force build, dist -ErrorAction SilentlyContinue
+
+$dataArgs = @(
+    "--add-data", "static;static",
+    "--add-data", "scripts;scripts",
+    "--add-data", "mcp_servers;mcp_servers",
+    "--add-data", "services/hwfit/data;services/hwfit/data",
+    "--add-data", "config;config",
+    "--add-data", ".env.example;.env.example"
+)
+
+& $pyExe -m PyInstaller --noconfirm --clean --onedir --noconsole --icon=static/icon.ico --name Odysseus @dataArgs launcher.py
+if ($LASTEXITCODE -ne 0) { Fail "PyInstaller build failed." }
+
+Write-Host ""
+Write-Host "Build complete." -ForegroundColor Green
+Write-Host "Portable app folder: $PSScriptRoot\dist\Odysseus" -ForegroundColor Green
+Write-Host "Distribute the whole folder (or zip it) so static assets and scripts stay with the exe." -ForegroundColor Green

core/auth.py

@@ -20,6 +20,7 @@ logger = logging.getLogger(__name__)
 
 
 from core.atomic_io import atomic_write_json as _atomic_write_json  # noqa: E402
+from core.middleware import INTERNAL_TOOL_USER  # noqa: E402
 
 DEFAULT_PRIVILEGES = {
     "can_use_agent": True,
@@ -47,7 +48,7 @@ ADMIN_PRIVILEGES["allowed_models_restricted"] = False
 # backwards for this sentinel.
 ADMIN_PRIVILEGES["block_all_models"] = False
 
-from src.constants import AUTH_FILE
+from src.constants import AUTH_FILE, PASSWORD_MIN_LENGTH
 DEFAULT_AUTH_PATH = AUTH_FILE
 TOKEN_TTL = 60 * 60 * 24 * 7  # 7 days
 
@@ -65,7 +66,7 @@ TOKEN_TTL = 60 * 60 * 24 * 7  # 7 days
 # of those names would be denied an assistant and inconsistently owner-scoped.
 # Refuse to create or rename into any of them so the sentinels can't be
 # impersonated. (Keep this in sync with that synthetic-owner set.)
-RESERVED_USERNAMES = frozenset({"internal-tool", "api", "demo", "system"})
+RESERVED_USERNAMES = frozenset({INTERNAL_TOOL_USER, "api", "demo", "system"})
 
 
 def normalize_known_username(users: Dict[str, Any], username: str | None) -> Optional[str]:
@@ -243,6 +244,15 @@ class AuthManager:
     def is_configured(self) -> bool:
         return len(self.users) > 0
 
+    def policy(self) -> dict:
+        """Return public auth policy constants for the frontend."""
+        return {
+            "password_min_length": PASSWORD_MIN_LENGTH,
+            "reserved_usernames": sorted(RESERVED_USERNAMES),
+            "signup_enabled": self.signup_enabled,
+            "session_days": TOKEN_TTL // 86400,
+        }
+
     # ------------------------------------------------------------------
     # Account management
     # ------------------------------------------------------------------
@@ -573,11 +583,15 @@ class AuthManager:
             return None
         return self.create_session_trusted(username)
 
-    def create_session_trusted(self, username: str) -> str:
+    def create_session_trusted(self, username: str) -> Optional[str]:
         """Issue a session token for an already-verified user.
         Call only after verify_password (and TOTP if enabled) have passed."""
         username = username.strip().lower()
         token = secrets.token_hex(32)
+        with self._config_lock:
+            if username not in self.users:
+                logger.warning("Refused to issue session for missing user '%s'", username)
+                return None
             with self._sessions_lock:
                 self._sessions[token] = {
                     "username": username,

core/middleware.py

@@ -15,6 +15,8 @@ from starlette.responses import Response
 # same value from this module. Never persisted or exposed externally.
 INTERNAL_TOOL_TOKEN = os.environ.get("ODYSSEUS_INTERNAL_TOKEN") or secrets.token_hex(32)
 INTERNAL_TOOL_HEADER = "X-Odysseus-Internal-Token"
+# Pseudo-username on in-process tool-loopback requests; require_admin trusts it and it is reserved.
+INTERNAL_TOOL_USER = "internal-tool"
 
 
 def is_cors_preflight(method: str, headers) -> bool:
@@ -39,7 +41,7 @@ def require_admin(request: Request):
         hdr = request.headers.get(INTERNAL_TOOL_HEADER)
         if hdr and secrets.compare_digest(hdr, INTERNAL_TOOL_TOKEN):
             return
-        if getattr(request.state, "current_user", None) == "internal-tool":
+        if getattr(request.state, "current_user", None) == INTERNAL_TOOL_USER:
             return
     except Exception:
         pass
@@ -65,10 +67,9 @@ class SecurityHeadersMiddleware(BaseHTTPMiddleware):
         response = await call_next(request)
         path = request.url.path
 
-        # Tool render endpoints are served inside iframes — allow framing by self
+        # Tool render endpoints
         is_tool_render = path.startswith("/api/tools/") and path.endswith("/render")
-        # PDF previews are embedded by the in-app document library. Keep the
+        # Document library PDF preview endpoint
-        # exception route-scoped so normal app pages remain unframeable.
         is_document_pdf_preview = path.startswith("/api/document/") and path.endswith("/render-pdf")
         # Visual report pages are self-contained HTML — need inline scripts + external images
         is_report = path.startswith("/api/research/report/")
@@ -95,9 +96,7 @@ class SecurityHeadersMiddleware(BaseHTTPMiddleware):
                 "frame-ancestors 'none'"
             )
         elif is_tool_render:
-            # Tool iframe content: skip all framing headers — the iframe's
+            # Skip framing headers for tools.
-            # sandbox="allow-scripts" attribute provides isolation.
-            # Don't overwrite the route's own restrictive CSP either.
             pass
         elif is_document_pdf_preview:
             response.headers["X-Frame-Options"] = "SAMEORIGIN"

docker-compose.gpu-amd.yml

@@ -60,6 +60,13 @@ services:
       - ODYSSEUS_INPROCESS_TASKS=${ODYSSEUS_INPROCESS_TASKS:-1}
       - ODYSSEUS_SCRIPT_HOST=${ODYSSEUS_SCRIPT_HOST:-localhost}
       - ODYSSEUS_CHAT_UPLOAD_MAX_BYTES=${ODYSSEUS_CHAT_UPLOAD_MAX_BYTES:-10485760}
+      - ODYSSEUS_GALLERY_UPLOAD_MAX_BYTES=${ODYSSEUS_GALLERY_UPLOAD_MAX_BYTES:-104857600}
+      - ODYSSEUS_GALLERY_TRANSFORM_UPLOAD_MAX_BYTES=${ODYSSEUS_GALLERY_TRANSFORM_UPLOAD_MAX_BYTES:-26214400}
+      - ODYSSEUS_MEMORY_IMPORT_MAX_BYTES=${ODYSSEUS_MEMORY_IMPORT_MAX_BYTES:-10485760}
+      - ODYSSEUS_PERSONAL_UPLOAD_MAX_BYTES=${ODYSSEUS_PERSONAL_UPLOAD_MAX_BYTES:-26214400}
+      - ODYSSEUS_EMAIL_COMPOSE_UPLOAD_MAX_BYTES=${ODYSSEUS_EMAIL_COMPOSE_UPLOAD_MAX_BYTES:-26214400}
+      - ODYSSEUS_STT_MAX_AUDIO_BYTES=${ODYSSEUS_STT_MAX_AUDIO_BYTES:-26214400}
+      - ODYSSEUS_ICS_MAX_BYTES=${ODYSSEUS_ICS_MAX_BYTES:-10485760}
       - DATA_BRAVE_API_KEY=${DATA_BRAVE_API_KEY:-}
       - GOOGLE_API_KEY=${GOOGLE_API_KEY:-}
       - GOOGLE_PSE_CX=${GOOGLE_PSE_CX:-}

docker-compose.gpu-nvidia.yml

@@ -59,6 +59,13 @@ services:
       - ODYSSEUS_INPROCESS_TASKS=${ODYSSEUS_INPROCESS_TASKS:-1}
       - ODYSSEUS_SCRIPT_HOST=${ODYSSEUS_SCRIPT_HOST:-localhost}
       - ODYSSEUS_CHAT_UPLOAD_MAX_BYTES=${ODYSSEUS_CHAT_UPLOAD_MAX_BYTES:-10485760}
+      - ODYSSEUS_GALLERY_UPLOAD_MAX_BYTES=${ODYSSEUS_GALLERY_UPLOAD_MAX_BYTES:-104857600}
+      - ODYSSEUS_GALLERY_TRANSFORM_UPLOAD_MAX_BYTES=${ODYSSEUS_GALLERY_TRANSFORM_UPLOAD_MAX_BYTES:-26214400}
+      - ODYSSEUS_MEMORY_IMPORT_MAX_BYTES=${ODYSSEUS_MEMORY_IMPORT_MAX_BYTES:-10485760}
+      - ODYSSEUS_PERSONAL_UPLOAD_MAX_BYTES=${ODYSSEUS_PERSONAL_UPLOAD_MAX_BYTES:-26214400}
+      - ODYSSEUS_EMAIL_COMPOSE_UPLOAD_MAX_BYTES=${ODYSSEUS_EMAIL_COMPOSE_UPLOAD_MAX_BYTES:-26214400}
+      - ODYSSEUS_STT_MAX_AUDIO_BYTES=${ODYSSEUS_STT_MAX_AUDIO_BYTES:-26214400}
+      - ODYSSEUS_ICS_MAX_BYTES=${ODYSSEUS_ICS_MAX_BYTES:-10485760}
       - DATA_BRAVE_API_KEY=${DATA_BRAVE_API_KEY:-}
       - GOOGLE_API_KEY=${GOOGLE_API_KEY:-}
       - GOOGLE_PSE_CX=${GOOGLE_PSE_CX:-}

docker-compose.yml

@@ -48,6 +48,13 @@ services:
       - ODYSSEUS_INPROCESS_TASKS=${ODYSSEUS_INPROCESS_TASKS:-1}
       - ODYSSEUS_SCRIPT_HOST=${ODYSSEUS_SCRIPT_HOST:-localhost}
       - ODYSSEUS_CHAT_UPLOAD_MAX_BYTES=${ODYSSEUS_CHAT_UPLOAD_MAX_BYTES:-10485760}
+      - ODYSSEUS_GALLERY_UPLOAD_MAX_BYTES=${ODYSSEUS_GALLERY_UPLOAD_MAX_BYTES:-104857600}
+      - ODYSSEUS_GALLERY_TRANSFORM_UPLOAD_MAX_BYTES=${ODYSSEUS_GALLERY_TRANSFORM_UPLOAD_MAX_BYTES:-26214400}
+      - ODYSSEUS_MEMORY_IMPORT_MAX_BYTES=${ODYSSEUS_MEMORY_IMPORT_MAX_BYTES:-10485760}
+      - ODYSSEUS_PERSONAL_UPLOAD_MAX_BYTES=${ODYSSEUS_PERSONAL_UPLOAD_MAX_BYTES:-26214400}
+      - ODYSSEUS_EMAIL_COMPOSE_UPLOAD_MAX_BYTES=${ODYSSEUS_EMAIL_COMPOSE_UPLOAD_MAX_BYTES:-26214400}
+      - ODYSSEUS_STT_MAX_AUDIO_BYTES=${ODYSSEUS_STT_MAX_AUDIO_BYTES:-26214400}
+      - ODYSSEUS_ICS_MAX_BYTES=${ODYSSEUS_ICS_MAX_BYTES:-10485760}
       - DATA_BRAVE_API_KEY=${DATA_BRAVE_API_KEY:-}
       - GOOGLE_API_KEY=${GOOGLE_API_KEY:-}
       - GOOGLE_PSE_CX=${GOOGLE_PSE_CX:-}

docker/entrypoint.sh

@@ -13,6 +13,8 @@ set -e
 
 PUID="${PUID:-1000}"
 PGID="${PGID:-1000}"
+GOSU_BIN="$(command -v gosu)"
+PYTHON_BIN="$(command -v python)"
 
 # Reuse an existing matching group/user if the host's UID/GID already
 # corresponds to one in /etc/passwd (e.g. when the image is rebuilt
@@ -24,26 +26,57 @@ if ! getent passwd "$PUID" >/dev/null 2>&1; then
     useradd -u "$PUID" -g "$PGID" -M -s /bin/sh -d /app odysseus
 fi
 
-# Repair ownership on every writable path the app touches at runtime.
+mount_root_for() {
-#
+    awk -v target="$1" '$5 == target { print $4; exit }' /proc/self/mountinfo 2>/dev/null || true
-# Bind-mounted dirs (/app/data, /app/logs) are the obvious ones, but
+}
-# the app ALSO writes inside the image's own source tree at runtime:
+
-#   - services/cache/{search,content}/*  (search cache LRU)
+is_broad_mount_root() {
-#   - services/search_analytics.json
+    case "$1" in
-#   - services/search_engine_error.log
+        /|/home|/srv|/var|/usr|/opt|/tmp|/mnt|/media)
-#   - services/tts cache, etc.
+            return 0
-# These dirs were created as root during `docker build`, so dropping
+            ;;
-# to PUID:PGID would otherwise crash on the first import that tries
+    esac
-# to mkdir them. Chown the whole /app tree — fast (<1s on this size)
+    return 1
-# and idempotent via the `-not -uid` filter so we only touch files
+}
-# that need fixing.
+
-for dir in /app /app/data /app/logs; do
+repair_tree_ownership() {
+    dir="$1"
     if [ -d "$dir" ]; then
-        # `find ... -not -uid` keeps this O(touched-files), not
+        find "$dir" -xdev -not -uid "$PUID" -print0 2>/dev/null \
-        # O(everything), so terabyte-sized maildirs don't slow startup.
-        find "$dir" -not -uid "$PUID" -print0 2>/dev/null \
             | xargs -0 -r chown "$PUID:$PGID" 2>/dev/null || true
     fi
+}
+
+repair_app_tree_ownership() {
+    if [ -d /app ]; then
+        find /app -xdev \
+            \( -path /app/data -o -path /app/logs -o -path /app/.ssh -o -path /app/.cache -o -path /app/.local \) -prune \
+            -o -not -uid "$PUID" -print0 2>/dev/null \
+            | xargs -0 -r chown "$PUID:$PGID" 2>/dev/null || true
+    fi
+}
+
+repair_bind_mount_ownership() {
+    dir="$1"
+    if [ ! -d "$dir" ]; then
+        return
+    fi
+
+    mount_root="$(mount_root_for "$dir")"
+    if is_broad_mount_root "$mount_root"; then
+        echo "Skipping recursive ownership repair for $dir because it maps to broad host path $mount_root" >&2
+        chown "$PUID:$PGID" "$dir" 2>/dev/null || true
+        return
+    fi
+
+    repair_tree_ownership "$dir"
+}
+
+# Repair image-owned writable paths without walking into bind-mounted host
+# trees, then repair the app-owned mount roots separately.
+repair_app_tree_ownership
+for dir in /app/data /app/logs /app/.ssh /app/.cache/huggingface /app/.local; do
+    repair_bind_mount_ownership "$dir"
 done
 
 # Cookbook installs vllm/etc. via `pip install --user`, which pulls
@@ -83,9 +116,9 @@ export PATH="/app/.local/bin:$PATH"
 # Run first-time setup as the app user so data/ files get the right ownership.
 # setup.py is idempotent — skips auth.json / .env if they already exist.
 # || true so a setup failure never prevents the container from starting.
-gosu "$PUID:$PGID" python /app/setup.py || true
+"$GOSU_BIN" "$PUID:$PGID" "$PYTHON_BIN" /app/setup.py || true
 
 # Drop root and run the actual app. `gosu` is preferred over `su` /
 # `sudo` because it cleans up the process tree (no extra shell layer)
 # so signals (SIGTERM from `docker stop`) reach uvicorn directly.
-exec gosu "$PUID:$PGID" "$@"
+exec "$GOSU_BIN" "$PUID:$PGID" "$@"

launch-windows.ps1

@@ -105,6 +105,14 @@ if (-not $pyExe) {
     }
 }
 
+if ($pyExe -like "*WindowsApps*python.exe") {
+    $pyCmd = Get-Command py -ErrorAction SilentlyContinue
+    if ($pyCmd) {
+        $pyExe = $pyCmd.Source
+        $pyArgs = @("-3.11")
+    }
+}
+
 if (-not $pyExe) {
     Fail "Couldn't find Python 3.11+ for Windows setup. Install Python 3.11+ (or open the Python launcher with 'py -3.11') from https://www.python.org/downloads/, then re-run this script."
 }

launcher.py

@@ -0,0 +1,142 @@
+# launcher.py
+"""Dedicated entrypoint for the standalone Windows portable launcher.
+
+Handles:
+- Immediate GUI splash screen creation using tkinter.
+- Suppressing console stream crashes in windowed GUI mode via NullWriter.
+- Spawning system tray icon via pystray and Pillow (lazy-loaded).
+- Auto-opening default browser pointing to the running backend.
+- Launching the FastAPI server (importing and running app.py).
+"""
+import os
+import sys
+import threading
+import time
+import webbrowser
+
+# Define a dummy NullWriter to suppress standard stream crashes (isatty etc.) in GUI mode
+class NullWriter:
+    def write(self, text):
+        pass
+    def flush(self):
+        pass
+    def isatty(self):
+        return False
+
+if sys.stdout is None:
+    sys.stdout = NullWriter()
+if sys.stderr is None:
+    sys.stderr = NullWriter()
+
+
+splash_root = None
+
+# If running from a frozen PyInstaller bundle, launch the splash screen IMMEDIATELY
+if getattr(sys, 'frozen', False):
+    import tkinter as tk
+
+    def show_splash_instantly():
+        global splash_root
+        try:
+            splash_root = tk.Tk()
+            splash_root.title("Odysseus")
+            splash_root.overrideredirect(True)
+            splash_root.configure(bg="#1a1c23")
+
+            # Accented borders
+            splash_root.config(highlightbackground="#e06c75", highlightcolor="#e06c75", highlightthickness=1)
+
+            w, h = 360, 160
+            ws = splash_root.winfo_screenwidth()
+            hs = splash_root.winfo_screenheight()
+            x = (ws - w) // 2
+            y = (hs - h) // 2
+            splash_root.geometry(f"{w}x{h}+{x}+{y}")
+
+            tk.Label(splash_root, text="⛵ Odysseus", font=("Segoe UI", 22, "bold"), bg="#1a1c23", fg="#e06c75").pack(pady=(22, 2))
+            tk.Label(splash_root, text="Launching background services...", font=("Segoe UI", 10), bg="#1a1c23", fg="#d1d4e0").pack(pady=2)
+            tk.Label(splash_root, text="Please wait, this will take a few seconds.", font=("Segoe UI", 8, "italic"), bg="#1a1c23", fg="#5c6370").pack(pady=(12, 0))
+
+            splash_root.attributes("-topmost", True)
+            splash_root.mainloop()
+        except Exception:
+            pass
+
+    # Launch the GUI splash screen immediately on a background thread
+    threading.Thread(target=show_splash_instantly, daemon=True).start()
+
+
+def create_tray_image():
+    # Generate a beautiful 64x64 icon matching Odysseus brand red accent (#e06c75)
+    from PIL import Image, ImageDraw
+    image = Image.new('RGBA', (64, 64), (0, 0, 0, 0))
+    dc = ImageDraw.Draw(image)
+    accent_red = (224, 108, 117, 255)
+    light_red = (224, 108, 117, 150)
+
+    # Draw premium sailing boat
+    dc.polygon([(32, 10), (32, 45), (12, 45)], fill=accent_red)
+    dc.polygon([(32, 18), (32, 45), (48, 45)], fill=light_red)
+    dc.polygon([(8, 48), (56, 48), (44, 56), (20, 56)], fill=accent_red)
+    return image
+
+
+def on_open_browser(icon, item, url):
+    webbrowser.open(url)
+
+
+def on_exit(icon, item):
+    icon.stop()
+    os._exit(0)
+
+
+def setup_system_tray(url):
+    try:
+        import pystray
+        icon_img = create_tray_image()
+        menu = (
+            pystray.MenuItem('Open Odysseus', lambda icon, item: on_open_browser(icon, item, url), default=True),
+            pystray.MenuItem('Exit', on_exit)
+        )
+        tray_icon = pystray.Icon(
+            "Odysseus",
+            icon_img,
+            "Odysseus",
+            menu
+        )
+        tray_icon.run()
+    except Exception:
+        pass
+
+
+def open_browser(url):
+    # Allow uvicorn and app lifecycles to complete warmups
+    time.sleep(3.5)
+
+    # Safely close the splash screen
+    try:
+        global splash_root
+        if splash_root:
+            splash_root.after(0, splash_root.destroy)
+    except Exception:
+        pass
+
+    webbrowser.open(url)
+
+
+if __name__ == "__main__":
+    import uvicorn
+    # Import the FastAPI app from app.py
+    from app import app
+
+    bind_host = os.getenv("APP_BIND", "127.0.0.1")
+    bind_port = int(os.getenv("APP_PORT", "7000"))
+    url = f"http://{bind_host}:{bind_port}"
+
+    if getattr(sys, 'frozen', False):
+        # Start browser manager thread
+        threading.Thread(target=open_browser, args=(url,), daemon=True).start()
+        # Start system tray manager thread
+        threading.Thread(target=setup_system_tray, args=(url,), daemon=True).start()
+
+    uvicorn.run(app, host=bind_host, port=bind_port, log_level="info")

mcp_servers/email_server.py

@@ -23,6 +23,7 @@ import os.path
 from pathlib import Path
 from datetime import datetime, timedelta
 import uuid
+from contextvars import ContextVar
 
 from mcp.server import Server
 from mcp.server.stdio import stdio_server
@@ -55,6 +56,8 @@ def _uid_fetch_rows(data) -> list:
 # flat keys when no DB row matches (legacy single-account behaviour).
 
 _ACCOUNT_CACHE: dict = {}  # key = normalized account selector -> config dict
+_MCP_OWNER_ARG = "_odysseus_owner"
+_CURRENT_OWNER: ContextVar[str | None] = ContextVar("email_mcp_owner", default=None)
 
 
 def _clean_header_value(value) -> str:
@@ -68,6 +71,45 @@ def _db_path() -> Path:
     return Path(APP_DB)
 
 
+def _current_owner() -> str:
+    owner = _CURRENT_OWNER.get()
+    return str(owner or "").strip()
+
+
+def _account_visible_to_owner(row: dict, owner: str) -> bool:
+    row_owner = str(row.get("owner") or "").strip()
+    if row_owner == owner:
+        return True
+    if row_owner:
+        return False
+    # Legacy ownerless accounts are only visible to a scoped caller when the
+    # mailbox itself matches the owner, mirroring the HTTP email route fallback.
+    owner_l = owner.lower()
+    return owner_l in {
+        str(row.get("imap_user") or "").strip().lower(),
+        str(row.get("from_address") or "").strip().lower(),
+    }
+
+
+def _filter_accounts_for_owner(rows: list[dict]) -> list[dict]:
+    owner = _current_owner()
+    if owner:
+        return [r for r in rows if _account_visible_to_owner(r, owner)]
+
+    owners = {str(r.get("owner") or "").strip() for r in rows if str(r.get("owner") or "").strip()}
+    if len(owners) > 1:
+        return []
+    return rows
+
+
+def _mcp_owner_required(rows: list[dict] | None = None) -> bool:
+    if _current_owner():
+        return False
+    rows = rows if rows is not None else _read_accounts_from_db()
+    owners = {str(r.get("owner") or "").strip() for r in rows if str(r.get("owner") or "").strip()}
+    return len(owners) > 1
+
+
 def _load_email_writing_style() -> str:
     """Return the existing Settings > Email > Writing Style value."""
     try:
@@ -121,9 +163,8 @@ def _default_document_owner() -> str | None:
         return None
 
 
-def _list_accounts_raw() -> list:
+def _read_accounts_from_db() -> list:
-    """Return list of dicts from the email_accounts table. Empty list if table
+    """Return all enabled email account rows. Empty list if missing. Never raises."""
-    missing or empty. Never raises."""
     path = _db_path()
     if not path.exists():
         return []
@@ -131,9 +172,10 @@ def _list_accounts_raw() -> list:
         conn = sqlite3.connect(str(path))
         conn.row_factory = sqlite3.Row
         columns = {r[1] for r in conn.execute("PRAGMA table_info(email_accounts)").fetchall()}
+        owner_select = "owner" if "owner" in columns else "NULL AS owner"
         smtp_security_select = "smtp_security" if "smtp_security" in columns else "'' AS smtp_security"
         rows = conn.execute(f"""
-            SELECT id, name, is_default, enabled,
+            SELECT id, {owner_select}, name, is_default, enabled,
                    imap_host, imap_port, imap_user, imap_password, imap_starttls,
                    smtp_host, smtp_port, {smtp_security_select}, smtp_user, smtp_password, from_address
             FROM email_accounts WHERE enabled = 1
@@ -147,11 +189,15 @@ def _list_accounts_raw() -> list:
         return []
 
 
-def _resolve_account(selector: str | None) -> dict | None:
+def _list_accounts_raw() -> list:
+    """Return owner-visible email account rows for the active MCP call."""
+    return _filter_accounts_for_owner(_read_accounts_from_db())
+
+
+def _resolve_account_from_rows(rows: list[dict], selector: str | None) -> dict | None:
     """Given a selector (None = default, or a name/user/id string), return the
     matching row or None. Matching is case-insensitive substring on name +
     imap_user + from_address, plus exact id match."""
-    rows = _list_accounts_raw()
     if not rows:
         return None
     if not selector:
@@ -186,6 +232,10 @@ def _resolve_account(selector: str | None) -> dict | None:
     return None
 
 
+def _resolve_account(selector: str | None) -> dict | None:
+    return _resolve_account_from_rows(_list_accounts_raw(), selector)
+
+
 def _load_config(account: str | None = None) -> dict:
     """Return the full config dict for the requested account (or default).
 
@@ -194,7 +244,7 @@ def _load_config(account: str | None = None) -> dict:
       2. env vars + settings.json flat keys (legacy)
       3. hardcoded fallbacks (localhost:31143 etc.)
     """
-    cache_key = (account or "").strip().lower() or "__default__"
+    cache_key = (_current_owner(), (account or "").strip().lower() or "__default__")
     if cache_key in _ACCOUNT_CACHE:
         return _ACCOUNT_CACHE[cache_key]
 
@@ -223,8 +273,11 @@ def _load_config(account: str | None = None) -> dict:
         "account_name": None,
     }
 
-    rows = _list_accounts_raw()
+    raw_rows = _read_accounts_from_db()
-    row = _resolve_account(account)
+    rows = _filter_accounts_for_owner(raw_rows)
+    row = _resolve_account_from_rows(rows, account)
+    if _current_owner() and raw_rows and not rows:
+        raise ValueError("No email account is configured for the authenticated owner")
     if account and rows and not row:
         available = ", ".join(
             f"{r.get('name') or r.get('imap_user')} <{r.get('imap_user') or r.get('from_address') or '?'}>"
@@ -953,7 +1006,7 @@ def _stash_agent_draft(*, to, subject, body, in_reply_to=None, references=None,
             now,
             account or None,
             "agent_draft",
-            "",
+            _current_owner(),
         ))
         conn.commit()
         conn.close()
@@ -1139,7 +1192,7 @@ def _create_email_draft_document(
     doc_id = str(uuid.uuid4())
     ver_id = str(uuid.uuid4())
     doc_title = (title or subject or "Email draft").strip() or "Email draft"
-    doc_owner = _default_document_owner()
+    doc_owner = _current_owner() or _default_document_owner()
 
     db = SessionLocal()
     try:
@@ -1925,10 +1978,22 @@ async def list_tools() -> list[Tool]:
 
 @server.call_tool()
 async def call_tool(name: str, arguments: dict) -> list[TextContent]:
+    arguments = dict(arguments) if isinstance(arguments, dict) else {}
+    owner = str(arguments.pop(_MCP_OWNER_ARG, "") or "").strip()
+    owner_token = _CURRENT_OWNER.set(owner or None)
     try:
+        all_db_accounts = _read_accounts_from_db()
+        if _mcp_owner_required(all_db_accounts):
+            return [TextContent(
+                type="text",
+                text="Error: email MCP requires an authenticated owner when multiple email account owners are configured.",
+            )]
+
         if name == "list_email_accounts":
-            rows = _list_accounts_raw()
+            rows = _filter_accounts_for_owner(all_db_accounts)
             if not rows:
+                if all_db_accounts and owner:
+                    return [TextContent(type="text", text="No email accounts configured for this owner.")]
                 return [TextContent(type="text", text="No email accounts configured. Legacy single-account mode active.")]
             lines = [f"Found {len(rows)} email account(s):\n"]
             for r in rows:
@@ -2108,6 +2173,16 @@ async def call_tool(name: str, arguments: dict) -> list[TextContent]:
                 bcc=arguments.get("bcc"),
                 account=acct,
             )
+            if "error" in result:
+                return [TextContent(type="text", text=f"Error: {result['error']}")]
+            if result.get("pending"):
+                return [TextContent(
+                    type="text",
+                    text=(
+                        f"Draft staged for approval (pending id: {result.get('pending_id')}). "
+                        "Nothing has been sent yet. Review and approve it in Odysseus before delivery."
+                    ),
+                )]
             acct_note = f" (from {result['account']})" if result.get("account") else ""
             return [TextContent(type="text", text=f"Sent email to {result['to']} with subject '{result['subject']}'{acct_note}.")]
 
@@ -2283,6 +2358,8 @@ async def call_tool(name: str, arguments: dict) -> list[TextContent]:
 
     except Exception as e:
         return [TextContent(type="text", text=f"Error: {e}")]
+    finally:
+        _CURRENT_OWNER.reset(owner_token)
 
 
 # ── Main ──

package-lock.json

@@ -4,93 +4,19 @@
   "requires": true,
   "packages": {
     "": {
-      "dependencies": {
-        "@anthropic-ai/sdk": "^0.104.1"
-      },
       "devDependencies": {
-        "@antithesishq/bombadil": "^0.5.0"
+        "@antithesishq/bombadil": "^0.6.1"
-      }
-    },
-    "node_modules/@anthropic-ai/sdk": {
-      "version": "0.104.1",
-      "resolved": "https://registry.npmjs.org/@anthropic-ai/sdk/-/sdk-0.104.1.tgz",
-      "integrity": "sha512-gGACa/+IaiXzRRmF96aOhamoBgapKRBiFWbmmTFP8aMkpaEcuStF+Q61bjo4vPxBM7gqWJNZqsngslRdnLHv0Q==",
-      "license": "MIT",
-      "dependencies": {
-        "json-schema-to-ts": "^3.1.1",
-        "standardwebhooks": "^1.0.0"
-      },
-      "bin": {
-        "anthropic-ai-sdk": "bin/cli"
-      },
-      "peerDependencies": {
-        "zod": "^3.25.0 || ^4.0.0"
-      },
-      "peerDependenciesMeta": {
-        "zod": {
-          "optional": true
-        }
       }
     },
     "node_modules/@antithesishq/bombadil": {
-      "version": "0.5.0",
+      "version": "0.6.1",
-      "resolved": "https://registry.npmjs.org/@antithesishq/bombadil/-/bombadil-0.5.0.tgz",
+      "resolved": "https://registry.npmjs.org/@antithesishq/bombadil/-/bombadil-0.6.1.tgz",
-      "integrity": "sha512-s0zImmr0iyvSP6QcVLvf40CUiZYIdWBAxiq20uhzujwvfitYa3PGJN652k/pLtVccHM/JrGQxZdvLnihZpltHA==",
+      "integrity": "sha512-d1iufG3MI7gSMSiSmMeNdcMW+qR0yQXL2zdkVynC3n3DYgFJYlYXKUQzygmqU12m4RWlR5iOdQU1hsx5UT6+IA==",
       "dev": true,
       "license": "MIT",
       "bin": {
         "bombadil": "bin/bombadil.js"
       }
-    },
-    "node_modules/@babel/runtime": {
-      "version": "7.29.7",
-      "resolved": "https://registry.npmjs.org/@babel/runtime/-/runtime-7.29.7.tgz",
-      "integrity": "sha512-Nq8OhGWiZIZGV6hLHoyAKLLcJihP/xFeBMGJoUrxTX2psI8dCifzLhZISFb+VWS3wFMRDmCGw5R+dOySCqPLhw==",
-      "license": "MIT",
-      "engines": {
-        "node": ">=6.9.0"
-      }
-    },
-    "node_modules/@stablelib/base64": {
-      "version": "1.0.1",
-      "resolved": "https://registry.npmjs.org/@stablelib/base64/-/base64-1.0.1.tgz",
-      "integrity": "sha512-1bnPQqSxSuc3Ii6MhBysoWCg58j97aUjuCSZrGSmDxNqtytIi0k8utUenAwTZN4V5mXXYGsVUI9zeBqy+jBOSQ==",
-      "license": "MIT"
-    },
-    "node_modules/fast-sha256": {
-      "version": "1.3.0",
-      "resolved": "https://registry.npmjs.org/fast-sha256/-/fast-sha256-1.3.0.tgz",
-      "integrity": "sha512-n11RGP/lrWEFI/bWdygLxhI+pVeo1ZYIVwvvPkW7azl/rOy+F3HYRZ2K5zeE9mmkhQppyv9sQFx0JM9UabnpPQ==",
-      "license": "Unlicense"
-    },
-    "node_modules/json-schema-to-ts": {
-      "version": "3.1.1",
-      "resolved": "https://registry.npmjs.org/json-schema-to-ts/-/json-schema-to-ts-3.1.1.tgz",
-      "integrity": "sha512-+DWg8jCJG2TEnpy7kOm/7/AxaYoaRbjVB4LFZLySZlWn8exGs3A4OLJR966cVvU26N7X9TWxl+Jsw7dzAqKT6g==",
-      "license": "MIT",
-      "dependencies": {
-        "@babel/runtime": "^7.18.3",
-        "ts-algebra": "^2.0.0"
-      },
-      "engines": {
-        "node": ">=16"
-      }
-    },
-    "node_modules/standardwebhooks": {
-      "version": "1.0.0",
-      "resolved": "https://registry.npmjs.org/standardwebhooks/-/standardwebhooks-1.0.0.tgz",
-      "integrity": "sha512-BbHGOQK9olHPMvQNHWul6MYlrRTAOKn03rOe4A8O3CLWhNf4YHBqq2HJKKC+sfqpxiBY52pNeesD6jIiLDz8jg==",
-      "license": "MIT",
-      "dependencies": {
-        "@stablelib/base64": "^1.0.0",
-        "fast-sha256": "^1.3.0"
-      }
-    },
-    "node_modules/ts-algebra": {
-      "version": "2.0.0",
-      "resolved": "https://registry.npmjs.org/ts-algebra/-/ts-algebra-2.0.0.tgz",
-      "integrity": "sha512-FPAhNPFMrkwz76P7cdjdmiShwMynZYN6SgOujD1urY4oNm80Ou9oMdmbR45LotcKOXoy7wSmHkRFE6Mxbrhefw==",
-      "license": "MIT"
     }
   }
 }

package.json

@@ -4,9 +4,6 @@
     "url": "https://github.com/pewdiepie-archdaemon/odysseus.git"
   },
   "devDependencies": {
-    "@antithesishq/bombadil": "^0.5.0"
+    "@antithesishq/bombadil": "^0.6.1"
-  },
-  "dependencies": {
-    "@anthropic-ai/sdk": "^0.104.1"
   }
 }

routes/assistant_routes.py

@@ -16,6 +16,7 @@ from pydantic import BaseModel
 
 from core.database import SessionLocal, CrewMember, ScheduledTask
 from src.auth_helpers import get_current_user
+from core.auth import RESERVED_USERNAMES
 from src.task_scheduler import compute_next_run
 
 
@@ -89,11 +90,11 @@ def setup_assistant_routes(task_scheduler) -> APIRouter:
     # check-in tasks seeded. Hitting any /assistant route under one of these
     # used to seed a full CrewMember + Morning/Midday/Evening tasks under that
     # owner, which then double-fired alongside the real user's check-ins.
-    _SYNTHETIC_OWNERS = frozenset({"internal-tool", "api", "demo", "system", ""})
+    # RESERVED_USERNAMES covers the same set; the `not owner` guard handles "".
 
     async def _get_or_create(owner: str) -> CrewMember:
         """Return the per-owner assistant CrewMember, creating it on demand."""
-        if not owner or owner in _SYNTHETIC_OWNERS:
+        if not owner or owner in RESERVED_USERNAMES:
             raise HTTPException(status_code=400, detail=f"Cannot seed assistant for {owner!r}")
         db = SessionLocal()
         try:

routes/auth_routes.py

@@ -12,8 +12,8 @@ import re
 from pathlib import Path
 
 from core.atomic_io import atomic_write_json, atomic_write_text
-from core.auth import AuthManager, SetAdminResult
+from core.auth import AuthManager, RESERVED_USERNAMES, SetAdminResult, TOKEN_TTL
-from src.constants import DEEP_RESEARCH_DIR, MEMORY_FILE, SKILLS_DIR
+from src.constants import DEEP_RESEARCH_DIR, MEMORY_FILE, PASSWORD_MIN_LENGTH, SKILLS_DIR
 from src.rate_limiter import RateLimiter
 from src.settings_scrub import scrub_settings
 from src.settings import (
@@ -102,8 +102,12 @@ def setup_auth_routes(auth_manager: AuthManager) -> APIRouter:
             raise HTTPException(429, "Too many requests — try again later")
         if auth_manager.is_configured:
             raise HTTPException(400, "Already configured")
-        if len(body.password) < 8:
+        if len(body.password) < PASSWORD_MIN_LENGTH:
-            raise HTTPException(400, "Password must be at least 8 characters")
+            raise HTTPException(400, f"Password must be at least {PASSWORD_MIN_LENGTH} characters")
+        if len(body.username.strip()) < 1:
+            raise HTTPException(400, "Username is required")
+        if body.username.lower() in RESERVED_USERNAMES:
+            raise HTTPException(403, "Username is reserved")
         ok = await asyncio.to_thread(auth_manager.setup, body.username, body.password)
         if not ok:
             raise HTTPException(500, "Setup failed")
@@ -118,10 +122,12 @@ def setup_auth_routes(auth_manager: AuthManager) -> APIRouter:
             raise HTTPException(400, "Run setup first")
         if not auth_manager.signup_enabled:
             raise HTTPException(403, "Registration is disabled. Ask an admin for an account.")
-        if len(body.password) < 8:
+        if len(body.password) < PASSWORD_MIN_LENGTH:
-            raise HTTPException(400, "Password must be at least 8 characters")
+            raise HTTPException(400, f"Password must be at least {PASSWORD_MIN_LENGTH} characters")
         if len(body.username.strip()) < 1:
             raise HTTPException(400, "Username is required")
+        if body.username.lower() in RESERVED_USERNAMES:
+            raise HTTPException(403, "Username is reserved")
         ok = await asyncio.to_thread(auth_manager.create_user, body.username, body.password, is_admin=False)
         if not ok:
             raise HTTPException(409, "Username already taken")
@@ -144,6 +150,8 @@ def setup_auth_routes(auth_manager: AuthManager) -> APIRouter:
                 raise HTTPException(401, "Invalid 2FA code")
         # All checks passed — create session (password already verified above)
         token = await asyncio.to_thread(auth_manager.create_session_trusted, username)
+        if not token:
+            raise HTTPException(401, "Invalid credentials")
         cookie_kwargs = dict(
             key=SESSION_COOKIE,
             value=token,
@@ -153,7 +161,7 @@ def setup_auth_routes(auth_manager: AuthManager) -> APIRouter:
             path="/",
         )
         if body.remember:
-            cookie_kwargs["max_age"] = 60 * 60 * 24 * 7  # 7 days
+            cookie_kwargs["max_age"] = TOKEN_TTL
         response.set_cookie(**cookie_kwargs)
         return {"ok": True, "username": username}
 
@@ -182,13 +190,18 @@ def setup_auth_routes(auth_manager: AuthManager) -> APIRouter:
             pass
         return result
 
+    @router.get("/policy")
+    async def auth_policy():
+        """Return public auth policy constants for the frontend."""
+        return auth_manager.policy()
+
     @router.post("/change-password")
     async def change_password(body: ChangePasswordRequest, request: Request):
         user = _get_current_user(request)
         if not user:
             raise HTTPException(401, "Not authenticated")
-        if len(body.new_password) < 8:
+        if len(body.new_password) < PASSWORD_MIN_LENGTH:
-            raise HTTPException(400, "Password must be at least 8 characters")
+            raise HTTPException(400, f"Password must be at least {PASSWORD_MIN_LENGTH} characters")
         current_token = request.cookies.get(SESSION_COOKIE)
         ok = await asyncio.to_thread(auth_manager.change_password, user, body.current_password, body.new_password)
         if not ok:
@@ -268,8 +281,12 @@ def setup_auth_routes(auth_manager: AuthManager) -> APIRouter:
         user = _get_current_user(request)
         if not user or not auth_manager.is_admin(user):
             raise HTTPException(403, "Admin only")
-        if len(body.password) < 8:
+        if len(body.password) < PASSWORD_MIN_LENGTH:
-            raise HTTPException(400, "Password must be at least 8 characters")
+            raise HTTPException(400, f"Password must be at least {PASSWORD_MIN_LENGTH} characters")
+        if len(body.username.strip()) < 1:
+            raise HTTPException(400, "Username is required")
+        if body.username.lower() in RESERVED_USERNAMES:
+            raise HTTPException(403, "Username is reserved")
         ok = auth_manager.create_user(body.username, body.password, body.is_admin)
         if not ok:
             raise HTTPException(409, "Username already taken")
@@ -432,6 +449,23 @@ def setup_auth_routes(auth_manager: AuthManager) -> APIRouter:
         except Exception as e:
             logger.warning("Failed to rename upload owner references %s -> %s: %s", old_username, new_username, e)
 
+        # direct personal RAG uploads live in per-owner directories and the
+        # vector metadata also carries the username used for owner-filtered
+        # search. Keep both in sync with the auth rename.
+        try:
+            from routes.personal_routes import rename_personal_upload_owner
+            personal_docs_manager = getattr(request.app.state, "personal_docs_manager", None)
+            if personal_docs_manager is not None:
+                rag_manager = getattr(personal_docs_manager, "rag_manager", None)
+                rename_personal_upload_owner(
+                    old_username,
+                    new_username,
+                    personal_docs_manager=personal_docs_manager,
+                    rag_manager=rag_manager,
+                )
+        except Exception as e:
+            logger.warning("Failed to rename personal RAG upload owner references %s -> %s: %s", old_username, new_username, e)
+
         # skills: SKILL.md frontmatter carries owner: <username>; the usage
         # sidecar (_usage.json) keys entries as owner::skill-name. Both must
         # be updated or the renamed user's Skills panel goes empty.

routes/chat_helpers.py

@@ -23,6 +23,22 @@ from fastapi import HTTPException
 logger = logging.getLogger(__name__)
 
 
+# Strong references to in-flight fire-and-forget tasks scheduled from this
+# module. asyncio only keeps weak references to tasks created via
+# create_task, so without this the GC can collect a task mid-execution and
+# the background work (extraction, auto-naming) silently never runs.
+# Mirrors WebhookManager._spawn_tracked from src/webhook_manager.py.
+_BG_TASKS: set[asyncio.Task] = set()
+
+
+def _spawn_bg(coro) -> asyncio.Task:
+    """Schedule a background task and hold a strong reference until it finishes."""
+    task = asyncio.create_task(coro)
+    _BG_TASKS.add(task)
+    task.add_done_callback(_BG_TASKS.discard)
+    return task
+
+
 # ── Data containers ────────────────────────────────────────────────────── #
 
 @dataclass
@@ -159,17 +175,9 @@ async def auto_name_session(session_manager, sess):
             return
 
         owner = getattr(sess, "owner", None)
-        t_url, t_model, t_headers = resolve_task_endpoint(owner=owner)
+        t_url, t_model, t_headers = resolve_task_endpoint(
-        if not t_model:
+            sess.endpoint_url, sess.model, sess.headers, owner=owner
-            # If no task/utility model is configured at all, fall back to
+        )
-            # the session's own model so auto-naming still works even on
-            # minimal setups.
-            from src.endpoint_resolver import resolve_endpoint
-            _fallback = resolve_endpoint("default", owner=owner)
-            if _fallback and _fallback[1]:
-                t_url, t_model, t_headers = _fallback
-            else:
-                t_url, t_model, t_headers = sess.endpoint_url, sess.model, sess.headers
         if not t_model:
             logger.debug("[auto-name] No model provided, skipping")
             return
@@ -576,7 +584,8 @@ async def build_chat_context(
     if not incognito:
         fire_message_event(request, webhook_manager, session_id, sess, message, compare_mode)
 
-    # Resolve user prefs
+    # Resolve owner-scoped prefs/context. Browser requests keep the cookie user;
+    # bearer-token chat requests use the token owner instead of the "api" sentinel.
     user = effective_user(request)
     uprefs = load_prefs_for_user(user)
 
@@ -1112,7 +1121,7 @@ def run_post_response_tasks(
             )))
 
     if _extraction_jobs:
-        asyncio.create_task(_run_extraction_jobs_sequentially(session_id, _extraction_jobs))
+        _spawn_bg(_run_extraction_jobs_sequentially(session_id, _extraction_jobs))
 
     # Token accumulation
     if last_metrics:
@@ -1127,4 +1136,4 @@ def run_post_response_tasks(
 
     # Auto-name
     if needs_auto_name(sess.name):
-        asyncio.create_task(auto_name_session(session_manager, sess))
+        _spawn_bg(auto_name_session(session_manager, sess))

routes/codex_routes.py

@@ -46,8 +46,12 @@ def _ssh_prefix_for_task(task: dict) -> tuple[str, str]:
     shell metacharacters in ``remoteHost`` is rejected with 400 rather than
     injected.
     """
-    host = validate_remote_host((task.get("remoteHost") or "").strip() or None) or ""
+    raw_host = task.get("remoteHost")
-    ssh_port = validate_ssh_port((task.get("sshPort") or "").strip() or None) or ""
+    raw_port = task.get("sshPort")
+    host_value = str(raw_host).strip() if raw_host is not None else None
+    port_value = str(raw_port).strip() if raw_port is not None else None
+    host = validate_remote_host(host_value or None) or ""
+    ssh_port = validate_ssh_port(port_value or None) or ""
     port_flag = f"-p {ssh_port} " if ssh_port and ssh_port != "22" else ""
     return host, port_flag
 
@@ -306,7 +310,10 @@ def setup_codex_routes(
 
     @router.post("/emails/draft-document")
     async def codex_email_draft_document(request: Request, body: dict[str, Any] = Body(default_factory=dict)):
-        owner = _scope_owner_all(request, {"email:draft", "documents:write"})
+        owner = _scope_owner(request, EMAIL_DRAFT_SCOPES)
+        docs_owner = _scope_owner_all(request, DOCS_WRITE_SCOPES)
+        if docs_owner != owner:
+            raise HTTPException(403, "API token owner mismatch")
         if documents_create_endpoint is None:
             raise HTTPException(503, "Documents integration is not available")
         from routes.document_routes import DocumentCreate

routes/cookbook_routes.py

@@ -1284,6 +1284,11 @@ def setup_cookbook_routes() -> APIRouter:
         # LOCAL execution on a native-Windows host never uses tmux (detached
         # process path below), regardless of the UI-supplied platform.
         local_windows = IS_WINDOWS and not remote
+        if is_windows and remote and "diffusion_server.py" in req.cmd:
+            raise HTTPException(
+                400,
+                "Remote Windows Diffusers serving is not supported yet; use local Windows or a Linux remote server.",
+            )
 
         if not is_windows and not local_windows and not await _binary_available("tmux", remote, req.ssh_port):
             return {

routes/document_helpers.py

@@ -102,8 +102,11 @@ def _owner_session_filter(q, user):
 
     The owner backfill runs in init_db before the app serves requests, so
     by the time this filter is live there are no NULL-owner rows to leak;
-    we therefore match the owner strictly."""
+    we therefore match the owner strictly for authenticated callers."""
-    if user is None:
+    if not user:
+        from src.auth_helpers import _auth_disabled
+        if user == "" or _auth_disabled():
+            return q
         return q.filter(False)
     return q.filter(Document.owner == user)
 

routes/document_routes.py

@@ -1332,6 +1332,12 @@ def setup_document_routes(session_manager, upload_handler=None) -> APIRouter:
             if not pdf_path:
                 raise HTTPException(404, f"Source PDF {upload_id} not found")
 
+            # Fail fast with a clear 503 if the optional PyMuPDF dependency
+            # is missing — fill_fields/stamp_annotations will otherwise
+            # raise RuntimeError deep inside and bubble out as a 500.
+            # Mirrors the convention in _load_pdf_viewer_fitz above.
+            _load_pdf_viewer_fitz()
+
             values = parse_markdown_to_values(doc.current_content or "")
             out_path = tempfile.NamedTemporaryFile(suffix=".pdf", delete=False).name
             _to_unlink.append(out_path)

routes/email_routes.py

@@ -2171,12 +2171,10 @@ def setup_email_routes():
         try:
             conn = sqlite3.connect(SCHEDULED_DB)
             conn.row_factory = sqlite3.Row
-            # The MCP server can't easily set owner, so it stores '' — fall
-            # back to those rows in addition to the caller's owner.
             rows = conn.execute(
                 """SELECT id, to_addr, subject, body, created_at, account_id
                    FROM scheduled_emails
-                   WHERE status = 'agent_draft' AND (owner = ? OR owner = '')
+                   WHERE status = 'agent_draft' AND owner = ?
                    ORDER BY created_at DESC""",
                 (owner or "",),
             ).fetchall()
@@ -2197,7 +2195,7 @@ def setup_email_routes():
             cur = conn.execute(
                 """UPDATE scheduled_emails
                    SET status = 'pending', send_at = ?
-                   WHERE id = ? AND status = 'agent_draft' AND (owner = ? OR owner = '')""",
+                   WHERE id = ? AND status = 'agent_draft' AND owner = ?""",
                 (datetime.utcnow().isoformat(), sid, owner or ""),
             )
             conn.commit()
@@ -2218,7 +2216,7 @@ def setup_email_routes():
             conn = sqlite3.connect(SCHEDULED_DB)
             cur = conn.execute(
                 """UPDATE scheduled_emails SET status = 'cancelled'
-                   WHERE id = ? AND status = 'agent_draft' AND (owner = ? OR owner = '')""",
+                   WHERE id = ? AND status = 'agent_draft' AND owner = ?""",
                 (sid, owner or ""),
             )
             conn.commit()

routes/gallery_routes.py

@@ -67,14 +67,6 @@ def _gallery_image_path(filename: str) -> Path:
         raise HTTPException(400, "Unsafe gallery filename")
     if safe_name != original:
         raise HTTPException(400, "Unsafe gallery filename")
-    if not path.exists():
-        cwd_root = (Path.cwd() / "data" / "generated_images").resolve()
-        cwd_path = (cwd_root / safe_name).resolve()
-        try:
-            if os.path.commonpath([str(cwd_root), str(cwd_path)]) == str(cwd_root) and cwd_path.exists():
-                return cwd_path
-        except Exception:
-            pass
     return path
 
 

routes/memory_routes.py

@@ -273,65 +273,30 @@ def setup_memory_routes(memory_manager: MemoryManager, session_manager: SessionM
     async def api_audit_memories(request: Request, session: str = Form(None)):
         """Deduplicate and consolidate memories via LLM.
 
-        Uses the default model from settings, or falls back to a session's model.
+        Uses task/utility/default settings through the shared resolver, with
+        the active session as fallback when no task or utility model is set.
         Returns before and after memory counts.
         """
-        from routes.model_routes import _load_settings, _normalize_base, build_chat_url
-        from core.database import ModelEndpoint
-        import json as _json
-
-        endpoint_url = model = None
-        headers = {}
-
-        # Try utility model from settings first — memory audit is a background
-        # task and should prefer the lighter utility model over the main chat model.
-        from src.task_endpoint import resolve_task_endpoint
         user = _owner(request)
-        t_url, t_model, t_headers = resolve_task_endpoint(owner=user)
+        fallback_url = fallback_model = None
-        if t_url and t_model:
+        fallback_headers = None
-            endpoint_url, model, headers = t_url, t_model, t_headers
+        if session:
-        else:
-            # Fall back to default model if no task/utility model configured
-            settings = _load_settings()
-            ep_id = settings.get("default_endpoint_id", "")
-            default_model = settings.get("default_model", "")
-            if ep_id:
-                db = SessionLocal()
-                try:
-                    ep = db.query(ModelEndpoint).filter(
-                        ModelEndpoint.id == ep_id, ModelEndpoint.is_enabled == True
-                    ).first()
-                    if ep:
-                        base = _normalize_base(ep.base_url)
-                        endpoint_url = build_chat_url(base)
-                        model = default_model
-                        if not model and ep.models:
-                            try:
-                                models = _json.loads(ep.models) if isinstance(ep.models, str) else ep.models
-                                if models:
-                                    model = models[0]
-                            except Exception:
-                                pass
-                        if ep.api_key:
-                            headers = {"Authorization": f"Bearer {ep.api_key}"}
-                finally:
-                    db.close()
-
-            # Fall back to session model if no default configured
-            if not endpoint_url and session:
             try:
                 sess = session_manager.get_session(session)
-                    _assert_session_owner(sess, _owner(request))
+                _assert_session_owner(sess, user)
-                    endpoint_url = sess.endpoint_url
+                fallback_url = sess.endpoint_url
-                    model = sess.model
+                fallback_model = sess.model
-                    headers = sess.headers
+                fallback_headers = sess.headers
             except KeyError:
                 pass
 
+        endpoint_url, model, headers = resolve_task_endpoint(
+            fallback_url, fallback_model, fallback_headers, owner=user
+        )
+
         if not endpoint_url or not model:
             raise HTTPException(400, "No default model configured — set one in Settings")
 
-        user = _owner(request)
         result = await audit_memories(
             memory_manager,
             memory_vector,
@@ -369,18 +334,28 @@ def setup_memory_routes(memory_manager: MemoryManager, session_manager: SessionM
         model = None
         headers = {}
 
+        user = _owner(request)
+
         if session:
             try:
                 sess = session_manager.get_session(session)
-                _assert_session_owner(sess, _owner(request))
+                _assert_session_owner(sess, user)
-                endpoint_url, model, headers = resolve_task_endpoint(
-                    sess.endpoint_url, sess.model, sess.headers, owner=_owner(request)
-                )
             except KeyError:
-                logger.warning("Session %s not found, falling back to utility endpoint", session)
+                sess = None
-                endpoint_url, model, headers = resolve_endpoint("utility", owner=_owner(request))
+            except HTTPException as exc:
+                if exc.status_code != 404:
+                    raise
+                sess = None
+
+            if sess is None:
+                logger.warning("Session %s not found or inaccessible, falling back to utility endpoint", session)
+                endpoint_url, model, headers = resolve_endpoint("utility", owner=user)
             else:
-            endpoint_url, model, headers = resolve_task_endpoint(owner=_owner(request))
+                endpoint_url, model, headers = resolve_task_endpoint(
+                    sess.endpoint_url, sess.model, sess.headers, owner=user
+                )
+        else:
+            endpoint_url, model, headers = resolve_task_endpoint(owner=user)
     
         if not endpoint_url or not model:
             raise HTTPException(400, "No LLM model configured. Set a default model in Settings.")

routes/model_routes.py

@@ -5,6 +5,7 @@ import re
 import uuid
 import json
 import hashlib
+import ipaddress
 import socket
 import time as _time
 import logging
@@ -562,6 +563,8 @@ def _safe_build_models_url(base_url: str) -> str:
     """Build a /models URL without letting optional provider imports break probes."""
     try:
         return build_models_url(base_url)
+    except ValueError:
+        raise
     except Exception as exc:
         logger.debug("Model URL detection failed for %s: %s", base_url, exc)
         return f"{(base_url or '').rstrip('/')}/models"
@@ -633,7 +636,7 @@ def _probe_single_model(base: str, api_key: str, model_id: str, timeout: int = 1
 
     try:
         t0 = _time.time()
-        r = httpx.post(target_url, headers=h, json=payload, timeout=timeout)
+        r = httpx.post(target_url, headers=h, json=payload, timeout=timeout, verify=llm_verify())
         latency = round((_time.time() - t0) * 1000)
         if r.is_success:
             return {"status": "ok", "latency_ms": latency}
@@ -659,13 +662,20 @@ def _probe_single_model(base: str, api_key: str, model_id: str, timeout: int = 1
 
 # Hostnames / IP prefixes that indicate a local endpoint
 _LOCAL_HOSTS = {"localhost", "127.0.0.1", "0.0.0.0", "::1"}
-_PRIVATE_PREFIXES = ("10.", "172.16.", "172.17.", "172.18.", "172.19.",
+_PRIVATE_NETWORKS = (
-                     "172.20.", "172.21.", "172.22.", "172.23.", "172.24.",
+    ipaddress.ip_network("10.0.0.0/8"),
-                     "172.25.", "172.26.", "172.27.", "172.28.", "172.29.",
+    ipaddress.ip_network("172.16.0.0/12"),
-                     "172.30.", "172.31.", "192.168.")
+    ipaddress.ip_network("192.168.0.0/16"),
+)
+_TAILSCALE_CGNAT = ipaddress.ip_network("100.64.0.0/10")
 
 
-_TAILSCALE_RE = re.compile(r"^100\.(6[4-9]|[7-9]\d|1[01]\d|12[0-7])\.")
+def _local_ip_literal(host: str) -> bool:
+    try:
+        ip = ipaddress.ip_address(host)
+    except ValueError:
+        return False
+    return any(ip in network for network in _PRIVATE_NETWORKS) or ip in _TAILSCALE_CGNAT
 
 
 def _classify_endpoint(base_url: str, endpoint_kind: str = "auto") -> str:
@@ -679,9 +689,7 @@ def _classify_endpoint(base_url: str, endpoint_kind: str = "auto") -> str:
         return "api"
     try:
         host = urlparse(base_url).hostname or ""
-        if host in _LOCAL_HOSTS or host.startswith(_PRIVATE_PREFIXES):
+        if host in _LOCAL_HOSTS or _local_ip_literal(host):
-            return "local"
-        if _TAILSCALE_RE.match(host):
             return "local"
     except Exception:
         pass

routes/note_routes.py

@@ -10,6 +10,7 @@ from fastapi import APIRouter, HTTPException, Request
 from pydantic import BaseModel
 
 from core.database import SessionLocal, Note
+from core.middleware import INTERNAL_TOOL_USER
 from src.auth_helpers import require_user
 from src.constants import DATA_DIR
 from sqlalchemy.orm.attributes import flag_modified
@@ -582,7 +583,7 @@ def setup_note_routes(task_scheduler=None):
         return require_user(request) or None
 
     def _is_admin_or_single_user(request: Request, user: str | None) -> bool:
-        if user == "internal-tool":
+        if user == INTERNAL_TOOL_USER:
             return True
         if not user:
             # require_user() already admitted this request, which only happens

routes/personal_routes.py

@@ -2,8 +2,9 @@
 """Routes for personal documents management."""
 import os
 import logging
+import shutil
 import uuid
-from typing import List, Tuple
+from typing import Any, Dict, List, Tuple
 from fastapi import APIRouter, HTTPException, Query, Request, UploadFile, File, Depends
 from src.request_models import DirectoryRequest
 from core.constants import BASE_DIR, PERSONAL_DIR, PERSONAL_UPLOADS_DIR
@@ -18,13 +19,14 @@ UPLOADS_DIR = PERSONAL_UPLOADS_DIR
 logger = logging.getLogger(__name__)
 
 
-def _personal_upload_dir_for_owner(owner: str | None) -> str:
+def _personal_upload_dir_for_owner(owner: str | None, *, create: bool = True) -> str:
     """Return the per-owner upload directory used for direct RAG uploads."""
     owner_segment = secure_filename((owner or "local").strip())[:80] or "local"
     upload_dir = os.path.abspath(os.path.join(UPLOADS_DIR, owner_segment))
     base_abs = os.path.abspath(UPLOADS_DIR)
     if os.path.commonpath([upload_dir, base_abs]) != base_abs:
         raise ValueError("Unsafe upload owner path")
+    if create:
         os.makedirs(upload_dir, exist_ok=True)
     return upload_dir
 
@@ -44,6 +46,87 @@ def _unique_personal_upload_path(upload_dir: str, original_name: str | None) ->
         raise ValueError("Unsafe upload filename")
     return file_path, filename, safe_name
 
+
+def _unique_existing_target(path: str) -> str:
+    """Return a non-existing sibling path for rename collision handling."""
+    if not os.path.exists(path):
+        return path
+    stem, ext = os.path.splitext(path)
+    while True:
+        candidate = f"{stem}-{uuid.uuid4().hex[:10]}{ext}"
+        if not os.path.exists(candidate):
+            return candidate
+
+
+def _remove_empty_tree(path: str) -> None:
+    """Best-effort removal of empty directories under ``path``."""
+    if not os.path.isdir(path):
+        return
+    for root, dirs, _files in os.walk(path, topdown=False):
+        for dirname in dirs:
+            candidate = os.path.join(root, dirname)
+            try:
+                os.rmdir(candidate)
+            except OSError:
+                pass
+    try:
+        os.rmdir(path)
+    except OSError:
+        pass
+
+
+def rename_personal_upload_owner(
+    old_owner: str,
+    new_owner: str,
+    *,
+    personal_docs_manager: Any = None,
+    rag_manager: Any = None,
+) -> Dict[str, Any]:
+    """Move direct personal uploads and rewrite RAG owner metadata on user rename."""
+    old_dir = _personal_upload_dir_for_owner(old_owner, create=False)
+    new_dir = _personal_upload_dir_for_owner(new_owner, create=False)
+    path_map: Dict[str, str] = {}
+    moved_files = 0
+
+    if os.path.isdir(old_dir) and old_dir != new_dir:
+        os.makedirs(new_dir, exist_ok=True)
+        for root, _dirs, files in os.walk(old_dir):
+            rel_root = os.path.relpath(root, old_dir)
+            target_root = new_dir if rel_root == "." else os.path.join(new_dir, rel_root)
+            os.makedirs(target_root, exist_ok=True)
+            for filename in files:
+                source = os.path.abspath(os.path.join(root, filename))
+                target = _unique_existing_target(os.path.abspath(os.path.join(target_root, filename)))
+                shutil.move(source, target)
+                path_map[source] = target
+                moved_files += 1
+        _remove_empty_tree(old_dir)
+
+    if personal_docs_manager is not None:
+        rename_directory = getattr(personal_docs_manager, "rename_directory", None)
+        if callable(rename_directory):
+            rename_directory(old_dir, new_dir, path_map=path_map)
+
+    rag_result = None
+    if rag_manager is not None:
+        rename_owner = getattr(rag_manager, "rename_owner", None)
+        if callable(rename_owner):
+            rag_result = rename_owner(
+                old_owner,
+                new_owner,
+                path_map=path_map,
+                path_prefixes=[(old_dir, new_dir)],
+            )
+
+    return {
+        "old_dir": old_dir,
+        "new_dir": new_dir,
+        "moved_files": moved_files,
+        "path_map": path_map,
+        "rag_result": rag_result,
+    }
+
+
 def setup_personal_routes(personal_docs_manager, rag_manager, rag_available):
     """
     Setup personal documents related routes.

routes/research_routes.py

@@ -12,8 +12,10 @@ from typing import Optional
 from fastapi import APIRouter, HTTPException, Query, Request
 from fastapi.responses import HTMLResponse, StreamingResponse
 from pydantic import BaseModel, Field
+from core.middleware import INTERNAL_TOOL_USER
 from src.endpoint_resolver import resolve_endpoint
 from src.auth_helpers import _auth_disabled, get_current_user
+from core.auth import RESERVED_USERNAMES
 from src.constants import DEEP_RESEARCH_DIR
 
 _SESSION_ID_RE = re.compile(r"^[a-zA-Z0-9-]{1,128}$")
@@ -385,9 +387,9 @@ def setup_research_routes(research_handler, session_manager=None) -> APIRouter:
         """Launch a research job from the dedicated panel."""
         from src.auth_helpers import require_privilege
         user = require_privilege(request, "can_use_research")
-        if user == "internal-tool":
+        if user == INTERNAL_TOOL_USER:
             tool_owner = (request.headers.get("X-Odysseus-Owner") or "").strip()
-            if tool_owner and tool_owner not in {"internal-tool", "api", "demo", "system"}:
+            if tool_owner and tool_owner not in RESERVED_USERNAMES:
                 auth_mgr = getattr(request.app.state, "auth_manager", None)
                 if auth_mgr is not None and getattr(auth_mgr, "is_configured", False):
                     try:

routes/session_routes.py

@@ -1004,6 +1004,7 @@ def setup_session_routes(session_manager: SessionManager, config: dict, webhook_
         """
         from src.llm_core import llm_call
         user = effective_user(request)
+        single_user_mode = not user and _auth_disabled()
         user_sessions = session_manager.get_sessions_for_user(user)
 
         # Delete empty and throwaway sessions before sorting
@@ -1022,7 +1023,12 @@ def setup_session_routes(session_manager: SessionManager, config: dict, webhook_
         }
         _THROWAWAY_MAX_MESSAGES = 4  # only delete if <= this many messages
         try:
-            rows = db.query(DbSession).filter(DbSession.archived == False, DbSession.owner == user).limit(2000).all()
+            rows_q = db.query(DbSession).filter(DbSession.archived == False)
+            if user:
+                rows_q = rows_q.filter(DbSession.owner == user)
+            elif not single_user_mode:
+                rows_q = rows_q.filter(DbSession.owner == user)
+            rows = rows_q.limit(2000).all()
             folder_map = {r.id: r.folder for r in rows}
             # Precompute per-session message counts in TWO aggregate queries
             # instead of 1–3 queries PER session — with many chats the per-row
@@ -1242,7 +1248,12 @@ def setup_session_routes(session_manager: SessionManager, config: dict, webhook_
         db = SessionLocal()
         try:
             for sid, folder_name in assignments.items():
-                db_session = db.query(DbSession).filter(DbSession.id == sid, DbSession.owner == user).first()
+                db_session_q = db.query(DbSession).filter(DbSession.id == sid)
+                if user:
+                    db_session_q = db_session_q.filter(DbSession.owner == user)
+                elif not single_user_mode:
+                    db_session_q = db_session_q.filter(DbSession.owner == user)
+                db_session = db_session_q.first()
                 if db_session:
                     db_session.folder = folder_name
                     db_session.updated_at = datetime.utcnow()

routes/shell_routes.py

@@ -15,6 +15,7 @@ from collections import namedtuple
 from pathlib import Path
 from typing import Dict, Any
 from core.platform_compat import IS_APPLE_SILICON, which_tool
+from core.middleware import INTERNAL_TOOL_USER
 from src.optional_deps import prepare_optional_dependency_import
 
 # POSIX-only: `pty`/`fcntl` transitively import `termios`, which does NOT exist
@@ -55,7 +56,7 @@ def _require_admin(request: Request):
     # In-process tool loopback. The AuthMiddleware already validated the
     # internal token + loopback client before setting this marker, so
     # honour it here as admin-equivalent.
-    if user == "internal-tool":
+    if user == INTERNAL_TOOL_USER:
         return
     if not user or user == "api":
         raise HTTPException(403, "Admin only")

routes/task_routes.py

@@ -11,6 +11,7 @@ from fastapi import APIRouter, HTTPException, Request
 from pydantic import BaseModel
 
 from core.database import SessionLocal, ScheduledTask, TaskRun
+from core.middleware import INTERNAL_TOOL_USER
 from core.constants import internal_api_base
 from src.auth_helpers import get_current_user
 from src.constants import DATA_DIR, EMAIL_URGENCY_CACHE_DIR
@@ -427,7 +428,7 @@ def setup_task_routes(task_scheduler) -> APIRouter:
         # In-process tool-loopback marker — AuthMiddleware validated
         # the internal token + loopback client before stamping this,
         # so treat as admin-equivalent.
-        if user == "internal-tool":
+        if user == INTERNAL_TOOL_USER:
             return True
         try:
             from core.auth import AuthManager

scripts/odysseus-calendar

@@ -103,9 +103,13 @@ def cmd_list(args) -> None:
     end = _parse_dt(args.end) if args.end else (start + timedelta(days=30))
     db = SessionLocal()
     try:
+        # Overlap semantics, matching the web route (routes/calendar_routes.py)
+        # and the recurring-expansion contract: an event is in the window when
+        # it starts before the window end AND ends after the window start. This
+        # includes multi-day / in-progress events that began before `start`.
         q = db.query(CalendarEvent).filter(
-            CalendarEvent.dtstart >= start,
             CalendarEvent.dtstart < end,
+            CalendarEvent.dtend > start,
         )
         if args.calendar:
             cal = db.query(CalendarCal).filter(CalendarCal.name == args.calendar).first()

services/hwfit/fit.py

@@ -130,6 +130,44 @@ def _lookup_bandwidth(system):
     return None
 
 
+def _canonical_cpu_backend(system):
+    """Return the canonical CPU backend for cpu_only speed estimation.
+
+    Normalizes CPU-architecture aliases separately from the GPU backend, and
+    overrides GPU-only backends (CUDA/ROCm/Metal) so they do not inherit a
+    discrete-GPU fallback constant when the model is actually running on CPU.
+    """
+    backend = (system.get("backend") or "").lower().strip()
+    cpu_arch = (system.get("cpu_arch") or "").lower().strip()
+    cpu_name = (system.get("cpu_name") or "").lower()
+    gpu_name = (system.get("gpu_name") or "").lower()
+
+    # Already-canonical CPU backends
+    if backend in ("cpu_x86", "cpu_arm"):
+        return backend
+
+    # Raw CPU-architecture aliases. Treat plain "arm" as 32-bit ARM, not the
+    # ARM64-class CPU fallback used for Apple Silicon/aarch64 machines.
+    if backend in ("x86_64", "amd64", "i386", "i686"):
+        return "cpu_x86"
+    if backend in ("arm64", "aarch64"):
+        return "cpu_arm"
+
+    # Prefer an explicit CPU architecture field when present
+    if cpu_arch:
+        if cpu_arch in ("x86_64", "amd64", "x86", "i386", "i686"):
+            return "cpu_x86"
+        if cpu_arch in ("arm64", "aarch64"):
+            return "cpu_arm"
+
+    # Apple Silicon enters ranking as backend="metal"; its CPU path is ARM.
+    if backend in ("metal", "mps", "apple") or "apple" in cpu_name or "apple" in gpu_name:
+        return "cpu_arm"
+
+    # Conservative default for CUDA/ROCm/discrete GPU backends and unknowns.
+    return "cpu_x86"
+
+
 def _estimate_speed(model, quant, run_mode, system, offload_frac=0.0):
     """Estimate tok/s. Uses active params for MoE (only active experts run per token).
 
@@ -147,6 +185,11 @@ def _estimate_speed(model, quant, run_mode, system, offload_frac=0.0):
     bw = _lookup_bandwidth(system)
     backend = system.get("backend", "cpu_x86")
 
+    # CPU-only inference must never inherit a GPU backend's fallback constant,
+    # even if the detected system happens to report a CUDA/Metal/ROCm backend.
+    if run_mode == "cpu_only":
+        backend = _canonical_cpu_backend(system)
+
     if bw and run_mode in ("gpu", "cpu_offload"):
         bpp = QUANT_BYTES_PER_PARAM.get(quant, 0.5)
         model_gb = pb * bpp

services/hwfit/hardware.py

@@ -320,7 +320,7 @@ def _detect_apple_silicon():
 
     # Only Apple Silicon (arm64) has a Metal GPU worth serving LLMs on; Intel
     # Macs fall through to the CPU path.
-    if "arm" not in arch and "aarch64" not in arch:
+    if _canonical_cpu_arch(arch) != "arm64":
         return None
 
     # Chip name, e.g. "Apple M4 Max" — carries the Pro/Max/Ultra variant that
@@ -503,6 +503,25 @@ def _get_cpu_count():
     return os.cpu_count() or 1
 
 
+def _canonical_cpu_arch(value):
+    arch = str(value or "").lower().strip().replace("-", "_")
+    if arch in ("x86_64", "amd64", "x64"):
+        return "x86_64"
+    if arch in ("i386", "i686", "x86"):
+        return "x86"
+    if arch in ("arm64", "aarch64"):
+        return "arm64"
+    if arch == "arm" or arch.startswith("armv"):
+        return "arm"
+    return arch
+
+
+def _get_cpu_arch():
+    if _remote_host:
+        return _canonical_cpu_arch(_run(["uname", "-m"]) or "")
+    return _canonical_cpu_arch(platform.machine())
+
+
 def _powershell_exe():
     """Pick the best PowerShell executable for LOCAL execution: prefer pwsh
     (PowerShell 7+), fall back to Windows PowerShell 5.1. Returns an absolute
@@ -528,6 +547,7 @@ def _detect_windows():
         $r.cpu_name = $cpu.Name
         $r.cpu_cores = (Get-CimInstance Win32_Processor | Measure-Object -Property NumberOfLogicalProcessors -Sum).Sum
         $r.arch = $cpu.AddressWidth
+        $r.cpu_arch = if ($env:PROCESSOR_ARCHITEW6432) { $env:PROCESSOR_ARCHITEW6432 } else { $env:PROCESSOR_ARCHITECTURE }
         # GPU detection via nvidia-smi (fastest) or WMI fallback
         try { 
             $nv = nvidia-smi --query-gpu=memory.total,name --format=csv,noheader,nounits 2>$null
@@ -599,6 +619,7 @@ def _detect_windows():
             "available_ram_gb": d.get("avail_gb", 0),
             "cpu_cores": _as_int(d.get("cpu_cores"), 1),
             "cpu_name": _cpu_name,
+            "cpu_arch": _canonical_cpu_arch(d.get("cpu_arch")),
             "has_gpu": bool(d.get("gpu_name")),
             "gpu_name": d.get("gpu_name"),
             "gpu_vram_gb": d.get("gpu_vram_gb"),
@@ -794,6 +815,7 @@ def detect_system(host="", ssh_port="", platform="", fresh=False):
     available_ram = round(_get_available_ram_gb(), 1)
     cpu_cores = _get_cpu_count()
     cpu_name = _get_cpu_name()
+    cpu_arch = _get_cpu_arch()
 
     gpu_info = _detect_apple_silicon() or _detect_nvidia() or _detect_amd()
 
@@ -803,6 +825,7 @@ def detect_system(host="", ssh_port="", platform="", fresh=False):
             "available_ram_gb": available_ram,
             "cpu_cores": cpu_cores,
             "cpu_name": cpu_name,
+            "cpu_arch": cpu_arch,
             "has_gpu": True,
             "gpu_name": gpu_info["gpu_name"],
             "gpu_vram_gb": gpu_info["gpu_vram_gb"],
@@ -817,17 +840,13 @@ def detect_system(host="", ssh_port="", platform="", fresh=False):
             "unified_memory": gpu_info.get("unified_memory", False),
         }
     else:
-        if _remote_host:
+        backend = "cpu_arm" if cpu_arch == "arm64" else "cpu_x86"
-            arch_out = _run(["uname", "-m"]) or ""
-        else:
-            import platform as _platform
-            arch_out = _platform.machine().lower()
-        backend = "cpu_arm" if "aarch64" in arch_out or "arm" in arch_out else "cpu_x86"
         result = {
             "total_ram_gb": total_ram,
             "available_ram_gb": available_ram,
             "cpu_cores": cpu_cores,
             "cpu_name": cpu_name,
+            "cpu_arch": cpu_arch,
             "has_gpu": False,
             "gpu_name": None,
             "gpu_vram_gb": None,

setup.py

@@ -16,8 +16,9 @@ sys.path.insert(0, BASE_DIR)
 from src.constants import (
     DATA_DIR, AUTH_FILE, UPLOAD_DIR, PERSONAL_DIR, PERSONAL_UPLOADS_DIR,
     TTS_CACHE_DIR, GENERATED_IMAGES_DIR, DEEP_RESEARCH_DIR, CHROMA_DIR,
-    RAG_DIR, MEMORY_VECTORS_DIR,
+    RAG_DIR, MEMORY_VECTORS_DIR, PASSWORD_MIN_LENGTH,
 )
+from core.auth import RESERVED_USERNAMES
 
 DIRS = [
     DATA_DIR,
@@ -59,15 +60,23 @@ def _prompt_admin_credentials():
     print("  (Press Enter to accept defaults)")
     print()
 
+    while True:
         username = input("  Username [admin]: ").strip().lower()
         if not username:
             username = "admin"
+        if username in RESERVED_USERNAMES:
+            print(f"  '{username}' is a reserved username. Choose another.")
+            continue
+        break
 
     while True:
         password = getpass.getpass("  Password: ")
         if not password:
             print("  Password cannot be empty.")
             continue
+        if len(password) < PASSWORD_MIN_LENGTH:
+            print(f"  Password must be at least {PASSWORD_MIN_LENGTH} characters.")
+            continue
         confirm = getpass.getpass("  Confirm password: ")
         if password != confirm:
             print("  Passwords don't match. Try again.")
@@ -93,8 +102,13 @@ def create_default_admin():
         password = os.getenv("ODYSSEUS_ADMIN_PASSWORD", "").strip()
 
         if username and password:
-            # Both provided via env — use them directly
+            # Both provided via env — validate before using
-            pass
+            if username in RESERVED_USERNAMES:
+                print(f"  [error] ODYSSEUS_ADMIN_USER '{username}' is a reserved username")
+                return "failed"
+            if len(password) < PASSWORD_MIN_LENGTH:
+                print(f"  [error] ODYSSEUS_ADMIN_PASSWORD must be at least {PASSWORD_MIN_LENGTH} characters")
+                return "failed"
         elif sys.stdin.isatty() and not os.getenv("ODYSSEUS_SKIP_ADMIN_PROMPT"):
             # Interactive terminal — ask the user
             username, password = _prompt_admin_credentials()

specs/architecture-runtime-inventory.md

@@ -0,0 +1,412 @@
+# Architecture Runtime Inventory
+
+> **Purpose**: Phase 0 planning baseline for codebase readability improvements (#4071).
+> **Parent issue**: [#4082](https://github.com/pewdiepie-archdaemon/odysseus/issues/4082)
+> **Last updated**: dev@b58af42 | 2026-06-16
+> **Status**: Draft — to be reviewed before follow-up slices open.
+> **Snapshot basis**: Importer / file / import-line counts are refreshed to `dev@b58af42` (2026-06-16) and are recomputable via the commands in §3.4. **Line counts** in §2.1 / §2.2 are a snapshot from an earlier baseline and drift as `dev` moves — recompute any of them with `wc -l <file>`. This inventory tracks structure and risk, not live metrics.
+
+This document maps the current runtime module structure, identifies high-risk boundaries, and recommends safe first refactor slices. It does **not** move files, change imports, or alter runtime behavior.
+
+---
+
+## 1. Current Structure Overview
+
+### 1.1 Top-Level Layout
+
+```
+odysseus/
+├── app.py                    # FastAPI app entrypoint (1,145 lines)
+├── conf/                     # Configuration (config.py, settings.py, settings_scrub.py)
+├── src/                      # 95 flat .py files + 2 subdirectories
+│   ├── agent_tools/          # Tool helpers: document, filesystem, subprocess, web
+│   └── search/               # Search subsystem
+├── routes/                   # 54 flat .py files — HTTP route handlers
+├── core/                     # 10 files — database models, auth, middleware, session
+├── mcp_servers/              # 5 files — MCP server implementations
+├── scripts/                  # CLI tools and one-shot scripts
+├── static/                   # Frontend HTML/CSS/JS
+├── tests/                    # 583 test files (~54,800 lines)
+└── services/                 # (exists as needed)
+```
+
+### 1.2 Directory Flatness Metric
+
+| Directory | Flat `.py` Files | Subdirectories | Concern |
+|-----------|-----------------|----------------|---------|
+| `src/` | **95** | 2 (`agent_tools/`, `search/`) | No domain grouping; 95 files in one directory |
+| `routes/` | **54** | 0 | All route handlers in one flat directory |
+| `core/` | 10 | 0 | Manageable, but `database.py` is oversized |
+
+---
+
+## 2. Largest Runtime Modules
+
+### 2.1 Python Backend
+
+| Rank | File | Lines | Classes | Functions | Risk |
+|------|------|-------|---------|-----------|------|
+| 1 | `src/tool_implementations.py` | **4,032** | 0 | ~48 | **HIGH** |
+| 2 | `routes/email_routes.py` | **3,245** | — | — | **MEDIUM** |
+| 3 | `routes/cookbook_routes.py` | **2,969** | — | — | **MEDIUM** |
+| 4 | `src/agent_loop.py` | **2,961** | 0 | ~24 | **HIGH** |
+| 5 | `src/task_scheduler.py` | **2,330** | — | 5 | MEDIUM |
+| 6 | `routes/model_routes.py` | **2,266** | — | — | MEDIUM |
+| 7 | `core/database.py` | **2,265** | 28 | ~59 helpers | **HIGH** |
+| 8 | `src/builtin_actions.py` | **2,262** | 2 | ~24 | MEDIUM |
+| 9 | `src/llm_core.py` | **2,164** | — | — | MEDIUM |
+| 10 | `mcp_servers/email_server.py` | 2,197 | — | — | LOW (separate process) |
+| 11 | `src/visual_report.py` | 1,918 | — | — | LOW |
+| 12 | `routes/gallery_routes.py` | 1,896 | — | — | LOW |
+| 13 | `src/ai_interaction.py` | 1,846 | — | — | MEDIUM |
+| 14 | `routes/document_routes.py` | 1,717 | — | — | LOW |
+| 15 | `routes/skills_routes.py` | 1,648 | — | — | LOW |
+
+**Heuristic**: Files > 2,000 lines with 20+ public symbols and many importers are the highest-risk splits. Files 1,000–2,000 lines are medium-risk if tightly coupled.
+
+### 2.2 Frontend
+
+| File | Lines | Concern |
+|------|-------|---------|
+| `static/style.css` | **36,653** | Entire app CSS in one file (tracked separately in #2617) |
+| `static/js/document.js` | **9,776** | Single JS file for document functionality |
+| `static/js/slashCommands.js` | 6,498 | |
+| `static/js/settings.js` | 5,266 | |
+| `static/js/emailLibrary.js` | 5,217 | |
+| `static/js/notes.js` | 5,124 | |
+| `static/js/chat.js` | 4,985 | |
+| `static/app.js` | 4,090 | |
+
+**Note**: Frontend modularization is tracked separately in #2617 (CSS) and is not the focus of this Phase 0 inventory. Frontend is listed here for completeness but follow-up slices should target Python backend boundaries first.
+
+---
+
+## 3. Import Dependency Graph
+
+### 3.1 Who Depends on `core/database.py`
+
+**102 files** import from `core.database` — this is the most depended-upon module:
+
+- All route handlers (`routes/*.py`)
+- Most `src/*.py` files
+- `core/session_manager.py`, `core/auth.py`
+- Multiple test files
+
+**Implication**: Any split of `core/database.py` is the highest-risk refactor. It should be tackled **last**, never first.
+
+### 3.2 Who Depends on `src/tool_implementations.py`
+
+**17 files** import from `src.tool_implementations`:
+- `src/agent_loop.py`, `src/builtin_actions.py`, `src/tool_index.py`
+- `src/task_scheduler.py`, `src/tool_policy.py`
+- Various tests
+
+### 3.3 Who Depends on `src/agent_loop.py`
+
+**22 files** import from `src.agent_loop`:
+
+- `src/tool_policy.py`, `src/teacher_escalation.py`, `src/bg_monitor.py`
+- `src/task_scheduler.py`
+- Multiple test files
+
+### 3.4 Cross-Layer Import Violations
+
+**`src/` importing from `routes/`** (backwards dependency — domain logic depending on HTTP layer):
+
+```
+src/tool_implementations.py ──→ routes/calendar_routes.py
+src/tool_implementations.py ──→ routes/cookbook_helpers.py
+src/tool_implementations.py ──→ routes/email_helpers.py
+src/tool_implementations.py ──→ routes/email_pollers.py
+src/tool_implementations.py ──→ routes/email_routes.py
+src/tool_implementations.py ──→ routes/model_routes.py
+src/tool_implementations.py ──→ routes/note_routes.py
+src/tool_implementations.py ──→ routes/prefs_routes.py
+```
+
+> These are **runtime imports** (inside function bodies, not at module top), which mitigates circular import risk but indicates fuzzy layer boundaries. Function-level inline imports from the HTTP layer into business logic are a code smell.
+
+**Import counts (top-level)**:
+| Direction | Count | Notes |
+|-----------|-------|-------|
+| `routes/` → `src/` | **374** | Expected: HTTP handlers call domain logic |
+| `routes/` → `core/` | **126** | Expected: handlers access DB models |
+| `src/` → `routes/` | **31** | **Unexpected**: domain logic reaching into HTTP layer (direct grep of import lines referencing `routes/`) |
+| `src/` → `core/` | **106** | Acceptable but could be reduced with a data-access layer |
+
+> **How the metrics in this document are computed** — recompute against current `dev` before treating any count as authoritative (the tree drifts; these numbers are a snapshot, not a live value):
+> - `src/` flat `.py` files: `find src -maxdepth 1 -name '*.py' | wc -l`
+> - `tests/` test files: `find tests -name 'test_*.py' | wc -l`
+> - `core.database` importers: `grep -rlE '(from|import) +core\.database' --include='*.py' . | grep -v core/database.py | wc -l`
+> - `src.agent_loop` importers: `grep -rlE '(from|import) +src\.agent_loop' --include='*.py' . | grep -v src/agent_loop.py | wc -l`
+> - Cross-layer import lines: `grep -rhE '(from|import) +<pkg>' --include='*.py' <dir>/ | wc -l` (e.g. `(from|import) +routes` over `src/`)
+
+---
+
+## 4. Route Ownership Map
+
+Routes can be grouped into logical feature domains. Current flat structure obscures these boundaries:
+
+| Domain | Route Files | Total Lines | Review Complexity |
+|--------|-------------|-------------|-------------------|
+| **Email** | `email_routes.py`, `email_helpers.py`, `email_pollers.py` | 5,936 | HIGH — most complex domain |
+| **Chat / Agent** | `chat_routes.py`, `chat_helpers.py`, `shell_routes.py`, `codex_routes.py`, `skills_routes.py` | 6,365 | HIGH — core interaction surface |
+| **Cookbook** | `cookbook_routes.py`, `cookbook_helpers.py`, `cookbook_output.py` | 4,110 | MEDIUM |
+| **Model / LLM** | `model_routes.py`, `assistant_routes.py`, `copilot_routes.py` | 2,764 | MEDIUM |
+| **Calendar / Contacts** | `calendar_routes.py`, `contacts_routes.py` | 2,336 | MEDIUM |
+| **Documents** | `document_routes.py`, `document_helpers.py` | 1,954 | LOW |
+| **Auth** | `auth_routes.py`, `api_token_routes.py`, `device_flow.py` | 1,171 | LOW |
+| **Tasks** | `task_routes.py` (standalone) | 1,157 | LOW |
+| **Session** | `session_routes.py` (standalone) | 1,287 | LOW |
+| **Gallery** | `gallery_routes.py`, `gallery_helpers.py` | 1,896 | LOW |
+| **Memory** | `memory_routes.py` | — | LOW |
+| **Research** | `research_routes.py` | — | LOW |
+| **MCP** | `mcp_routes.py` | — | LOW |
+| **Notes** | `note_routes.py` | — | LOW |
+| **Other** | `prefs_routes.py`, `upload_routes.py`, `vault_routes.py`, `webhook_routes.py`, `workspace_routes.py`, `search_routes.py`, `history_routes.py`, `hwfit_routes.py`, `preset_routes.py`, `signature_routes.py`, `backup_routes.py`, `cleanup_routes.py`, `diagnostics_routes.py`, `embedding_routes.py`, `emoji_routes.py`, `font_routes.py`, `stt_routes.py`, `tts_routes.py`, `compare_routes.py`, `personal_routes.py`, `editor_draft_routes.py`, `admin_wipe_routes.py`, `chatgpt_subscription_routes.py` | 2,000+ | LOW individual, HIGH cumulative |
+
+---
+
+## 5. Tool Registry & Implementation Boundaries
+
+### 5.1 Current Tool Architecture
+
+| Component | File | Lines | Role |
+|-----------|------|-------|------|
+| Tool schemas | `src/tool_schemas.py` | 1,392 | JSON Schema tool definitions (Duck-TypedDict) |
+| Tool index | `src/tool_index.py` | 542 | RAG-based tool retrieval from ChromaDB |
+| Tool implementations | `src/tool_implementations.py` | 4,032 | 33 `do_*` functions — all tool execution logic |
+| Tool security | `src/tool_security.py` | — | Owner-scoped tool blocking |
+| Tool policy | `src/tool_policy.py` | — | Guide-only directive, plan-mode disabled tools |
+| Tool utils | `src/tool_utils.py` | — | Shared tool helpers |
+
+### 5.2 Tool Implementation Categories
+
+The 33 `do_*` functions in `tool_implementations.py` fall into natural domain groups — the basis for slice 1's split in §6.2:
+
+| Category | `do_*` functions | Count |
+|----------|------------------|-------|
+| **System / config** | `do_manage_skills`, `do_manage_tasks`, `do_manage_endpoints`, `do_manage_mcp`, `do_manage_webhooks`, `do_manage_tokens`, `do_manage_settings`, `do_api_call`, `do_app_api` | 9 |
+| **Cookbook / model serving** | `do_download_model`, `do_serve_model`, `do_list_served_models`, `do_stop_served_model`, `do_tail_serve_output`, `do_list_downloads`, `do_cancel_download`, `do_search_hf_models`, `do_adopt_served_model`, `do_list_cookbook_servers`, `do_list_serve_presets`, `do_serve_preset`, `do_list_cached_models` | 13 |
+| **Notes** | `do_manage_notes` | 1 |
+| **Calendar** | `do_manage_calendar` | 1 |
+| **Search** | `do_search_chats` | 1 |
+| **Research** | `do_manage_research`, `do_trigger_research` | 2 |
+| **Contacts** | `do_resolve_contact`, `do_manage_contact` | 2 |
+| **Vault** | `do_vault_search`, `do_vault_get`, `do_vault_unlock` | 3 |
+| **Image** | `do_edit_image` | 1 |
+| | **Total** | **33** |
+
+> Low-level tools (filesystem, subprocess, web fetch, document parsing) live in `src/agent_tools/`, **not** in `tool_implementations.py` — out of scope for this split.
+
+---
+
+## 6. Risk Assessment & Candidate Slice Ranking
+
+> **Candidate proposals, not a committed plan.** The rankings, package shapes (e.g. `src/pkg/`, `src/domain/`, `src/infra/`, `src/api/`), split ordering, and route-grouping strategy below are **options for maintainer discussion**. Per #4082/#4071, slice ownership and order are settled by maintainers before any follow-up PR. §1–§3 above are the factual current-state inventory.
+
+### 6.1 Risk Scale
+
+| Level | Criteria |
+|-------|----------|
+| **LOW** | File has ≤3 importers AND ≤500 lines, OR is a pure refactor with clear boundaries |
+| **MEDIUM** | File has 4–15 importers OR 500–1,500 lines |
+| **HIGH** | File has 16+ importers OR >2,000 lines, OR has cross-layer import violations |
+
+### 6.2 Ranked Split Candidates
+
+| Priority | Target | Risk | Rationale |
+|----------|--------|------|-----------|
+| **1** | `src/tool_implementations.py` → `src/tools/*.py` | **MEDIUM** | 4,032 lines → ~10 files by tool category. Already has natural boundaries. 17 importers, tracked in #3629. Use `__init__.py` shim to keep existing imports working. |
+| **2** | `routes/` → domain subdirectories (one domain per PR) | **MEDIUM** | 54 flat files. Done **one domain at a time** (e.g. a standalone PR for the email domain, then chat, …), not a broad reorganization — route modules carry helper imports, registration assumptions, and test import paths. |
+| **3** | `src/agent_loop.py` → `src/agent/loop.py` + submodules | **MEDIUM-HIGH** | 2,961 lines, 24 functions. Can extract prompt building, classification, verification, and runaway detection. Tracked in #3266. |
+| **4** | `src/` → `src/pkg/`, `src/domain/`, `src/infra/`, `src/api/` | **MEDIUM** | Structural reorganization. Split flat `src/` into layered packages. Must come after routes and tools are stable. |
+| **5** | `routes/email_*.py` consolidation | **LOW** | Already grouped by filename prefix. Low-risk cleanup within the email domain. |
+| **6** | `core/database.py` → `src/infra/database/models/*.py` | **HIGH** | 28 classes, 102 importers. Highest-risk split. Must be **last** in any sequence. Requires careful import shim strategy. |
+| **7** | Frontend CSS modularization | **MEDIUM** | 36,653 lines. Tracked in #2617. Separate timeline from backend work. |
+| **8** | Frontend JS modularization | **MEDIUM** | 9,776 lines in `document.js`. Introduce ES modules at minimum. |
+
+### 6.3 Candidate First 3 Behavior-Preserving Slices
+
+**Slice 1: Split `tool_implementations.py`** (Lowest-risk high-impact)
+
+- Create `src/tools/` package with one file per tool category
+- Add `src/tools/__init__.py` re-exporting all symbols with current names
+- Update 17 importers to use new paths (can be deferred via shim)
+- Validation: `python -m pytest tests/ -x -q` + manual smoke test of tool execution
+- Reference: #3629
+
+**Slice 2: Group `routes/` by domain** (one domain per PR, not a broad sweep)
+
+Route modules carry helper imports, router registration assumptions, and test import paths, so this must be done **one domain at a time** rather than as a single reorganization PR. Example sequence (each its own PR):
+
+- PR 2a: move the **email** domain (`email_routes.py`, `email_helpers.py`, `email_pollers.py`) → `routes/email/` + shim
+- PR 2b: move the **chat/agent** domain → `routes/chat/` + shim
+- PR 2c: move the **cookbook** domain → `routes/cookbook/` + shim
+- …and so on per domain from §4
+
+Each PR: add `__init__.py` re-exporting old names, update `app.py` router imports, validation `python app.py` starts clean. **No behavior change** — pure file reorganization.
+
+**Slice 3: Extract `agent_loop.py` submodules** (Improve reviewability)
+
+- Move prompt assembly → `src/agent/prompt.py`
+- Move request classification → `src/agent/classifier.py`
+- Move sub-agent verification → `src/agent/verifier.py`
+- Move runaway detection → `src/agent/runaway.py`
+- Move context management → `src/agent/context.py`
+- Keep `src/agent/loop.py` as the main orchestration module
+- Validation: `python -m pytest tests/test_agent_loop.py tests/test_loop_breaker_runaway.py -v`
+
+---
+
+## 7. Safety Guardrails for Follow-Up Work
+
+Per maintainer guidance in #4082 and #4071:
+
+- [ ] **One domain/slice per PR** — never mix multiple reorganizations
+- [ ] **No behavior changes** mixed with file moves — pure reorganization only
+- [ ] **Keep compatibility shims** — `__init__.py` re-exports for all existing import paths
+- [ ] **Add or identify focused tests** before risky splits
+- [ ] **Do not start with `core/database.py`** or broad route movement unless this inventory shows a safe boundary
+- [ ] **Prefer small, reviewable slices** over large restructures
+- [ ] **No packaging/runtime/tooling migration** mixed into file moves
+- [ ] **No frontend framework migration** inside this stabilization lane
+- [ ] **Validate with `python -m compileall`** — every PR must pass CI checks
+- [ ] **Validate with `pytest`** — run the full test suite before opening each PR
+
+---
+
+## 8. Validation Commands
+
+Each follow-up PR should be verifiable with these commands before submission:
+
+```bash
+# Syntax check — must pass with zero errors
+python -m compileall src/ routes/ core/ conf/
+
+# Full test suite — must match baseline pass rate
+python -m pytest tests/ -x -q
+
+# Import shim verification — existing import paths must still work
+python -c "from src.tool_implementations import do_search_chats; print('OK')"
+
+# App startup smoke test (if backend touched)
+timeout 5 python app.py 2>&1 | head -5 || true
+```
+
+---
+
+## 9. Open Questions
+
+1. Is `#2538` (specs ground truth) the canonical behavior map baseline, and should this inventory be kept in sync with those specs once merged?
+2. Should route grouping follow the domain map proposed here, or is there a different taxonomy preferred by maintainers?
+3. For the `tool_implementations.py` split (#3629), is the tool categorization in §5.2 acceptable, or should it follow a different grouping?
+4. Should compatibility shims (`__init__.py`) be temporary (removed in a follow-up wave) or permanent?
+5. Should an ADR (Architecture Decision Record) document be started to track decisions made during this process?
+
+---
+
+## 10. Future Direction (NOT current state)
+
+The following are **future refactor targets** (candidate directions **pending maintainer agreement**, not committed), recorded here so this inventory does not imply they exist today. None of them are present in the current `dev` tree:
+
+- `main.py` — proposed rename of the `app.py` entrypoint. Today the app boots via `app.py`.
+- `src/agent/` — proposed package to hold `agent_loop.py` submodules (prompt/classifier/verifier/runaway/context). Today `agent_loop.py` is a single flat file in `src/`.
+- `src/infra/`, `src/domain/`, `src/pkg/`, `src/api/` — proposed layered reorganization of the flat `src/` directory (slice 4 in §6).
+
+These become real only when the corresponding slices land.
+
+---
+
+## Appendix A: File Listing
+
+### `src/` (95 files — 61 shown; run `ls src/*.py` for the full list)
+
+```
+agent_loop.py          tool_implementations.py   tool_schemas.py
+tool_index.py          tool_security.py          tool_policy.py
+tool_utils.py          builtin_actions.py        task_scheduler.py
+llm_core.py            model_context.py          model_discovery.py
+session_search.py      context_budget.py         context_compactor.py
+ai_interaction.py      action_intents.py         agent_runs.py
+app_helpers.py         app_initializer.py        config.py
+database.py            memory.py                 memory_provider.py
+secret_storage.py      prompt_security.py        url_security.py
+url_safety.py          rate_limiter.py           cleanup_service.py
+readiness.py           service_health.py         exceptions.py
+request_models.py      assistant_log.py          bg_monitor.py
+builtin_mcp.py         chat_helpers.py           chroma_client.py
+document_processor.py  embedding_lanes.py        deep_research.py
+research_handler.py    research_utils.py         personal_docs.py
+rag_manager.py         rag_singleton.py          topic_analyzer.py
+visual_report.py       youtube_handler.py        pdf_forms.py
+pdf_form_doc.py        pdf_runtime.py            caldav_writeback.py
+email_thread_parser.py text_helpers.py           user_time.py
+teacher_escalation.py  cookbook_serve_lifecycle.py
+chatgpt_subscription.py  mcp_manager.py
+```
+
+### `routes/` (54 files)
+
+```
+__init__.py    _validators.py
+auth_routes.py              api_token_routes.py       device_flow.py
+chat_routes.py              chat_helpers.py           shell_routes.py
+codex_routes.py             skills_routes.py
+email_routes.py             email_helpers.py          email_pollers.py
+cookbook_routes.py          cookbook_helpers.py       cookbook_output.py
+model_routes.py             assistant_routes.py       copilot_routes.py
+calendar_routes.py          contacts_routes.py
+document_routes.py          document_helpers.py
+gallery_routes.py           gallery_helpers.py
+task_routes.py              session_routes.py
+note_routes.py              memory_routes.py          research_routes.py
+mcp_routes.py               search_routes.py          history_routes.py
+webhook_routes.py           workspace_routes.py       upload_routes.py
+vault_routes.py             prefs_routes.py           preset_routes.py
+signature_routes.py         personal_routes.py        hwfit_routes.py
+backup_routes.py            cleanup_routes.py         diagnostics_routes.py
+embedding_routes.py         emoji_routes.py           font_routes.py
+stt_routes.py               tts_routes.py             compare_routes.py
+editor_draft_routes.py      chatgpt_subscription_routes.py    admin_wipe_routes.py
+```
+
+### `core/` (10 files)
+
+```
+__init__.py    constants.py    database.py    models.py
+auth.py        middleware.py   session_manager.py   exceptions.py
+atomic_io.py   platform_compat.py
+```
+
+---
+
+## Appendix B: Key Import Relationships
+
+```
+core/database.py  ←── 102 importers (routes/*, src/*, core/*, tests/*)
+    ↑
+    ├── routes/auth_routes.py
+    ├── routes/email_routes.py
+    ├── src/builtin_actions.py
+    ├── src/task_scheduler.py
+    ├── src/tool_implementations.py (inline)
+    └── ...97 more
+
+src/tool_implementations.py  ←── 17 importers
+    ↑
+    ├── src/agent_loop.py
+    ├── src/builtin_actions.py
+    ├── src/tool_index.py
+    ├── src/task_scheduler.py
+    ├── src/tool_policy.py
+    └── ...12 more (mostly tests)
+
+src/agent_loop.py  ←── 22 importers
+    ↑
+    ├── src/tool_policy.py
+    ├── src/teacher_escalation.py
+    ├── src/bg_monitor.py
+    ├── src/task_scheduler.py
+    └── 18 more (incl. tests)
+```

src/agent_loop.py

@@ -267,6 +267,10 @@ _DOMAIN_RULES = {
 - Use `resolve_contact` to look up a contact's email or phone number by name. Searches the CardDAV address book and sent email history.
 - Use `manage_contact` to list, add, update, or delete contacts in the address book.
 - Do NOT use `manage_memory` for contact lookups — contact details live in the address book, not memory.""",
+    "integrations": """\
+## Integration/API rules
+- To query or control a configured service integration (Home Assistant, Miniflux, Gitea, Linkding, Jellyfin, or any other registered service), use `api_call` with the integration name, HTTP method, path, and optional JSON body.
+- Do not use shell, curl, or `app_api` to reach a user's connected integration when `api_call` is available.""",
 }
 
 _DOMAIN_TOOL_MAP = {
@@ -277,9 +281,10 @@ _DOMAIN_TOOL_MAP = {
     "notes_calendar_tasks": {"manage_notes", "manage_calendar", "manage_tasks"},
     "ui": {"ui_control"},
     "sessions": {"create_session", "list_sessions", "manage_session", "send_to_session", "search_chats"},
-    "files": {"bash", "python", "read_file", "write_file", "edit_file", "grep", "glob", "ls", "get_workspace"},
+    "files": {"bash", "python", "read_file", "write_file", "edit_file", "grep", "glob", "ls", "get_workspace", "manage_bg_jobs"},
     "settings": {"manage_settings", "manage_endpoints", "manage_mcp", "manage_webhooks", "manage_tokens", "app_api"},
     "contacts": {"resolve_contact", "manage_contact"},
+    "integrations": {"api_call"},
 }
 
 def _domain_rules_for_tools(tool_names: set) -> list[str]:
@@ -811,10 +816,25 @@ def _classify_agent_request(messages: List[Dict], last_user: str) -> Dict[str, o
         domains.add("sessions")
     if has(r"\b(file|folder|directory|repo|git|grep|find in files|read file|edit file|shell|terminal|bash|python)\b"):
         domains.add("files")
+    # Managing detached bash jobs: "kill the background job", "stop the job",
+    # "kill that job", "check the job output", "is the bg job done".
+    if (has(r"\b(background|bg)\s+(jobs?|task)\b")
+            or has(r"\b(kill|stop|cancel|terminate|check|tail|show|list)\b.{0,16}\bjobs?\b")
+            or has(r"\bjobs?\b.{0,16}\b(output|status|done|finished|running)\b")):
+        domains.add("files")
     if has(r"\b(endpoint|api token|mcp|webhook|preference|configure|config|setting)\b"):
         domains.add("settings")
     if has(r"\b(contact|contacts|phone|phone number|address book|vcard)\b"):
         domains.add("contacts")
+    # API-integration intent — calling a configured service via the api_call
+    # tool. Without this the #3794 repro ("Use the api_call tool to call Home
+    # Assistant GET /api/states") matched no domain, classified as low-signal,
+    # and the tool never reached the schema filter. Detect it explicitly so the
+    # "integrations" domain seeds api_call deterministically (see
+    # _DOMAIN_TOOL_MAP), independent of embedding retrieval.
+    if has(r"\bapi[ _]call\b", r"\bintegrations?\b",
+           r"\b(?:home ?assistant|miniflux|gitea|linkding|jellyfin)\b"):
+        domains.add("integrations")
 
     low_signal = not continuation and not domains
     return {
@@ -843,8 +863,11 @@ def _recent_context_for_retrieval(messages: List[Dict], max_user: int = 3, max_c
         if isinstance(content, list):
             content = " ".join(b.get("text", "") for b in content if isinstance(b, dict))
         content = (content or "").strip()
-        # Skip injected tool-result envelopes — role=user but not human intent.
+        # Skip injected envelopes — role=user but not human intent. Tool results
-        if not content or content.startswith("[Tool execution results]"):
+        # are now wrapped via untrusted_context_message (metadata.trusted=False);
+        # keep the legacy "[Tool execution results]" prefix for older histories.
+        meta = msg.get("metadata") or {}
+        if not content or meta.get("trusted") is False or content.startswith("[Tool execution results]"):
             continue
         collected.append(content)
         if len(collected) >= max_user:
@@ -1562,8 +1585,14 @@ def _append_tool_results(
         if round_reasoning:
             msg["reasoning_content"] = round_reasoning
         messages.append(msg)
+        # Tool output (shell/python stdout, file reads, fetched pages, email
+        # bodies, MCP results) is sourced from outside the server. Wrap it as
+        # untrusted data so prompt-injection inside a tool result is treated as
+        # data, not instructions — same hardening as skills (#788) and the
+        # web/RAG context. THREAT_MODEL.md lists tool output as a surface that
+        # must go through untrusted_context_message.
         messages.append(
-            {"role": "user", "content": f"[Tool execution results]\n\n{tool_output_text}"}
+            untrusted_context_message("tool execution results", tool_output_text)
         )
 
 

src/agent_tools/__init__.py

@@ -22,6 +22,9 @@ from .subprocess_tools import BashTool, PythonTool
 from .web_tools import WebSearchTool, WebFetchTool
 from .filesystem_tools import ReadFileTool, WriteFileTool, EditFileTool, LsTool, GlobTool, GrepTool, GetWorkspaceTool
 from .document_tools import CreateDocumentTool, UpdateDocumentTool, EditDocumentTool, SuggestDocumentTool, ManageDocumentTool
+from .model_interaction_tools import ChatWithModelTool, AskTeacherTool, ListModelsTool
+from .bg_job_tools import ManageBgJobsTool
+from .session_tools import CreateSessionTool, ListSessionsTool, SendToSessionTool, ManageSessionTool
 
 TOOL_HANDLERS = {
     "bash": BashTool().execute,
@@ -40,6 +43,14 @@ TOOL_HANDLERS = {
     "suggest_document": SuggestDocumentTool().execute,
     "manage_documents": ManageDocumentTool().execute,
     "get_workspace": GetWorkspaceTool().execute,
+    "chat_with_model": ChatWithModelTool().execute,
+    "ask_teacher": AskTeacherTool().execute,
+    "list_models": ListModelsTool().execute,
+    "manage_bg_jobs": ManageBgJobsTool().execute,
+    "create_session": CreateSessionTool().execute,
+    "list_sessions": ListSessionsTool().execute,
+    "send_to_session": SendToSessionTool().execute,
+    "manage_session": ManageSessionTool().execute,
 }
 
 # ---------------------------------------------------------------------------
@@ -52,7 +63,7 @@ PYTHON_TIMEOUT = 30
 
 # Tool types that trigger execution
 TOOL_TAGS = {"bash", "python", "web_search", "web_fetch", "read_file", "write_file", "edit_file",
-             "grep", "glob", "ls", "get_workspace",
+             "grep", "glob", "ls", "get_workspace", "manage_bg_jobs",
              "create_document", "update_document", "edit_document",
              "search_chats",
              "chat_with_model", "create_session", "list_sessions",

src/agent_tools/bg_job_tools.py

@@ -0,0 +1,98 @@
+"""Agent tool to inspect and control detached background `bash` jobs.
+
+`bash` blocks prefixed with a `#!bg` marker run detached via `src.bg_jobs`; the
+agent is auto-re-invoked with the output when they finish. This tool covers the
+gaps in that flow: list the jobs in the current chat, read a still-running job's
+output on demand, and kill a runaway job instead of waiting out its max-runtime.
+
+Registry tool (`TOOL_HANDLERS["manage_bg_jobs"]`). Jobs are scoped to the chat
+that launched them, so every action requires the caller's `session_id` and a job
+from another session is treated as not found.
+"""
+
+import json
+import time
+from typing import Any, Dict, List
+
+_LIST_ACTIONS = {"list", "ls", "jobs"}
+_OUTPUT_ACTIONS = {"output", "get", "read", "tail", "status", "show"}
+_KILL_ACTIONS = {"kill", "stop", "cancel", "terminate"}
+
+
+def _age(rec: Dict[str, Any]) -> str:
+    start = rec.get("started_at")
+    if not start:
+        return "?"
+    secs = int(time.time() - start)
+    if secs < 60:
+        return f"{secs}s"
+    if secs < 3600:
+        return f"{secs // 60}m"
+    return f"{secs // 3600}h{(secs % 3600) // 60}m"
+
+
+def _status_label(rec: Dict[str, Any]) -> str:
+    status = rec.get("status", "?")
+    if rec.get("killed"):
+        return "killed"
+    if rec.get("timed_out"):
+        return "timed out"
+    if rec.get("died"):
+        return "died"
+    if status in ("done", "failed"):
+        return f"{status} (exit {rec.get('exit_code')})"
+    return status
+
+
+def _row(rec: Dict[str, Any]) -> str:
+    cmd = (rec.get("command") or "").strip().splitlines()[0][:80]
+    return f"[{rec.get('id')}] {_status_label(rec)} | {_age(rec)} | {cmd}"
+
+
+class ManageBgJobsTool:
+    async def execute(self, content: str, ctx: dict) -> dict:
+        from src import bg_jobs
+
+        session_id = ctx.get("session_id")
+        raw = (content or "").strip()
+        try:
+            args = json.loads(raw) if raw else {}
+        except (ValueError, TypeError):
+            args = {}
+        if not isinstance(args, dict):
+            args = {}
+        action = str(args.get("action", "list")).strip().lower()
+        job_id = str(args.get("job_id") or args.get("id") or "").strip()
+
+        if not session_id:
+            return {"error": "manage_bg_jobs: no active chat session; background jobs are scoped to a chat.", "exit_code": 1}
+
+        if action in _LIST_ACTIONS:
+            jobs: List[Dict[str, Any]] = bg_jobs.list_for_session(session_id)
+            if not jobs:
+                return {"output": "No background jobs in this chat.", "exit_code": 0}
+            jobs.sort(key=lambda r: r.get("started_at") or 0, reverse=True)
+            lines = "\n".join(_row(r) for r in jobs)
+            return {"output": f"{len(jobs)} background job(s):\n{lines}", "exit_code": 0}
+
+        if action in _OUTPUT_ACTIONS or action in _KILL_ACTIONS:
+            if not job_id:
+                return {"error": f"manage_bg_jobs: action '{action}' requires a job_id (see action='list').", "exit_code": 1}
+            rec = bg_jobs.get(job_id)
+            # Scope: only the chat that launched a job may see or control it.
+            if rec is None or rec.get("session_id") != session_id:
+                return {"error": f"manage_bg_jobs: no background job '{job_id}' in this chat.", "exit_code": 1}
+
+            if action in _KILL_ACTIONS:
+                if rec.get("status") != "running":
+                    return {"output": f"Job `{job_id}` already {_status_label(rec)}; nothing to kill.", "exit_code": 0}
+                killed = bg_jobs.kill(job_id)
+                return {"output": f"Killed background job `{job_id}` ({(killed or {}).get('command', '').splitlines()[0][:80]}).", "exit_code": 0}
+
+            out = rec.get("output") or "(no output yet)"
+            return {
+                "output": f"Job `{job_id}` [{_status_label(rec)}, {_age(rec)}]\nCommand: {rec.get('command')}\n\nOutput:\n{out}",
+                "exit_code": 0,
+            }
+
+        return {"error": f"manage_bg_jobs: unknown action '{action}'. Use list, output, or kill.", "exit_code": 1}

src/agent_tools/filesystem_tools.py

@@ -1,6 +1,7 @@
 import asyncio
 import json
 import os
+import re
 import difflib
 import fnmatch
 import shutil
@@ -16,6 +17,31 @@ _CODENAV_SKIP_DIRS = frozenset({
 _CODENAV_MAX_HITS = 200
 _CODENAV_MAX_LINE = 400
 
+
+def _glob_to_regex(pat: str) -> "re.Pattern":
+    """Translate a forward-slash glob (**, *, ?) into a compiled regex.
+    `**/` matches zero or more complete directories.
+    `*` matches within a single path segment (does not cross /).
+    """
+    i, n, out = 0, len(pat), []
+    while i < n:
+        if pat[i : i + 3] == "**/":
+            out.append("(?:[^/]+/)*")
+            i += 3
+        elif pat[i : i + 2] == "**":
+            out.append(".*")
+            i += 2
+        elif pat[i] == "*":
+            out.append("[^/]*")
+            i += 1
+        elif pat[i] == "?":
+            out.append("[^/]")
+            i += 1
+        else:
+            out.append(re.escape(pat[i]))
+            i += 1
+    return re.compile("".join(out))
+
 def _unified_diff(old: str, new: str, path: str) -> Optional[Dict[str, Any]]:
     if old == new:
         return None
@@ -259,23 +285,38 @@ class GlobTool:
             return {"error": f"glob: {e}", "exit_code": 1}
 
         def _glob():
-            from pathlib import Path
+            base = os.path.abspath(root)
-            base = Path(root)
+            if not os.path.isdir(base):
-            if not base.is_dir():
                 return None, f"glob: {root}: not a directory"
+            norm_pat = pattern.replace("\\", "/")
+            # Fast path: literal pattern (no wildcards) → direct path lookup.
+            if not any(c in norm_pat for c in "*?["):
+                cand = os.path.normpath(os.path.join(base, norm_pat))
+                if os.path.exists(cand):
+                    return [cand], None
+                # Literal not at exact path — fall through to walk so
+                # e.g. "foo.py" still matches at any depth (like rglob).
+            # Compile glob to regex: * stays within one segment, **/ spans dirs.
+            regex = _glob_to_regex(norm_pat)
             matched = []
+            cap = _CODENAV_MAX_HITS * 5
             try:
-                for p in base.rglob(pattern):
+                for dp, dns, fns in os.walk(base):
-                    if set(p.relative_to(base).parts) & _CODENAV_SKIP_DIRS:
+                    # Prune skipped dirs before descending (unlike rglob which
-                        continue
+                    # descends first then filters — fatal on large node_modules).
+                    dns[:] = [d for d in dns if d not in _CODENAV_SKIP_DIRS]
+                    for name in fns + dns:
+                        full = os.path.join(dp, name)
+                        rel = os.path.relpath(full, base).replace(os.sep, "/")
+                        if regex.fullmatch(rel) or regex.fullmatch(name):
                             try:
-                        mtime = p.stat().st_mtime
+                                mtime = os.stat(full).st_mtime
                             except OSError:
                                 mtime = 0
-                    matched.append((mtime, str(p)))
+                            matched.append((mtime, full))
-                    if len(matched) > _CODENAV_MAX_HITS * 5:
+                    if len(matched) > cap:
                         break
-            except (OSError, ValueError) as _e:
+            except OSError as _e:
                 return None, f"glob: {_e}"
             matched.sort(key=lambda t: t[0], reverse=True)
             return [pth for _, pth in matched[:_CODENAV_MAX_HITS]], None

src/agent_tools/model_interaction_tools.py

@@ -0,0 +1,208 @@
+"""model_interaction_tools.py - agent tools for talking to other models.
+
+Owns the model-interaction tool implementations (chat_with_model, ask_teacher,
+list_models) and their handler classes, registered in ``TOOL_HANDLERS``. Part
+of the tool -> registry migration (#3629): the implementations were moved here
+out of ``src.ai_interaction`` so dispatch flows through the registry instead of
+the elif chain / dispatch_ai_tool in tool_execution.py.
+
+Shared helpers that still live in ``src.ai_interaction`` and are used by tools
+not yet migrated (``_resolve_model``, ``AI_CHAT_TIMEOUT``) are imported lazily
+inside the functions to avoid an import cycle at module load.
+"""
+import logging
+from typing import Dict, Optional
+
+logger = logging.getLogger(__name__)
+
+
+_TEACHER_SYSTEM_PROMPT = (
+    "You are a senior AI mentor. A less capable model is stuck on a problem and asking for help. "
+    "Provide clear, actionable guidance:\n"
+    "1. Brief analysis of the problem\n"
+    "2. Recommended approach (step by step)\n"
+    "3. Key things to watch out for\n\n"
+    "Be concise and practical. No preamble."
+)
+
+
+async def chat_with_model(content: str, session_id: Optional[str] = None, owner: Optional[str] = None) -> Dict:
+    """Send a message to a specific model and return its response.
+
+    Content format:
+      Line 1: model_name (or model_name@endpoint_name)
+      Line 2+: the message to send
+    """
+    from src.ai_interaction import _resolve_model, AI_CHAT_TIMEOUT
+    from src.llm_core import llm_call_async
+
+    lines = content.strip().split("\n", 1)
+    if not lines or not lines[0].strip():
+        return {"error": "First line must be the model name"}
+
+    model_spec = lines[0].strip()
+    message = lines[1].strip() if len(lines) > 1 else ""
+    if not message:
+        return {"error": "No message provided (line 2+ is the message)"}
+
+    try:
+        url, model, headers = _resolve_model(model_spec, owner=owner)
+    except ValueError as e:
+        return {"error": str(e)}
+
+    try:
+        response = await llm_call_async(
+            url, model,
+            [{"role": "user", "content": message}],
+            headers=headers,
+            timeout=AI_CHAT_TIMEOUT,
+        )
+        # Truncate very long responses
+        if len(response) > 10000:
+            response = response[:10000] + "\n... (truncated)"
+        return {"model": model, "response": response}
+    except Exception as e:
+        logger.error(f"chat_with_model failed: {e}")
+        return {"error": f"Failed to get response from {model_spec}: {e}"}
+
+
+async def ask_teacher(content: str, session_id: Optional[str] = None, owner: Optional[str] = None) -> Dict:
+    """Ask a more capable model for help.
+
+    Content format:
+      Line 1: model_name (or 'auto')
+      Line 2+: the problem description
+    """
+    from src.ai_interaction import _resolve_model, AI_CHAT_TIMEOUT
+    from src.llm_core import llm_call_async
+    from src.settings import get_setting
+
+    lines = content.strip().split("\n", 1)
+    model_spec = lines[0].strip() if lines else "auto"
+    problem = lines[1].strip() if len(lines) > 1 else ""
+
+    if not problem:
+        return {"error": "No problem description provided"}
+
+    if model_spec.lower() in ("auto", ""):
+        model_spec = get_setting("teacher_model", "")
+        if not model_spec:
+            return {"error": "No teacher model configured. Specify a model name or set teacher_model in settings."}
+
+    try:
+        url, model, headers = _resolve_model(model_spec, owner=owner)
+    except ValueError as e:
+        return {"error": str(e)}
+
+    try:
+        response = await llm_call_async(
+            url, model,
+            [
+                {"role": "system", "content": _TEACHER_SYSTEM_PROMPT},
+                {"role": "user", "content": f"Problem:\n{problem}"},
+            ],
+            headers=headers,
+            timeout=AI_CHAT_TIMEOUT,
+        )
+        if len(response) > 8000:
+            response = response[:8000] + "\n... (truncated)"
+        return {"model": model, "response": response, "teacher": True}
+    except Exception as e:
+        logger.error(f"ask_teacher failed: {e}")
+        return {"error": f"Teacher call failed ({model_spec}): {e}"}
+
+
+async def list_models(content: str, session_id: Optional[str] = None, owner: Optional[str] = None) -> Dict:
+    """List all available models across configured endpoints.
+
+    Content = optional filter keyword.
+    """
+    import json
+    import httpx
+    from src.database import SessionLocal, ModelEndpoint
+    from src.llm_core import _detect_provider, ANTHROPIC_MODELS
+    from src.auth_helpers import owner_filter
+    from src.endpoint_resolver import resolve_endpoint_runtime, build_headers, build_models_url
+
+    keyword = content.strip().lower() if content.strip() else None
+
+    db = SessionLocal()
+    try:
+        query = db.query(ModelEndpoint).filter(ModelEndpoint.is_enabled == True)
+        if owner:
+            query = owner_filter(query, ModelEndpoint, owner)
+        endpoints = query.all()
+        if not endpoints:
+            return {"results": "No enabled model endpoints configured."}
+
+        result_lines = []
+        total_models = 0
+
+        for ep in endpoints:
+            try:
+                base, api_key = resolve_endpoint_runtime(ep, owner=owner)
+            except Exception:
+                continue
+            provider = _detect_provider(base)
+            headers = build_headers(api_key, base)
+
+            model_ids = []
+            if provider == "anthropic":
+                model_ids = list(ANTHROPIC_MODELS)
+            else:
+                try:
+                    models_url = build_models_url(base)
+                    if models_url:
+                        r = httpx.get(models_url, headers=headers, timeout=5)
+                        r.raise_for_status()
+                        data = r.json()
+                        model_ids = [m.get("id") for m in (data.get("data") or []) if m.get("id")]
+                        if not model_ids:
+                            model_ids = [
+                                m.get("name") or m.get("model")
+                                for m in (data.get("models") or [])
+                                if m.get("name") or m.get("model")
+                            ]
+                    else:
+                        model_ids = json.loads(ep.cached_models or "[]")
+                except Exception:
+                    model_ids = ["(endpoint offline)"]
+
+            if keyword:
+                model_ids = [m for m in model_ids if keyword in m.lower() or keyword in (ep.name or "").lower()]
+
+            if model_ids:
+                result_lines.append(f"\n**{ep.name or base}** ({provider}):")
+                for mid in model_ids:
+                    result_lines.append(f"  - `{mid}`")
+                    total_models += 1
+
+        if not result_lines:
+            return {"results": "No models found" + (f" matching '{keyword}'" if keyword else "") + "."}
+
+        header = f"Available models ({total_models} total):"
+        return {"results": header + "\n".join(result_lines)}
+    except Exception as e:
+        logger.error(f"list_models failed: {e}")
+        return {"error": str(e)}
+    finally:
+        db.close()
+
+
+# ---------------------------------------------------------------------------
+# Handler classes registered in TOOL_HANDLERS
+# ---------------------------------------------------------------------------
+
+class ChatWithModelTool:
+    async def execute(self, content: str, ctx: dict) -> Dict:
+        return await chat_with_model(content, ctx.get("session_id"), owner=ctx.get("owner"))
+
+
+class AskTeacherTool:
+    async def execute(self, content: str, ctx: dict) -> Dict:
+        return await ask_teacher(content, ctx.get("session_id"), owner=ctx.get("owner"))
+
+
+class ListModelsTool:
+    async def execute(self, content: str, ctx: dict) -> Dict:
+        return await list_models(content, ctx.get("session_id"), owner=ctx.get("owner"))

src/agent_tools/session_tools.py

@@ -0,0 +1,464 @@
+"""session_tools.py - agent tools for AI-to-AI session management.
+
+Owns create_session, list_sessions, send_to_session and manage_session, moved
+out of src.ai_interaction as part of the tool -> registry migration (#3629), and
+their handler classes registered in TOOL_HANDLERS.
+
+The session manager is a runtime-set singleton in src.ai_interaction, so each
+function fetches it via get_session_manager() (imported here); _resolve_model and
+AI_CHAT_TIMEOUT are reused from there too.
+"""
+import json
+import logging
+import uuid
+from typing import Dict, Optional
+
+from src.ai_interaction import get_session_manager, _resolve_model, AI_CHAT_TIMEOUT
+
+logger = logging.getLogger(__name__)
+
+
+async def create_session(content: str, session_id: Optional[str] = None, owner: Optional[str] = None) -> Dict:
+    """Create a new chat session.
+
+    Content format:
+      Line 1: session name
+      Line 2: model_name (or model_name@endpoint_name)
+    """
+    _session_manager = get_session_manager()
+    if not _session_manager:
+        return {"error": "Session manager not available"}
+
+    lines = content.strip().split("\n")
+    if len(lines) < 2:
+        return {"error": "Need 2 lines: session name, then model spec"}
+
+    name = lines[0].strip()
+    model_spec = lines[1].strip()
+
+    if not name:
+        return {"error": "Session name cannot be empty"}
+
+    try:
+        url, model, headers = _resolve_model(model_spec, owner=owner)
+    except ValueError as e:
+        return {"error": str(e)}
+
+    sid = str(uuid.uuid4())[:8]
+    try:
+        _session_manager.create_session(
+            session_id=sid,
+            name=name,
+            endpoint_url=url,
+            model=model,
+            rag=False,
+            owner=owner,
+        )
+        # Store headers on session for future calls
+        sess = _session_manager.get_session(sid)
+        if sess and headers:
+            sess.headers = headers
+        try:
+            from src.event_bus import fire_event
+            fire_event("session_created", owner)
+        except Exception:
+            logger.debug("session_created event dispatch failed", exc_info=True)
+
+        return {"session_id": sid, "name": name, "model": model, "endpoint_url": url}
+    except Exception as e:
+        logger.error(f"create_session failed: {e}")
+        return {"error": f"Failed to create session: {e}"}
+
+async def list_sessions(content: str, session_id: Optional[str] = None, owner: Optional[str] = None) -> Dict:
+    """List sessions sorted by most-recently-active first.
+
+    Output includes a relative "last active" timestamp per row so the
+    agent can answer "open my last chat" without guessing from titles.
+    The most-recent session is always first in the list.
+
+    Content = optional filter keyword (matches session name).
+    """
+    _session_manager = get_session_manager()
+    if not _session_manager:
+        return {"error": "Session manager not available"}
+
+    keyword = content.strip().lower() if content.strip() else None
+
+    try:
+        from core.database import SessionLocal, Session as DbSession
+        from datetime import datetime, timezone
+
+        # Pull every session's last_accessed from the DB so we can sort
+        # by recency. In-memory sessions hold name + model + msg_count;
+        # the DB row holds the timestamps.
+        db = SessionLocal()
+        try:
+            db_rows = {r.id: r for r in db.query(DbSession).all()}
+        finally:
+            db.close()
+
+        # SECURITY: scope to the caller's sessions. Passing None returned
+        # every user's sessions, which the agent tool then exposed via the
+        # "list my chats" reply.
+        sessions = _session_manager.get_sessions_for_user(owner)
+        rows = []
+        for sid, sess in sessions.items():
+            if keyword and keyword not in (sess.name or "").lower():
+                continue
+            db_row = db_rows.get(sid)
+            # Prefer last_accessed; fall back to updated_at, then created_at.
+            ts = None
+            if db_row:
+                ts = getattr(db_row, 'last_accessed', None) or getattr(db_row, 'updated_at', None) or getattr(db_row, 'created_at', None)
+            rows.append((ts, sid, sess))
+
+        # Sort by timestamp DESC; rows without a timestamp sink to the bottom.
+        rows.sort(key=lambda r: r[0] or datetime.min, reverse=True)
+
+        def _rel(ts):
+            if not ts:
+                return 'never'
+            now = datetime.utcnow()
+            try:
+                if ts.tzinfo is not None:
+                    now = datetime.now(timezone.utc)
+                diff = (now - ts).total_seconds()
+            except Exception:
+                return 'unknown'
+            if diff < 60: return 'just now'
+            if diff < 3600: return f'{int(diff / 60)}m ago'
+            if diff < 86400: return f'{int(diff / 3600)}h ago'
+            if diff < 86400 * 7: return f'{int(diff / 86400)}d ago'
+            return ts.strftime('%Y-%m-%d')
+
+        lines = []
+        for i, (ts, sid, sess) in enumerate(rows):
+            if i >= 50:
+                lines.append(f"... and {len(rows) - 50} more (showing first 50)")
+                break
+            safe_name = (sess.name or "Untitled").replace("[", "\\[").replace("]", "\\]")
+            msg_count = getattr(sess, "message_count", 0) or 0
+            model = getattr(sess, "model", "unknown")
+            marker = " ← most recent" if i == 0 else ""
+            lines.append(f"- **[{safe_name}](#session-{sid})** (id: `{sid}`, model: {model}, {msg_count} msgs, last active {_rel(ts)}){marker}")
+
+        if not lines:
+            return {"results": "No sessions found" + (f" matching '{keyword}'" if keyword else "") + "."}
+
+        return {
+            "results": (
+                f"Found {len(rows)} session(s), sorted most-recent first:\n"
+                + "\n".join(lines)
+                + "\n\nAssistant: when replying to the user, preserve the chat-title markdown links exactly as shown, e.g. `[Chat](#session-id)`. Do not rewrite this as a plain, non-clickable table."
+            )
+        }
+    except Exception as e:
+        logger.error(f"list_sessions failed: {e}")
+        return {"error": str(e)}
+
+async def send_to_session(content: str, session_id: Optional[str] = None, owner: Optional[str] = None) -> Dict:
+    """Send a message to an existing session and get a response.
+
+    Content format:
+      Line 1: session_id
+      Line 2+: message
+    """
+    _session_manager = get_session_manager()
+    from src.llm_core import llm_call_async
+    from core.models import ChatMessage
+
+    if not _session_manager:
+        return {"error": "Session manager not available"}
+
+    lines = content.strip().split("\n", 1)
+    if len(lines) < 2:
+        return {"error": "Need 2 lines: session_id, then message"}
+
+    target_sid = lines[0].strip()
+    message = lines[1].strip()
+
+    sess = _session_manager.get_session(target_sid)
+    if not sess:
+        return {"error": f"Session '{target_sid}' not found"}
+
+    # Owner-scope: reject access to another user's session
+    if owner and getattr(sess, "owner", None) and sess.owner != owner:
+        return {"error": f"Session '{target_sid}' not found"}
+
+    if not message:
+        return {"error": "No message provided"}
+
+    try:
+        # Build context from session history
+        context = sess.get_context_messages()
+        context.append({"role": "user", "content": message})
+
+        response = await llm_call_async(
+            sess.endpoint_url, sess.model, context,
+            headers=sess.headers,
+            timeout=AI_CHAT_TIMEOUT,
+        )
+
+        # Save both messages to session
+        sess.add_message(ChatMessage("user", message))
+        sess.add_message(ChatMessage("assistant", response))
+
+        # Truncate for tool output
+        if len(response) > 10000:
+            response = response[:10000] + "\n... (truncated)"
+
+        return {
+            "session_id": target_sid,
+            "session_name": sess.name,
+            "response": response,
+        }
+    except Exception as e:
+        logger.error(f"send_to_session failed: {e}")
+        return {"error": f"Failed to send to session: {e}"}
+
+async def manage_session(content: str, session_id: Optional[str] = None, owner: Optional[str] = None) -> Dict:
+    """Manage sessions: rename, archive, delete, important, truncate, fork.
+
+    Content format:
+      Line 1: action (rename|archive|unarchive|delete|important|unimportant|truncate|fork)
+      Line 2: target session_id (or "current" to use the active session)
+      Line 3+: action-specific params (e.g. new name for rename, keep_count for truncate)
+    """
+    _session_manager = get_session_manager()
+    if not _session_manager:
+        return {"error": "Session manager not available"}
+
+    from src.database import SessionLocal, Session as DbSession
+
+    # Accept BOTH the structured JSON args the tool schema advertises
+    # ({action, session_id, value}) AND the legacy line-based format
+    # (line1=action, line2=session_id, line3=value). Native function-calling
+    # models send JSON; fenced-block callers send lines. Previously only the
+    # line format was parsed, so a model that followed the schema (JSON) got
+    # "Need at least 2 lines" / "Rename needs line 3" and couldn't drive it.
+    _raw = (content or "").strip()
+    action = ""
+    target_sid = ""
+    value = None      # the action param: new name (rename) / keep_count (truncate, fork)
+    _list_filter = ""
+    _parsed = None
+    if _raw.startswith("{"):
+        try:
+            _parsed = json.loads(_raw)
+        except Exception:
+            _parsed = None
+    if isinstance(_parsed, dict):
+        action = str(_parsed.get("action") or "").strip().lower()
+        target_sid = str(_parsed.get("session_id") or _parsed.get("session") or _parsed.get("id") or "").strip()
+        _v = _parsed.get("value")
+        if _v is None:
+            _v = (_parsed.get("name") or _parsed.get("new_name")
+                  or _parsed.get("title") or _parsed.get("keep_count"))
+        value = None if _v is None else str(_v).strip()
+        _list_filter = str(_parsed.get("filter") or "").strip()
+    else:
+        lines = _raw.split("\n")
+        if not lines or not lines[0].strip():
+            return {"error": "Missing action (rename|archive|delete|important|truncate|fork|list|switch)"}
+        action = lines[0].strip().lower()
+        target_sid = lines[1].strip() if len(lines) >= 2 else ""
+        value = lines[2].strip() if len(lines) >= 3 else None
+        _list_filter = "\n".join(lines[1:]).strip()
+
+    if not action:
+        return {"error": "Missing action (rename|archive|delete|important|truncate|fork|list|switch)"}
+
+    # `list` alias - dispatch to list_sessions so the agent's natural
+    # first guess (every other manage_* tool has a `list` action) works.
+    if action == "list":
+        return await list_sessions(_list_filter, session_id, owner=owner)
+
+    if not target_sid:
+        return {"error": "Need a session_id (or 'current' for the active chat)"}
+
+    # Allow "current" to refer to the active session
+    if target_sid.lower() == "current" and session_id:
+        target_sid = session_id
+
+    # `switch` / `open` / `select` / `view` - the agent reaches for
+    # these when the user asks to "open" or "switch to" a session.
+    # There's no server-side way to make the browser navigate, so we
+    # just return a clickable anchor link the user can click. The
+    # frontend's chat-history click delegate routes `#session-<id>`
+    # to selectSession(). The agent's reply naturally embeds this
+    # result so the user sees a single clickable line.
+    def _session_query(db):
+        query = db.query(DbSession).filter(DbSession.id == target_sid)
+        if owner is not None:
+            query = query.filter(DbSession.owner == owner)
+        return query
+
+    if action in ("switch", "open", "select", "view"):
+        db = SessionLocal()
+        try:
+            db_sess = _session_query(db).first()
+            if not db_sess:
+                return {"error": f"Session '{target_sid}' not found. Use list_sessions and pass the exact id it returned."}
+            name = db_sess.name or target_sid
+        finally:
+            db.close()
+        return {
+            "action": action,
+            "session_id": target_sid,
+            "name": name,
+            "results": f"[{name}](#session-{target_sid}) - click to open.",
+        }
+
+    db = SessionLocal()
+    try:
+        if action == "rename":
+            if not value:
+                return {"error": "rename needs a new name (the `value` arg, or line 3 in the legacy format)"}
+            new_name = value
+            db_sess = _session_query(db).first()
+            if not db_sess:
+                return {"error": f"Session '{target_sid}' not found. Use list_sessions and pass the exact id it returned."}
+            db_sess.name = new_name
+            db.commit()
+            _session_manager.update_session_name(target_sid, new_name)
+            return {"action": "rename", "session_id": target_sid, "name": new_name,
+                    "results": f"Session renamed to '{new_name}'"}
+
+        elif action == "archive":
+            db_sess = _session_query(db).first()
+            if not db_sess:
+                return {"error": f"Session '{target_sid}' not found. Use list_sessions and pass the exact id it returned."}
+            db_sess.archived = True
+            db.commit()
+            return {"action": "archive", "session_id": target_sid,
+                    "results": f"Session '{db_sess.name}' archived"}
+
+        elif action == "unarchive":
+            db_sess = _session_query(db).first()
+            if not db_sess:
+                return {"error": f"Session '{target_sid}' not found. Use list_sessions and pass the exact id it returned."}
+            db_sess.archived = False
+            db.commit()
+            return {"action": "unarchive", "session_id": target_sid,
+                    "results": f"Session '{db_sess.name}' unarchived"}
+
+        elif action == "delete":
+            if target_sid == session_id:
+                return {"error": "Cannot delete the current session while chatting in it. Delete other sessions first."}
+            db_sess = _session_query(db).first()
+            if not db_sess:
+                return {"error": f"Session '{target_sid}' not found. Refusing to delete an unknown chat id; use the exact id from list_sessions."}
+            if db_sess and db_sess.is_important:
+                return {"error": f"Session '{db_sess.name}' is starred/favorited. Unstar it first before deleting."}
+            try:
+                ok = _session_manager.delete_session(target_sid)
+                if not ok:
+                    return {"error": f"Session '{target_sid}' was not deleted because it no longer exists."}
+                return {"action": "delete", "session_id": target_sid,
+                        "results": f"Session '{db_sess.name or target_sid}' deleted"}
+            except Exception as e:
+                return {"error": f"Failed to delete session: {e}"}
+
+        elif action in ("important", "unimportant"):
+            is_important = action == "important"
+            db_sess = _session_query(db).first()
+            if not db_sess:
+                return {"error": f"Session '{target_sid}' not found. Use list_sessions and pass the exact id it returned."}
+            # Prevent AI from unstarring sessions - only the user can do that manually
+            if not is_important and db_sess.is_important:
+                return {"error": f"Session '{db_sess.name}' is starred by the user. Only the user can unstar sessions manually."}
+            db_sess.is_important = is_important
+            db.commit()
+            status = "marked as important" if is_important else "unmarked as important"
+            return {"action": action, "session_id": target_sid,
+                    "results": f"Session '{db_sess.name}' {status}"}
+
+        elif action == "truncate":
+            db_sess = _session_query(db).first()
+            if not db_sess:
+                return {"error": f"Session '{target_sid}' not found. Use list_sessions and pass the exact id it returned."}
+            keep_count = 10
+            if value:
+                try:
+                    keep_count = int(value)
+                except ValueError:
+                    pass
+            success = _session_manager.truncate_messages(target_sid, keep_count)
+            if success:
+                return {"action": "truncate", "session_id": target_sid,
+                        "results": f"Session truncated to last {keep_count} messages"}
+            return {"error": f"Failed to truncate session '{target_sid}'"}
+
+        elif action == "fork":
+            db_sess = _session_query(db).first()
+            if not db_sess:
+                return {"error": f"Session '{target_sid}' not found. Use list_sessions and pass the exact id it returned."}
+            keep_count = 0  # 0 = all messages
+            if value:
+                try:
+                    keep_count = int(value)
+                except ValueError:
+                    pass
+
+            source = _session_manager.get_session(target_sid)
+            if not source:
+                return {"error": f"Session '{target_sid}' not found"}
+
+            new_sid = str(uuid.uuid4())[:8]
+            _session_manager.create_session(
+                session_id=new_sid,
+                name=f"Fork: {source.name}",
+                endpoint_url=source.endpoint_url,
+                model=source.model,
+                rag=False,
+                owner=owner,
+            )
+            # Copy messages
+            history = source.get_context_messages()
+            if keep_count > 0:
+                history = history[:keep_count]
+            from core.models import ChatMessage as InMemoryMsg
+            new_sess = _session_manager.get_session(new_sid)
+            for msg in history:
+                new_sess.add_message(InMemoryMsg(msg["role"], msg["content"]))
+            try:
+                from src.event_bus import fire_event
+                fire_event("session_created", owner)
+            except Exception:
+                logger.debug("session_created event dispatch failed", exc_info=True)
+
+            return {"action": "fork", "session_id": new_sid,
+                    "source_session": target_sid, "messages_copied": len(history),
+                    "results": f"Forked session '{source.name}' -> new session {new_sid} ({len(history)} messages)"}
+
+        else:
+            return {"error": f"Unknown action '{action}'. Use: list, switch, rename, archive, unarchive, delete, important, unimportant, truncate, fork"}
+    except Exception as e:
+        logger.error(f"manage_session failed: {e}")
+        return {"error": str(e)}
+    finally:
+        db.close()
+
+
+# ---------------------------------------------------------------------------
+# Handler classes registered in TOOL_HANDLERS
+# ---------------------------------------------------------------------------
+
+class CreateSessionTool:
+    async def execute(self, content: str, ctx: dict) -> Dict:
+        return await create_session(content, ctx.get("session_id"), owner=ctx.get("owner"))
+
+
+class ListSessionsTool:
+    async def execute(self, content: str, ctx: dict) -> Dict:
+        return await list_sessions(content, ctx.get("session_id"), owner=ctx.get("owner"))
+
+
+class SendToSessionTool:
+    async def execute(self, content: str, ctx: dict) -> Dict:
+        return await send_to_session(content, ctx.get("session_id"), owner=ctx.get("owner"))
+
+
+class ManageSessionTool:
+    async def execute(self, content: str, ctx: dict) -> Dict:
+        return await manage_session(content, ctx.get("session_id"), owner=ctx.get("owner"))

src/ai_interaction.py

@@ -1,8 +1,14 @@
 """
 ai_interaction.py
 
-AI-to-AI interaction tools: chat_with_model, create_session, list_sessions,
+AI-to-AI interaction tools: pipeline and manage_memory, plus shared model
-send_to_session, pipeline.
+resolution (_resolve_model), the session-manager singleton, and dispatch_ai_tool.
+
+As part of the tool -> registry migration (#3629), chat_with_model, ask_teacher
+and list_models moved to src/agent_tools/model_interaction_tools.py, and
+create_session, list_sessions, send_to_session and manage_session moved to
+src/agent_tools/session_tools.py. Those modules reuse get_session_manager /
+_resolve_model / AI_CHAT_TIMEOUT from here.
 
 These are agent tools — the LLM writes fenced code blocks and they execute
 through the standard agent_tools.py pipeline.
@@ -159,440 +165,6 @@ def _resolve_model(spec: str, owner: Optional[str] = None) -> Tuple[str, str, Di
 # Tool implementations
 # ---------------------------------------------------------------------------
 
-async def do_chat_with_model(content: str, session_id: Optional[str] = None, owner: Optional[str] = None) -> Dict:
-    """Send a message to a specific model and return its response.
-
-    Content format:
-      Line 1: model_name (or model_name@endpoint_name)
-      Line 2+: the message to send
-    """
-    from src.llm_core import llm_call_async
-
-    lines = content.strip().split("\n", 1)
-    if not lines or not lines[0].strip():
-        return {"error": "First line must be the model name"}
-
-    model_spec = lines[0].strip()
-    message = lines[1].strip() if len(lines) > 1 else ""
-    if not message:
-        return {"error": "No message provided (line 2+ is the message)"}
-
-    try:
-        url, model, headers = _resolve_model(model_spec, owner=owner)
-    except ValueError as e:
-        return {"error": str(e)}
-
-    try:
-        response = await llm_call_async(
-            url, model,
-            [{"role": "user", "content": message}],
-            headers=headers,
-            timeout=AI_CHAT_TIMEOUT,
-        )
-        # Truncate very long responses
-        if len(response) > 10000:
-            response = response[:10000] + "\n... (truncated)"
-        return {"model": model, "response": response}
-    except Exception as e:
-        logger.error(f"chat_with_model failed: {e}")
-        return {"error": f"Failed to get response from {model_spec}: {e}"}
-
-
-_TEACHER_SYSTEM_PROMPT = (
-    "You are a senior AI mentor. A less capable model is stuck on a problem and asking for help. "
-    "Provide clear, actionable guidance:\n"
-    "1. Brief analysis of the problem\n"
-    "2. Recommended approach (step by step)\n"
-    "3. Key things to watch out for\n\n"
-    "Be concise and practical. No preamble."
-)
-
-
-async def do_ask_teacher(content: str, session_id: Optional[str] = None, owner: Optional[str] = None) -> Dict:
-    """Ask a more capable model for help.
-
-    Content format:
-      Line 1: model_name (or 'auto')
-      Line 2+: the problem description
-    """
-    from src.llm_core import llm_call_async
-    from src.settings import get_setting
-
-    lines = content.strip().split("\n", 1)
-    model_spec = lines[0].strip() if lines else "auto"
-    problem = lines[1].strip() if len(lines) > 1 else ""
-
-    if not problem:
-        return {"error": "No problem description provided"}
-
-    if model_spec.lower() in ("auto", ""):
-        model_spec = get_setting("teacher_model", "")
-        if not model_spec:
-            return {"error": "No teacher model configured. Specify a model name or set teacher_model in settings."}
-
-    try:
-        url, model, headers = _resolve_model(model_spec, owner=owner)
-    except ValueError as e:
-        return {"error": str(e)}
-
-    try:
-        response = await llm_call_async(
-            url, model,
-            [
-                {"role": "system", "content": _TEACHER_SYSTEM_PROMPT},
-                {"role": "user", "content": f"Problem:\n{problem}"},
-            ],
-            headers=headers,
-            timeout=AI_CHAT_TIMEOUT,
-        )
-        if len(response) > 8000:
-            response = response[:8000] + "\n... (truncated)"
-        return {"model": model, "response": response, "teacher": True}
-    except Exception as e:
-        logger.error(f"ask_teacher failed: {e}")
-        return {"error": f"Teacher call failed ({model_spec}): {e}"}
-
-
-async def do_second_opinion(content: str, session_id: Optional[str] = None, owner: Optional[str] = None) -> Dict:
-    """Get a second opinion from another model, then have the original model
-    evaluate the feedback and produce a unified version.
-
-    Content format:
-      Line 1: model_name (or model_name@endpoint_name)
-      Line 2+ (optional): specific question or focus area
-
-    Flow:
-      1. Pull recent conversation context
-      2. Send to reviewer model → get honest feedback
-      3. Send feedback back to the session's own model → evaluate & unify
-      4. Return both the review and the unified response
-    """
-    from src.llm_core import llm_call_async
-
-    lines = content.strip().split("\n", 1)
-    if not lines or not lines[0].strip():
-        return {"error": "First line must be the model name"}
-
-    model_spec = lines[0].strip()
-    focus = lines[1].strip() if len(lines) > 1 else ""
-
-    try:
-        reviewer_url, reviewer_model, reviewer_headers = _resolve_model(model_spec, owner=owner)
-    except ValueError as e:
-        return {"error": str(e)}
-
-    # Pull recent conversation context from current session
-    context_text = ""
-    sess = None
-    if session_id and _session_manager:
-        sess = _session_manager.get_session(session_id)
-        if sess:
-            messages = sess.get_context_messages()
-            recent = messages[-15:] if len(messages) > 15 else messages
-            parts = []
-            for m in recent:
-                role = m.get("role", "unknown").upper()
-                text = m.get("content", "")
-                if isinstance(text, list):
-                    text = " ".join(
-                        p.get("text", "") for p in text if isinstance(p, dict)
-                    )
-                if text:
-                    parts.append(f"[{role}]: {text[:2000]}")
-            context_text = "\n\n".join(parts)
-
-    if not context_text:
-        return {"error": "No conversation context found to review"}
-
-    # ── Step 1: Get the reviewer's feedback ──
-    reviewer_system = (
-        "You are giving a second opinion on a conversation between a user and an AI assistant. "
-        "Your job is to be genuinely helpful and honest — not a yes-man, but not a contrarian either.\n\n"
-        "Guidelines:\n"
-        "- If the plan/idea is solid, say so clearly. Don't manufacture problems that aren't there.\n"
-        "- If you spot a real flaw, blind spot, or simpler approach — call it out directly.\n"
-        "- Be practical. Don't over-engineer or over-analyze. Real-world tradeoffs matter.\n"
-        "- If there's a meaningfully better way to do something, suggest it concretely.\n"
-        "- Give credit where it's due — highlight what's working well.\n"
-        "- Keep it concise and actionable. No fluff.\n"
-        "- You're a second pair of eyes, not a professor grading a paper."
-    )
-
-    reviewer_message = f"Here's the conversation so far:\n\n{context_text}"
-    if focus:
-        reviewer_message += f"\n\n---\nSpecifically, I want your take on: {focus}"
-    else:
-        reviewer_message += "\n\n---\nGive me your honest second opinion on what's being discussed."
-
-    try:
-        review = await llm_call_async(
-            reviewer_url, reviewer_model,
-            [
-                {"role": "system", "content": reviewer_system},
-                {"role": "user", "content": reviewer_message},
-            ],
-            headers=reviewer_headers,
-            timeout=AI_CHAT_TIMEOUT,
-        )
-        if len(review) > 8000:
-            review = review[:8000] + "\n... (truncated)"
-    except Exception as e:
-        logger.error(f"second_opinion reviewer call failed: {e}")
-        return {"error": f"Failed to get second opinion from {model_spec}: {e}"}
-
-    # ── Step 2: Send review back to session's own model for evaluation ──
-    unified = ""
-    original_model = "unknown"
-    if sess:
-        original_url = sess.endpoint_url
-        original_model = sess.model
-        original_headers = getattr(sess, "headers", None) or {}
-
-        unify_system = (
-            "Another AI model just reviewed the conversation you've been having with the user. "
-            "Read their feedback carefully, then respond with:\n\n"
-            "1. **What you agree with** — acknowledge valid points honestly.\n"
-            "2. **What you disagree with** — explain why, briefly.\n"
-            "3. **Unified version** — produce an updated/refined version of whatever was being discussed, "
-            "incorporating the feedback you found valid. Don't accept every note blindly — "
-            "use your judgment on what actually improves things vs what's unnecessary.\n\n"
-            "Be concise and practical. The user wants a better result, not a meta-discussion."
-        )
-
-        unify_message = (
-            f"Here's the conversation context:\n\n{context_text}\n\n"
-            f"---\n\n"
-            f"**Review from {reviewer_model}:**\n\n{review}\n\n"
-            f"---\n\n"
-            f"Evaluate this feedback and produce a unified improved version."
-        )
-
-        try:
-            unified = await llm_call_async(
-                original_url, original_model,
-                [
-                    {"role": "system", "content": unify_system},
-                    {"role": "user", "content": unify_message},
-                ],
-                headers=original_headers,
-                timeout=AI_CHAT_TIMEOUT,
-            )
-            if len(unified) > 10000:
-                unified = unified[:10000] + "\n... (truncated)"
-        except Exception as e:
-            logger.error(f"second_opinion unify call failed: {e}")
-            unified = f"(Failed to get unified response: {e})"
-
-    # Build combined result
-    combined = (
-        f"## Second Opinion from {reviewer_model}\n\n{review}"
-        f"\n\n---\n\n"
-        f"## {original_model}'s Response\n\n{unified}"
-    )
-
-    return {
-        "model": reviewer_model,
-        "response": combined,
-        "instruction": "Present these results to the user exactly as they are. Do NOT call second_opinion again. The user can continue the conversation from here.",
-    }
-
-
-async def do_create_session(content: str, session_id: Optional[str] = None, owner: Optional[str] = None) -> Dict:
-    """Create a new chat session.
-
-    Content format:
-      Line 1: session name
-      Line 2: model_name (or model_name@endpoint_name)
-    """
-    if not _session_manager:
-        return {"error": "Session manager not available"}
-
-    lines = content.strip().split("\n")
-    if len(lines) < 2:
-        return {"error": "Need 2 lines: session name, then model spec"}
-
-    name = lines[0].strip()
-    model_spec = lines[1].strip()
-
-    if not name:
-        return {"error": "Session name cannot be empty"}
-
-    try:
-        url, model, headers = _resolve_model(model_spec, owner=owner)
-    except ValueError as e:
-        return {"error": str(e)}
-
-    sid = str(uuid.uuid4())[:8]
-    try:
-        _session_manager.create_session(
-            session_id=sid,
-            name=name,
-            endpoint_url=url,
-            model=model,
-            rag=False,
-            owner=owner,
-        )
-        # Store headers on session for future calls
-        sess = _session_manager.get_session(sid)
-        if sess and headers:
-            sess.headers = headers
-        try:
-            from src.event_bus import fire_event
-            fire_event("session_created", owner)
-        except Exception:
-            logger.debug("session_created event dispatch failed", exc_info=True)
-
-        return {"session_id": sid, "name": name, "model": model, "endpoint_url": url}
-    except Exception as e:
-        logger.error(f"create_session failed: {e}")
-        return {"error": f"Failed to create session: {e}"}
-
-
-async def do_list_sessions(content: str, session_id: Optional[str] = None, owner: Optional[str] = None) -> Dict:
-    """List sessions sorted by most-recently-active first.
-
-    Output includes a relative "last active" timestamp per row so the
-    agent can answer "open my last chat" without guessing from titles.
-    The most-recent session is always first in the list.
-
-    Content = optional filter keyword (matches session name).
-    """
-    if not _session_manager:
-        return {"error": "Session manager not available"}
-
-    keyword = content.strip().lower() if content.strip() else None
-
-    try:
-        from core.database import SessionLocal, Session as DbSession
-        from datetime import datetime, timezone
-
-        # Pull every session's last_accessed from the DB so we can sort
-        # by recency. In-memory sessions hold name + model + msg_count;
-        # the DB row holds the timestamps.
-        db = SessionLocal()
-        try:
-            db_rows = {r.id: r for r in db.query(DbSession).all()}
-        finally:
-            db.close()
-
-        # SECURITY: scope to the caller's sessions. Passing None returned
-        # every user's sessions, which the agent tool then exposed via the
-        # "list my chats" reply.
-        sessions = _session_manager.get_sessions_for_user(owner)
-        rows = []
-        for sid, sess in sessions.items():
-            if keyword and keyword not in (sess.name or "").lower():
-                continue
-            db_row = db_rows.get(sid)
-            # Prefer last_accessed; fall back to updated_at, then created_at.
-            ts = None
-            if db_row:
-                ts = getattr(db_row, 'last_accessed', None) or getattr(db_row, 'updated_at', None) or getattr(db_row, 'created_at', None)
-            rows.append((ts, sid, sess))
-
-        # Sort by timestamp DESC; rows without a timestamp sink to the bottom.
-        rows.sort(key=lambda r: r[0] or datetime.min, reverse=True)
-
-        def _rel(ts):
-            if not ts:
-                return 'never'
-            now = datetime.utcnow()
-            try:
-                if ts.tzinfo is not None:
-                    now = datetime.now(timezone.utc)
-                diff = (now - ts).total_seconds()
-            except Exception:
-                return 'unknown'
-            if diff < 60: return 'just now'
-            if diff < 3600: return f'{int(diff / 60)}m ago'
-            if diff < 86400: return f'{int(diff / 3600)}h ago'
-            if diff < 86400 * 7: return f'{int(diff / 86400)}d ago'
-            return ts.strftime('%Y-%m-%d')
-
-        lines = []
-        for i, (ts, sid, sess) in enumerate(rows):
-            if i >= 50:
-                lines.append(f"... and {len(rows) - 50} more (showing first 50)")
-                break
-            safe_name = (sess.name or "Untitled").replace("[", "\\[").replace("]", "\\]")
-            msg_count = getattr(sess, "message_count", 0) or 0
-            model = getattr(sess, "model", "unknown")
-            marker = " ← most recent" if i == 0 else ""
-            lines.append(f"- **[{safe_name}](#session-{sid})** (id: `{sid}`, model: {model}, {msg_count} msgs, last active {_rel(ts)}){marker}")
-
-        if not lines:
-            return {"results": "No sessions found" + (f" matching '{keyword}'" if keyword else "") + "."}
-
-        return {
-            "results": (
-                f"Found {len(rows)} session(s), sorted most-recent first:\n"
-                + "\n".join(lines)
-                + "\n\nAssistant: when replying to the user, preserve the chat-title markdown links exactly as shown, e.g. `[Chat](#session-id)`. Do not rewrite this as a plain, non-clickable table."
-            )
-        }
-    except Exception as e:
-        logger.error(f"list_sessions failed: {e}")
-        return {"error": str(e)}
-
-
-async def do_send_to_session(content: str, session_id: Optional[str] = None, owner: Optional[str] = None) -> Dict:
-    """Send a message to an existing session and get a response.
-
-    Content format:
-      Line 1: session_id
-      Line 2+: message
-    """
-    from src.llm_core import llm_call_async
-    from core.models import ChatMessage
-
-    if not _session_manager:
-        return {"error": "Session manager not available"}
-
-    lines = content.strip().split("\n", 1)
-    if len(lines) < 2:
-        return {"error": "Need 2 lines: session_id, then message"}
-
-    target_sid = lines[0].strip()
-    message = lines[1].strip()
-
-    sess = _session_manager.get_session(target_sid)
-    if not sess:
-        return {"error": f"Session '{target_sid}' not found"}
-
-    # Owner-scope: reject access to another user's session
-    if owner and getattr(sess, "owner", None) and sess.owner != owner:
-        return {"error": f"Session '{target_sid}' not found"}
-
-    if not message:
-        return {"error": "No message provided"}
-
-    try:
-        # Build context from session history
-        context = sess.get_context_messages()
-        context.append({"role": "user", "content": message})
-
-        response = await llm_call_async(
-            sess.endpoint_url, sess.model, context,
-            headers=sess.headers,
-            timeout=AI_CHAT_TIMEOUT,
-        )
-
-        # Save both messages to session
-        sess.add_message(ChatMessage("user", message))
-        sess.add_message(ChatMessage("assistant", response))
-
-        # Truncate for tool output
-        if len(response) > 10000:
-            response = response[:10000] + "\n... (truncated)"
-
-        return {
-            "session_id": target_sid,
-            "session_name": sess.name,
-            "response": response,
-        }
-    except Exception as e:
-        logger.error(f"send_to_session failed: {e}")
-        return {"error": f"Failed to send to session: {e}"}
 
 
 async def stream_ai_tool(tool: str, content: str, session_id: Optional[str] = None, owner: Optional[str] = None):
@@ -715,229 +287,6 @@ async def do_pipeline(content: str, session_id: Optional[str] = None, owner: Opt
 # Session management tool
 # ---------------------------------------------------------------------------
 
-async def do_manage_session(content: str, session_id: Optional[str] = None, owner: Optional[str] = None) -> Dict:
-    """Manage sessions: rename, archive, delete, important, truncate, fork.
-
-    Content format:
-      Line 1: action (rename|archive|unarchive|delete|important|unimportant|truncate|fork)
-      Line 2: target session_id (or "current" to use the active session)
-      Line 3+: action-specific params (e.g. new name for rename, keep_count for truncate)
-    """
-    if not _session_manager:
-        return {"error": "Session manager not available"}
-
-    from src.database import SessionLocal, Session as DbSession
-
-    # Accept BOTH the structured JSON args the tool schema advertises
-    # ({action, session_id, value}) AND the legacy line-based format
-    # (line1=action, line2=session_id, line3=value). Native function-calling
-    # models send JSON; fenced-block callers send lines. Previously only the
-    # line format was parsed, so a model that followed the schema (JSON) got
-    # "Need at least 2 lines" / "Rename needs line 3" and couldn't drive it.
-    _raw = (content or "").strip()
-    action = ""
-    target_sid = ""
-    value = None      # the action param: new name (rename) / keep_count (truncate, fork)
-    _list_filter = ""
-    _parsed = None
-    if _raw.startswith("{"):
-        try:
-            _parsed = json.loads(_raw)
-        except Exception:
-            _parsed = None
-    if isinstance(_parsed, dict):
-        action = str(_parsed.get("action") or "").strip().lower()
-        target_sid = str(_parsed.get("session_id") or _parsed.get("session") or _parsed.get("id") or "").strip()
-        _v = _parsed.get("value")
-        if _v is None:
-            _v = (_parsed.get("name") or _parsed.get("new_name")
-                  or _parsed.get("title") or _parsed.get("keep_count"))
-        value = None if _v is None else str(_v).strip()
-        _list_filter = str(_parsed.get("filter") or "").strip()
-    else:
-        lines = _raw.split("\n")
-        if not lines or not lines[0].strip():
-            return {"error": "Missing action (rename|archive|delete|important|truncate|fork|list|switch)"}
-        action = lines[0].strip().lower()
-        target_sid = lines[1].strip() if len(lines) >= 2 else ""
-        value = lines[2].strip() if len(lines) >= 3 else None
-        _list_filter = "\n".join(lines[1:]).strip()
-
-    if not action:
-        return {"error": "Missing action (rename|archive|delete|important|truncate|fork|list|switch)"}
-
-    # `list` alias — dispatch to do_list_sessions so the agent's natural
-    # first guess (every other manage_* tool has a `list` action) works.
-    if action == "list":
-        return await do_list_sessions(_list_filter, session_id, owner=owner)
-
-    if not target_sid:
-        return {"error": "Need a session_id (or 'current' for the active chat)"}
-
-    # Allow "current" to refer to the active session
-    if target_sid.lower() == "current" and session_id:
-        target_sid = session_id
-
-    # `switch` / `open` / `select` / `view` — the agent reaches for
-    # these when the user asks to "open" or "switch to" a session.
-    # There's no server-side way to make the browser navigate, so we
-    # just return a clickable anchor link the user can click. The
-    # frontend's chat-history click delegate routes `#session-<id>`
-    # to selectSession(). The agent's reply naturally embeds this
-    # result so the user sees a single clickable line.
-    def _session_query(db):
-        query = db.query(DbSession).filter(DbSession.id == target_sid)
-        if owner is not None:
-            query = query.filter(DbSession.owner == owner)
-        return query
-
-    if action in ("switch", "open", "select", "view"):
-        db = SessionLocal()
-        try:
-            db_sess = _session_query(db).first()
-            if not db_sess:
-                return {"error": f"Session '{target_sid}' not found. Use list_sessions and pass the exact id it returned."}
-            name = db_sess.name or target_sid
-        finally:
-            db.close()
-        return {
-            "action": action,
-            "session_id": target_sid,
-            "name": name,
-            "results": f"[{name}](#session-{target_sid}) — click to open.",
-        }
-
-    db = SessionLocal()
-    try:
-        if action == "rename":
-            if not value:
-                return {"error": "rename needs a new name (the `value` arg, or line 3 in the legacy format)"}
-            new_name = value
-            db_sess = _session_query(db).first()
-            if not db_sess:
-                return {"error": f"Session '{target_sid}' not found. Use list_sessions and pass the exact id it returned."}
-            db_sess.name = new_name
-            db.commit()
-            _session_manager.update_session_name(target_sid, new_name)
-            return {"action": "rename", "session_id": target_sid, "name": new_name,
-                    "results": f"Session renamed to '{new_name}'"}
-
-        elif action == "archive":
-            db_sess = _session_query(db).first()
-            if not db_sess:
-                return {"error": f"Session '{target_sid}' not found. Use list_sessions and pass the exact id it returned."}
-            db_sess.archived = True
-            db.commit()
-            return {"action": "archive", "session_id": target_sid,
-                    "results": f"Session '{db_sess.name}' archived"}
-
-        elif action == "unarchive":
-            db_sess = _session_query(db).first()
-            if not db_sess:
-                return {"error": f"Session '{target_sid}' not found. Use list_sessions and pass the exact id it returned."}
-            db_sess.archived = False
-            db.commit()
-            return {"action": "unarchive", "session_id": target_sid,
-                    "results": f"Session '{db_sess.name}' unarchived"}
-
-        elif action == "delete":
-            if target_sid == session_id:
-                return {"error": "Cannot delete the current session while chatting in it. Delete other sessions first."}
-            db_sess = _session_query(db).first()
-            if not db_sess:
-                return {"error": f"Session '{target_sid}' not found. Refusing to delete an unknown chat id; use the exact id from list_sessions."}
-            if db_sess and db_sess.is_important:
-                return {"error": f"Session '{db_sess.name}' is starred/favorited. Unstar it first before deleting."}
-            try:
-                ok = _session_manager.delete_session(target_sid)
-                if not ok:
-                    return {"error": f"Session '{target_sid}' was not deleted because it no longer exists."}
-                return {"action": "delete", "session_id": target_sid,
-                        "results": f"Session '{db_sess.name or target_sid}' deleted"}
-            except Exception as e:
-                return {"error": f"Failed to delete session: {e}"}
-
-        elif action in ("important", "unimportant"):
-            is_important = action == "important"
-            db_sess = _session_query(db).first()
-            if not db_sess:
-                return {"error": f"Session '{target_sid}' not found. Use list_sessions and pass the exact id it returned."}
-            # Prevent AI from unstarring sessions — only the user can do that manually
-            if not is_important and db_sess.is_important:
-                return {"error": f"Session '{db_sess.name}' is starred by the user. Only the user can unstar sessions manually."}
-            db_sess.is_important = is_important
-            db.commit()
-            status = "marked as important" if is_important else "unmarked as important"
-            return {"action": action, "session_id": target_sid,
-                    "results": f"Session '{db_sess.name}' {status}"}
-
-        elif action == "truncate":
-            db_sess = _session_query(db).first()
-            if not db_sess:
-                return {"error": f"Session '{target_sid}' not found. Use list_sessions and pass the exact id it returned."}
-            keep_count = 10
-            if value:
-                try:
-                    keep_count = int(value)
-                except ValueError:
-                    pass
-            success = _session_manager.truncate_messages(target_sid, keep_count)
-            if success:
-                return {"action": "truncate", "session_id": target_sid,
-                        "results": f"Session truncated to last {keep_count} messages"}
-            return {"error": f"Failed to truncate session '{target_sid}'"}
-
-        elif action == "fork":
-            db_sess = _session_query(db).first()
-            if not db_sess:
-                return {"error": f"Session '{target_sid}' not found. Use list_sessions and pass the exact id it returned."}
-            keep_count = 0  # 0 = all messages
-            if value:
-                try:
-                    keep_count = int(value)
-                except ValueError:
-                    pass
-
-            source = _session_manager.get_session(target_sid)
-            if not source:
-                return {"error": f"Session '{target_sid}' not found"}
-
-            new_sid = str(uuid.uuid4())[:8]
-            _session_manager.create_session(
-                session_id=new_sid,
-                name=f"Fork: {source.name}",
-                endpoint_url=source.endpoint_url,
-                model=source.model,
-                rag=False,
-                owner=owner,
-            )
-            # Copy messages
-            history = source.get_context_messages()
-            if keep_count > 0:
-                history = history[:keep_count]
-            from core.models import ChatMessage as InMemoryMsg
-            new_sess = _session_manager.get_session(new_sid)
-            for msg in history:
-                new_sess.add_message(InMemoryMsg(msg["role"], msg["content"]))
-            try:
-                from src.event_bus import fire_event
-                fire_event("session_created", owner)
-            except Exception:
-                logger.debug("session_created event dispatch failed", exc_info=True)
-
-            return {"action": "fork", "session_id": new_sid,
-                    "source_session": target_sid, "messages_copied": len(history),
-                    "results": f"Forked session '{source.name}' -> new session {new_sid} ({len(history)} messages)"}
-
-        else:
-            return {"error": f"Unknown action '{action}'. Use: list, switch, rename, archive, unarchive, delete, important, unimportant, truncate, fork"}
-    except Exception as e:
-        logger.error(f"manage_session failed: {e}")
-        return {"error": str(e)}
-    finally:
-        db.close()
-
-
 # ---------------------------------------------------------------------------
 # Memory management tool
 # ---------------------------------------------------------------------------
@@ -1104,83 +453,6 @@ async def do_manage_memory(content: str, session_id: Optional[str] = None, owner
         return {"error": f"Unknown action '{action}'. Use: list, add, edit, delete, search"}
 
 
-# ---------------------------------------------------------------------------
-# List models tool
-# ---------------------------------------------------------------------------
-
-async def do_list_models(content: str, session_id: Optional[str] = None, owner: Optional[str] = None) -> Dict:
-    """List all available models across configured endpoints.
-
-    Content = optional filter keyword.
-    """
-    import httpx
-    from src.database import SessionLocal, ModelEndpoint
-    from src.llm_core import _detect_provider, ANTHROPIC_MODELS
-    from src.auth_helpers import owner_filter
-
-    keyword = content.strip().lower() if content.strip() else None
-
-    db = SessionLocal()
-    try:
-        query = db.query(ModelEndpoint).filter(ModelEndpoint.is_enabled == True)
-        if owner:
-            query = owner_filter(query, ModelEndpoint, owner)
-        endpoints = query.all()
-        if not endpoints:
-            return {"results": "No enabled model endpoints configured."}
-
-        result_lines = []
-        total_models = 0
-
-        for ep in endpoints:
-            try:
-                base, api_key = resolve_endpoint_runtime(ep, owner=owner)
-            except Exception:
-                continue
-            provider = _detect_provider(base)
-            headers = build_headers(api_key, base)
-
-            model_ids = []
-            if provider == "anthropic":
-                model_ids = list(ANTHROPIC_MODELS)
-            else:
-                try:
-                    models_url = build_models_url(base)
-                    if models_url:
-                        r = httpx.get(models_url, headers=headers, timeout=5)
-                        r.raise_for_status()
-                        data = r.json()
-                        model_ids = [m.get("id") for m in (data.get("data") or []) if m.get("id")]
-                        if not model_ids:
-                            model_ids = [
-                                m.get("name") or m.get("model")
-                                for m in (data.get("models") or [])
-                                if m.get("name") or m.get("model")
-                            ]
-                    else:
-                        model_ids = json.loads(ep.cached_models or "[]")
-                except Exception:
-                    model_ids = ["(endpoint offline)"]
-
-            if keyword:
-                model_ids = [m for m in model_ids if keyword in m.lower() or keyword in (ep.name or "").lower()]
-
-            if model_ids:
-                result_lines.append(f"\n**{ep.name or base}** ({provider}):")
-                for mid in model_ids:
-                    result_lines.append(f"  - `{mid}`")
-                    total_models += 1
-
-        if not result_lines:
-            return {"results": "No models found" + (f" matching '{keyword}'" if keyword else "") + "."}
-
-        header = f"Available models ({total_models} total):"
-        return {"results": header + "\n".join(result_lines)}
-    except Exception as e:
-        logger.error(f"list_models failed: {e}")
-        return {"error": str(e)}
-    finally:
-        db.close()
 
 
 # ---------------------------------------------------------------------------
@@ -1831,55 +1103,20 @@ async def dispatch_ai_tool(
 ) -> Tuple[str, Dict]:
     """Dispatch an AI interaction tool. Returns (description, result_dict)."""
 
-    if tool == "chat_with_model":
+    if tool == "pipeline":
-        model_spec = content.split("\n")[0].strip()[:60]
-        desc = f"chat_with_model: {model_spec}"
-        result = await do_chat_with_model(content, session_id, owner=owner)
-
-    elif tool == "create_session":
-        name = content.split("\n")[0].strip()[:60]
-        desc = f"create_session: {name}"
-        result = await do_create_session(content, session_id, owner=owner)
-
-    elif tool == "list_sessions":
-        keyword = content.strip()[:40]
-        desc = f"list_sessions{': ' + keyword if keyword else ''}"
-        result = await do_list_sessions(content, session_id, owner=owner)
-
-    elif tool == "send_to_session":
-        sid = content.split("\n")[0].strip()[:20]
-        desc = f"send_to_session: {sid}"
-        result = await do_send_to_session(content, session_id, owner=owner)
-
-    elif tool == "pipeline":
         desc = "pipeline: running steps"
         result = await do_pipeline(content, session_id, owner=owner)
 
-    elif tool == "manage_session":
-        action = content.split("\n")[0].strip()[:40]
-        desc = f"manage_session: {action}"
-        result = await do_manage_session(content, session_id, owner=owner)
-
     elif tool == "manage_memory":
         action = content.split("\n")[0].strip()[:40]
         desc = f"manage_memory: {action}"
         result = await do_manage_memory(content, session_id, owner=owner)
 
-    elif tool == "list_models":
-        keyword = content.strip()[:40]
-        desc = f"list_models{': ' + keyword if keyword else ''}"
-        result = await do_list_models(content, session_id, owner=owner)
-
     elif tool == "ui_control":
         action = content.split("\n")[0].strip()[:60]
         desc = f"ui_control: {action}"
         result = await do_ui_control(content, session_id, owner=owner)
 
-    elif tool == "ask_teacher":
-        problem = content.split("\n", 1)[-1].strip()[:60]
-        desc = f"ask_teacher: {problem}"
-        result = await do_ask_teacher(content, session_id, owner=owner)
-
     else:
         desc = f"unknown ai tool: {tool}"
         result = {"error": f"Unknown AI interaction tool: {tool}"}

src/bg_jobs.py

@@ -263,10 +263,32 @@ def list_for_session(session_id: str) -> List[Dict[str, Any]]:
     return [r for r in refresh().values() if r.get("session_id") == session_id]
 
 
+def kill(job_id: str) -> Optional[Dict[str, Any]]:
+    """Terminate a running job's process tree and mark it killed. Returns the
+    updated record, or None if the id is unknown. Idempotent: a job that already
+    finished is returned unchanged. Sets followed_up so the monitor does not also
+    fire an auto-continue for a job the agent deliberately stopped."""
+    jobs = _load()
+    rec = jobs.get(job_id)
+    if rec is None:
+        return None
+    if rec.get("status") == "running":
+        _kill(rec.get("pid"))
+        rec["status"] = "failed"
+        rec["exit_code"] = -1
+        rec["ended_at"] = time.time()
+        rec["killed"] = True
+        rec["followed_up"] = True
+        _save(jobs)
+    return rec
+
+
 def result_text(rec: Dict[str, Any]) -> str:
     """Human/agent-readable summary of a finished job, for the follow-up."""
     out = _read_output(rec)
-    if rec.get("timed_out"):
+    if rec.get("killed"):
+        head = "Background job was killed."
+    elif rec.get("timed_out"):
         head = f"Background job timed out after {rec.get('max_runtime_s')}s."
     elif rec.get("died"):
         head = "Background job process died unexpectedly (no exit code)."

src/constants.py

@@ -102,6 +102,9 @@ SEARXNG_INSTANCE = os.getenv("SEARXNG_INSTANCE", "http://localhost:8080")
 CLEANUP_ENABLED = os.getenv("CLEANUP_ENABLED", "True").lower() == "true"
 CLEANUP_INTERVAL_HOURS = int(os.getenv("CLEANUP_INTERVAL_HOURS", "24"))
 
+# Auth policy
+PASSWORD_MIN_LENGTH = 8
+
 # Default parameters
 DEFAULT_TEMPERATURE = 1.0
 DEFAULT_MAX_TOKENS = 0

src/endpoint_resolver.py

@@ -161,6 +161,32 @@ def normalize_base(url: str) -> str:
     return url
 
 
+def _validated_endpoint_base(url: str) -> str:
+    """Return a base URL that is safe for endpoint path appends."""
+    base = (url or "").strip().rstrip("/")
+    if "?" in base or "#" in base:
+        raise ValueError("Endpoint base URL must not include query or fragment")
+    return urlunparse(urlparse(base)._replace(query="", fragment="")).rstrip("/")
+
+
+def _prepare_endpoint_base(base: str) -> str:
+    base = _validated_endpoint_base(normalize_base(base))
+    return _validated_endpoint_base(normalize_base(resolve_url(base)))
+
+
+def _append_endpoint_path(base: str, suffix: str) -> str:
+    parsed = urlparse(base)
+    current = (parsed.path or "").rstrip("/")
+    extra = "/" + suffix.lstrip("/")
+    path = f"{current}{extra}" if current else extra
+    return urlunparse(parsed._replace(path=path, query="", fragment=""))
+
+
+def _pathless_host(base: str, host: str) -> bool:
+    parsed = urlparse(base)
+    return (parsed.hostname or "").lower() == host and not (parsed.path or "").strip("/")
+
+
 def _anthropic_api_root(base: str) -> str:
     """Return Anthropic's API root, preserving /v1 for OpenAI-compatible APIs elsewhere."""
     base = (base or "").strip().rstrip("/")
@@ -171,15 +197,17 @@ def _anthropic_api_root(base: str) -> str:
 
 def build_chat_url(base: str) -> str:
     """Return the correct chat endpoint URL for a given base."""
-    base = resolve_url(base)
+    base = _prepare_endpoint_base(base)
     provider = _detect_provider(base)
     if provider == "anthropic":
-        return _anthropic_api_root(base) + "/v1/messages"
+        return _append_endpoint_path(_anthropic_api_root(base), "/v1/messages")
     if provider == "ollama":
-        return _ollama_api_root(base) + "/chat"
+        return _append_endpoint_path(_ollama_api_root(base), "/chat")
     if provider == "chatgpt-subscription":
-        return base.rstrip("/") + "/responses"
+        return _append_endpoint_path(base, "/responses")
-    return base + "/chat/completions"
+    if _pathless_host(base, "api.openai.com"):
+        base = _append_endpoint_path(base, "/v1")
+    return _append_endpoint_path(base, "/chat/completions")
 
 
 def build_models_url(base: str) -> Optional[str]:
@@ -193,12 +221,12 @@ def build_models_url(base: str) -> Optional[str]:
     untouched (so custom prefixes like ``/openai`` or ``/api/openai/v1`` keep
     their semantics).
     """
-    base = normalize_base(resolve_url(base))
+    base = _prepare_endpoint_base(base)
     provider = _detect_provider(base)
     if provider == "anthropic":
-        return _anthropic_api_root(base) + "/v1/models"
+        return _append_endpoint_path(_anthropic_api_root(base), "/v1/models")
     if provider == "ollama":
-        return _ollama_api_root(base) + "/tags"
+        return _append_endpoint_path(_ollama_api_root(base), "/tags")
     if provider == "chatgpt-subscription":
         return None
     # Generic OpenAI-compatible fallback: local model servers with no explicit
@@ -208,10 +236,10 @@ def build_models_url(base: str) -> Optional[str]:
     parsed = urlparse(base)
     host = (parsed.hostname or "").lower()
     is_local = host in {"localhost", "127.0.0.1", "::1", "host.docker.internal"}
-    uses_v1_models_by_default = is_local or host in {"api.deepseek.com"}
+    uses_v1_models_by_default = is_local or host in {"api.deepseek.com", "api.openai.com"}
     if not parsed.path and uses_v1_models_by_default:
-        base = base + "/v1"
+        base = _append_endpoint_path(base, "/v1")
-    return base + "/models"
+    return _append_endpoint_path(base, "/models")
 
 
 def build_headers(api_key: Optional[str], base: str) -> Dict[str, str]:

src/integrations.py

@@ -4,6 +4,7 @@ import uuid
 import logging
 import re
 from typing import Dict, List, Optional, Any
+from urllib.parse import urljoin, urlparse, urlunparse
 
 import httpx
 from fastapi import HTTPException
@@ -202,6 +203,22 @@ def mask_integration_secret(integration: Dict[str, Any]) -> Dict[str, Any]:
     return safe
 
 
+def _normalize_integration_base_url(base_url: Any) -> str:
+    if not isinstance(base_url, str) or not base_url.strip():
+        raise ValueError("Integration base URL is required")
+    cleaned = base_url.strip().rstrip("/")
+    if "?" in cleaned or "#" in cleaned:
+        raise ValueError("Integration base URL must not include query or fragment")
+    parsed = urlparse(cleaned)
+    if parsed.scheme.lower() not in ("http", "https") or not parsed.hostname:
+        raise ValueError("Integration base URL must be an HTTP(S) URL")
+    return urlunparse(parsed._replace(scheme=parsed.scheme.lower(), query="", fragment="")).rstrip("/")
+
+
+def _join_integration_url(base_url: str, path: str) -> str:
+    return urljoin(base_url.rstrip("/") + "/", path.lstrip("/"))
+
+
 def load_integrations() -> List[Dict[str, Any]]:
     """Load all integrations from disk with secrets decrypted for runtime use."""
     if not os.path.exists(DATA_FILE):
@@ -261,8 +278,10 @@ def add_integration(data: Dict[str, Any]) -> Dict[str, Any]:
 
     if not isinstance(integration.get("name"), str) or not integration["name"].strip():
         raise HTTPException(400, "Integration name is required")
-    if not isinstance(integration.get("base_url"), str) or not integration["base_url"].strip():
+    try:
-        raise HTTPException(400, "Integration base URL is required")
+        integration["base_url"] = _normalize_integration_base_url(integration.get("base_url"))
+    except ValueError as exc:
+        raise HTTPException(400, str(exc)) from exc
 
     integrations = load_integrations()
     integrations.append(integration)
@@ -272,10 +291,14 @@ def add_integration(data: Dict[str, Any]) -> Dict[str, Any]:
 
 def update_integration(integration_id: str, data: Dict[str, Any]) -> Optional[Dict[str, Any]]:
     """Update fields on an existing integration. Returns updated integration or None."""
+    data = dict(data)
     if "name" in data and (not isinstance(data["name"], str) or not data["name"].strip()):
         raise HTTPException(400, "Integration name is required")
-    if "base_url" in data and (not isinstance(data["base_url"], str) or not data["base_url"].strip()):
+    if "base_url" in data:
-        raise HTTPException(400, "Integration base URL is required")
+        try:
+            data["base_url"] = _normalize_integration_base_url(data["base_url"])
+        except ValueError as exc:
+            raise HTTPException(400, str(exc)) from exc
 
     integrations = load_integrations()
     for item in integrations:
@@ -341,9 +364,10 @@ async def execute_api_call(
     if not integration.get("enabled", True):
         return {"error": f"Integration '{integration.get('name')}' is disabled", "exit_code": 1}
 
-    base_url = integration.get("base_url", "").rstrip("/")
+    try:
-    if not base_url:
+        base_url = _normalize_integration_base_url(integration.get("base_url", ""))
-        return {"error": "Integration has no base_url configured", "exit_code": 1}
+    except ValueError as exc:
+        return {"error": str(exc), "exit_code": 1}
 
     # Strip common API path suffixes users might accidentally include
     # (e.g. "http://host/v1/" → "http://host"). The integration's preset
@@ -366,7 +390,10 @@ async def execute_api_call(
     if re.search(r"^https?://", path) or "://" in path:
         return {"error": "Path must not contain a protocol scheme", "exit_code": 1}
 
-    url = base_url + path
+    if "#" in path:
+        return {"error": "Path must not contain a fragment", "exit_code": 1}
+
+    url = _join_integration_url(base_url, path)
     method = method.upper()
 
     # Build headers

src/llm_core.py

@@ -87,7 +87,7 @@ _host_health_lock = threading.Lock()
 _model_activity: Dict[str, float] = {}
 
 _HARMONY_MARKER_RE = re.compile(
-    r"<\|channel\|>(analysis|final)"
+    r"<\|channel\|>(analysis|commentary|final)"
     r"|<\|start\|>(?:assistant|system|user|tool)?"
     r"|<\|message\|>"
     r"|<\|end\|>"
@@ -96,6 +96,7 @@ _HARMONY_MARKER_RE = re.compile(
 )
 _HARMONY_MARKERS = (
     "<|channel|>analysis",
+    "<|channel|>commentary",
     "<|channel|>final",
     "<|start|>assistant",
     "<|start|>system",
@@ -145,7 +146,10 @@ class _HarmonyStreamRouter:
             out.append((text, False))
             return
         if self._in_message:
-            out.append((text, self._channel == "analysis"))
+            # analysis + commentary (tool-call preambles / function-arg bodies)
+            # are internal, not user-facing — route them to thinking so they
+            # don't leak into the visible answer; only `final` is visible.
+            out.append((text, self._channel in ("analysis", "commentary")))
 
     def _handle_marker(self, match: re.Match[str]) -> None:
         marker = match.group(0)

src/model_context.py

@@ -17,10 +17,11 @@ import httpx
 logger = logging.getLogger(__name__)
 
 _LOCAL_HOSTS = {"localhost", "127.0.0.1", "0.0.0.0", "::1", "host.docker.internal"}
-_PRIVATE_PREFIXES = ("10.", "172.16.", "172.17.", "172.18.", "172.19.",
+_PRIVATE_NETWORKS = (
-                     "172.20.", "172.21.", "172.22.", "172.23.", "172.24.",
+    ipaddress.ip_network("10.0.0.0/8"),
-                     "172.25.", "172.26.", "172.27.", "172.28.", "172.29.",
+    ipaddress.ip_network("172.16.0.0/12"),
-                     "172.30.", "172.31.", "192.168.")
+    ipaddress.ip_network("192.168.0.0/16"),
+)
 
 # Tailscale uses the CGNAT range 100.64.0.0/10, NOT all of 100.0.0.0/8.
 # A bare "100." prefix would classify public addresses (e.g. AWS ranges
@@ -36,6 +37,14 @@ def _in_tailscale_range(host: str) -> bool:
         return False
 
 
+def _is_private_ip_literal(host: str) -> bool:
+    try:
+        ip = ipaddress.ip_address(host)
+    except ValueError:
+        return False
+    return any(ip in network for network in _PRIVATE_NETWORKS)
+
+
 def _normalize_base_for_compare(url: str) -> str:
     url = (url or "").strip().rstrip("/")
     for suffix in ("/chat/completions", "/models", "/completions", "/v1/messages"):
@@ -87,7 +96,7 @@ def is_local_endpoint(url: str) -> bool:
         return True
     try:
         host = urlparse(url).hostname or ""
-        return host in _LOCAL_HOSTS or host.startswith(_PRIVATE_PREFIXES) or _in_tailscale_range(host)
+        return host in _LOCAL_HOSTS or _is_private_ip_literal(host) or _in_tailscale_range(host)
     except Exception:
         return False
 

src/personal_docs.py

@@ -322,6 +322,47 @@ class PersonalDocsManager:
         else:
             logger.info(f"Directory not in index: {directory}")
 
+    def rename_directory(self, old_directory: str, new_directory: str, *, path_map: Dict[str, str] = None):
+        """Rewrite tracked directory and excluded-file paths after an owner rename."""
+        old_directory = os.path.abspath(old_directory)
+        new_directory = os.path.abspath(new_directory)
+        path_map = {os.path.abspath(k): os.path.abspath(v) for k, v in (path_map or {}).items()}
+
+        def rewrite(path: str) -> str:
+            abs_path = os.path.abspath(path)
+            mapped = path_map.get(abs_path)
+            if mapped:
+                return mapped
+            if abs_path == old_directory:
+                return new_directory
+            if abs_path.startswith(old_directory + os.sep):
+                return new_directory + abs_path[len(old_directory):]
+            return abs_path
+
+        changed_dirs = False
+        rewritten_dirs = []
+        for directory in self.indexed_directories:
+            rewritten = rewrite(directory)
+            changed_dirs = changed_dirs or rewritten != os.path.abspath(directory)
+            if rewritten not in rewritten_dirs:
+                rewritten_dirs.append(rewritten)
+        if changed_dirs:
+            self.indexed_directories = rewritten_dirs
+            self.save_directories()
+
+        changed_excluded = False
+        rewritten_excluded = set()
+        for path in self.excluded_files:
+            rewritten = rewrite(path)
+            changed_excluded = changed_excluded or rewritten != os.path.abspath(path)
+            rewritten_excluded.add(rewritten)
+        if changed_excluded:
+            self.excluded_files = rewritten_excluded
+            self._save_excluded()
+
+        if changed_dirs or changed_excluded:
+            self.refresh_index()
+
     def get_indexed_directories(self):
         """Get the list of all indexed directories."""
         return self.indexed_directories.copy()

src/rag_vector.py

@@ -50,6 +50,23 @@ def _generate_doc_id(text: str, owner: str = "") -> str:
     return f"doc_{hashlib.sha256(key.encode('utf-8')).hexdigest()[:16]}"
 
 
+def _rewrite_owner_path(value: str, path_map: Dict[str, str], path_prefixes: List[tuple]) -> str:
+    if not isinstance(value, str) or not value:
+        return value
+    abs_value = os.path.abspath(value)
+    mapped = path_map.get(abs_value)
+    if mapped:
+        return mapped
+    for old_prefix, new_prefix in path_prefixes:
+        old_abs = os.path.abspath(old_prefix)
+        new_abs = os.path.abspath(new_prefix)
+        if abs_value == old_abs:
+            return new_abs
+        if abs_value.startswith(old_abs + os.sep):
+            return new_abs + abs_value[len(old_abs):]
+    return value
+
+
 class VectorRAG:
     """RAG system using ChromaDB vector storage with hybrid search."""
 
@@ -250,6 +267,75 @@ class VectorRAG:
             "failed_count": len(docs) - len(valid),
         }
 
+    def rename_owner(
+        self,
+        old_owner: str,
+        new_owner: str,
+        *,
+        path_map: Optional[Dict[str, str]] = None,
+        path_prefixes: Optional[List[tuple]] = None,
+    ) -> Dict[str, Any]:
+        """Rewrite existing RAG metadata after an auth username rename."""
+        if not self.healthy:
+            return {"success": False, "updated_count": 0, "message": "Collection not initialized"}
+
+        old_owner = (old_owner or "").strip().lower()
+        new_owner = (new_owner or "").strip().lower()
+        if not old_owner or not new_owner or old_owner == new_owner:
+            return {"success": True, "updated_count": 0, "message": "No owner rename needed"}
+
+        path_map = {os.path.abspath(k): os.path.abspath(v) for k, v in (path_map or {}).items()}
+        path_prefixes = path_prefixes or []
+        updated_ids = set()
+        failed_count = 0
+
+        for lane_name, collection in self._collections_for_delete():
+            try:
+                results = collection.get(
+                    where={"owner": old_owner},
+                    include=["metadatas"],
+                )
+            except Exception as e:
+                logger.warning("rename_owner metadata scan failed in %s lane: %s", lane_name, e)
+                failed_count += 1
+                continue
+
+            ids = results.get("ids") or []
+            metadatas = results.get("metadatas") or []
+            if not ids:
+                continue
+
+            new_metas = []
+            selected_ids = []
+            for doc_id, meta in zip(ids, metadatas):
+                if not isinstance(meta, dict):
+                    continue
+                next_meta = dict(meta)
+                if str(next_meta.get("owner", "")).strip().lower() == old_owner:
+                    next_meta["owner"] = new_owner
+                for key in ("source", "directory"):
+                    next_meta[key] = _rewrite_owner_path(next_meta.get(key), path_map, path_prefixes)
+                selected_ids.append(doc_id)
+                new_metas.append(next_meta)
+
+            if not selected_ids:
+                continue
+
+            try:
+                collection.update(ids=selected_ids, metadatas=new_metas)
+                updated_ids.update(selected_ids)
+            except Exception as e:
+                logger.warning("rename_owner metadata update failed in %s lane: %s", lane_name, e)
+                failed_count += len(selected_ids)
+
+        success = failed_count == 0
+        return {
+            "success": success,
+            "updated_count": len(updated_ids),
+            "failed_count": failed_count,
+            "message": f"Updated {len(updated_ids)} RAG chunk(s)",
+        }
+
     # ------------------------------------------------------------------
     # Search — hybrid: vector similarity + keyword overlap
     # ------------------------------------------------------------------

src/task_scheduler.py

@@ -9,6 +9,8 @@ import uuid
 from datetime import datetime, timedelta, timezone
 from typing import Any, Awaitable, Callable, Dict, Tuple
 
+from core.auth import RESERVED_USERNAMES
+
 logger = logging.getLogger(__name__)
 
 
@@ -17,6 +19,34 @@ def _utcnow() -> datetime:
     return datetime.now(timezone.utc).replace(tzinfo=None)
 
 
+# Shell/file tools a scheduled task's agent should be offered by default,
+# mirroring the chat agent (where these are on unless a privilege or global
+# setting turns them off). The RAG tool selector + ASSISTANT_ALWAYS_AVAILABLE
+# never include bash/python, so on a host with an empty/degraded tool-embedding
+# index a task could not run shell or Python even for an admin owner. Offering
+# them here is safe: stream_agent_loop's blocked_tools_for_owner() still strips
+# this whole group for non-admin multi-user owners, and only admits it for
+# admins and single-user (AUTH_ENABLED=false) deployments.
+TASK_DEFAULT_SHELL_TOOLS = frozenset({
+    "bash", "python", "read_file", "write_file", "edit_file",
+    "grep", "glob", "ls", "get_workspace",
+})
+
+
+def compose_task_relevant_tools(rag_tools, assistant_always, disabled_tools):
+    """Compose the relevant-tools set offered to a scheduled task's agent.
+
+    Unions the RAG-retrieved tools, the assistant's always-available set, and
+    the default shell/file group, then removes anything the task's crew
+    explicitly disabled via its `enabled_tools` allowlist. Per-owner admin
+    gating is applied later by stream_agent_loop (blocked_tools_for_owner).
+    """
+    tools = set(rag_tools) | set(assistant_always) | set(TASK_DEFAULT_SHELL_TOOLS)
+    if disabled_tools:
+        tools -= set(disabled_tools)
+    return tools
+
+
 # ── Shared TTL cache (singleflight) ────────────────────────────────────────
 # Multiple scheduled tasks firing in the same minute often need the same
 # external data (Miniflux unreads, MCP tool snapshots, etc.). This cache
@@ -1389,15 +1419,28 @@ class TaskScheduler:
             time_str = _utcnow().strftime("%A, %B %d %Y, %H:%M UTC")
         system_prompt = f"Current time: {time_str}\n\n{system_prompt}"
 
-        # Compute tool filter from CrewMember.enabled_tools if set
+        # Compute the disabled-tools set: the crew's enabled_tools allowlist
-        disabled_tools = None
+        # (inverted) plus the operator's global disabled_tools setting. The
+        # global list must be merged here — chat does the same merge before
+        # entering the agent loop (routes/chat_routes.py) — otherwise an admin
+        # or AUTH_ENABLED=false scheduled task would still see and call shell/
+        # file tools after the operator disabled them globally, because the
+        # prompt/schema/execution gates only enforce what is passed in.
+        disabled_tools: set[str] = set()
         if crew and crew.enabled_tools:
             try:
                 enabled = json.loads(crew.enabled_tools)
                 if isinstance(enabled, list) and enabled:
                     from src.tool_index import BUILTIN_TOOL_DESCRIPTIONS
                     all_tools = set(BUILTIN_TOOL_DESCRIPTIONS.keys())
-                    disabled_tools = all_tools - set(enabled)
+                    disabled_tools |= all_tools - set(enabled)
+            except Exception:
+                pass
+        try:
+            from src.settings import get_setting
+            _global_disabled = get_setting("disabled_tools", [])
+            if isinstance(_global_disabled, list):
+                disabled_tools.update(_global_disabled)
         except Exception:
             pass
 
@@ -1409,10 +1452,10 @@ class TaskScheduler:
             tool_idx = get_tool_index()
             if tool_idx:
                 rag_tools = tool_idx.get_tools_for_query(task.prompt or "", k=8)
-                relevant_tools = (rag_tools | ASSISTANT_ALWAYS_AVAILABLE)
+                relevant_tools = compose_task_relevant_tools(
-                if disabled_tools:
+                    rag_tools, ASSISTANT_ALWAYS_AVAILABLE, disabled_tools
-                    relevant_tools -= disabled_tools
+                )
-                logger.info(f"[assistant] RAG selected {len(rag_tools)} tools + {len(ASSISTANT_ALWAYS_AVAILABLE)} always-available = {len(relevant_tools)} total for '{task.name}'")
+                logger.info(f"[assistant] RAG selected {len(rag_tools)} tools + {len(ASSISTANT_ALWAYS_AVAILABLE)} always-available + shell/file defaults = {len(relevant_tools)} total for '{task.name}'")
         except Exception as e:
             logger.warning(f"[assistant] RAG tool selection failed, using all: {e}")
 
@@ -1420,7 +1463,7 @@ class TaskScheduler:
         try:
             result = await self._run_agent_loop(
                 endpoint_url, model, task, session_id,
-                system_prompt=system_prompt, disabled_tools=disabled_tools,
+                system_prompt=system_prompt, disabled_tools=disabled_tools or None,
                 relevant_tools=relevant_tools,
             )
         except Exception as e:
@@ -2221,7 +2264,7 @@ class TaskScheduler:
         # check-ins seeded, which then double-fire alongside the human user's
         # check-ins. This was the root cause of the duplicate 'Morning check-in'
         # rows we had to manually clean up.
-        if not owner or owner in {"internal-tool", "api", "demo", "system"}:
+        if not owner or owner in RESERVED_USERNAMES:
             logger.info(f"ensure_assistant_defaults: skip synthetic owner {owner!r}")
             return
         from core.database import SessionLocal, CrewMember, ScheduledTask

src/tool_execution.py

@@ -323,6 +323,24 @@ _MCP_TOOL_MAP = {
     "web_fetch":      ("web_fetch",  "web_fetch"),
     "generate_image": ("image_gen",  "generate_image"),
 }
+_EMAIL_MCP_OWNER_ARG = "_odysseus_owner"
+
+
+def _parse_qualified_mcp_args(tool: str, content: str) -> tuple[Dict, Optional[str]]:
+    raw = (content or "").strip()
+    if not raw:
+        return {}, None
+    try:
+        parsed = json.loads(raw)
+    except (json.JSONDecodeError, TypeError):
+        if tool.startswith("mcp__email__"):
+            return {}, "Email MCP tool arguments must be a JSON object."
+        return {}, None
+    if not isinstance(parsed, dict):
+        if tool.startswith("mcp__email__"):
+            return {}, "Email MCP tool arguments must be a JSON object."
+        return {}, None
+    return parsed, None
 
 
 def _parse_generate_image(content: str) -> Dict:
@@ -453,6 +471,8 @@ async def _direct_fallback(
     tool: str,
     content: str,
     progress_cb: Optional[Callable[[Dict], Awaitable[None]]] = None,
+    session_id: Optional[str] = None,
+    owner: Optional[str] = None,
 ) -> Optional[Dict]:
     _subproc_env = {
         **os.environ,
@@ -466,6 +486,8 @@ async def _direct_fallback(
         ctx = {
             "progress_cb": progress_cb,
             "subproc_env": _subproc_env,
+            "session_id": session_id,
+            "owner": owner,
         }
 
         from src.agent_tools import TOOL_HANDLERS
@@ -713,10 +735,13 @@ async def _execute_tool_block_impl(
             desc = f"bash (background): {short}"
             result = {
                 "output": (
-                    f"Started background job `{rec['id']}`. It is running detached — "
+                    f"Started background job `{rec['id']}`. It is running detached; "
                     f"do NOT wait for it or poll it. You will be automatically re-invoked "
                     f"with its full output when it finishes. Continue with other work, or "
-                    f"end your turn now and resume when the result arrives."
+                    f"end your turn now and resume when the result arrives. If the user "
+                    f"later asks to check progress or stop it, call the manage_bg_jobs "
+                    f"tool yourself (output or kill); do not tell them to run a tool "
+                    f"command, and do not surface raw tool syntax in your reply."
                 ),
                 "exit_code": 0,
                 "bg_job_id": rec["id"],
@@ -737,6 +762,11 @@ async def _execute_tool_block_impl(
         desc = f"{tool}: {first_line}"
         result = await _direct_fallback(tool, content, progress_cb=progress_cb) \
             or {"error": f"{tool}: execution failed", "exit_code": 1}
+    elif tool == "manage_bg_jobs":
+        # Inspect/kill detached `bash` jobs; needs session_id to scope to chat.
+        desc = f"manage_bg_jobs: {content.split(chr(10))[0][:80]}"
+        result = await _direct_fallback(tool, content, session_id=session_id, owner=owner) \
+            or {"error": "manage_bg_jobs: execution failed", "exit_code": 1}
     elif tool in ("create_document", "update_document", "edit_document",
                   "suggest_document", "manage_documents"):
         desc = f"{tool}: {content.split(chr(10))[0][:80]}"
@@ -748,10 +778,24 @@ async def _execute_tool_block_impl(
         query = content.split("\n")[0].strip()
         desc = f"search_chats: {query[:80]}"
         result = await do_search_chats(query, owner=owner)
-    elif tool in ("chat_with_model", "create_session", "list_sessions",
+    elif tool in ("chat_with_model", "ask_teacher", "list_models"):
-                  "send_to_session", "pipeline",
+        # Migrated to the agent_tools registry (#3629): dispatched through
-                  "manage_session", "manage_memory", "list_models",
+        # TOOL_HANDLERS with the owner/session ctx these tools need, instead
-                  "ui_control", "ask_teacher"):
+        # of the legacy dispatch_ai_tool elif. The impls live in
+        # src/agent_tools/model_interaction_tools.py.
+        first_line = content.split(chr(10))[0].strip()[:60]
+        desc = f"{tool}: {first_line}" if first_line else tool
+        result = await _document_tool_dispatch(tool, content, session_id, owner) \
+            or {"error": f"{tool}: execution failed", "exit_code": 1}
+    elif tool in ("create_session", "list_sessions", "send_to_session", "manage_session"):
+        # Migrated to the agent_tools registry (#3629): dispatched through
+        # TOOL_HANDLERS with the owner/session ctx these tools need. The impls
+        # live in src/agent_tools/session_tools.py.
+        first_line = content.split(chr(10))[0].strip()[:60]
+        desc = f"{tool}: {first_line}" if first_line else tool
+        result = await _document_tool_dispatch(tool, content, session_id, owner) \
+            or {"error": f"{tool}: execution failed", "exit_code": 1}
+    elif tool in ("pipeline", "manage_memory", "ui_control"):
         from src.ai_interaction import dispatch_ai_tool
         desc, result = await dispatch_ai_tool(tool, content, session_id, owner=owner)
     elif tool == "manage_tasks":
@@ -858,11 +902,14 @@ async def _execute_tool_block_impl(
         # MCP tool dispatch
         mcp = get_mcp_manager()
         if mcp:
-            try:
-                args = json.loads(content) if content.strip().startswith("{") else {}
-            except (json.JSONDecodeError, TypeError):
-                args = {}
             desc = f"mcp: {tool}"
+            args, parse_error = _parse_qualified_mcp_args(tool, content)
+            if parse_error:
+                result = {"error": parse_error, "exit_code": 1}
+            else:
+                if tool.startswith("mcp__email__") and owner:
+                    args = dict(args)
+                    args[_EMAIL_MCP_OWNER_ARG] = owner
                 result = await mcp.call_tool(tool, args)
         else:
             desc = f"mcp: {tool}"

src/tool_implementations.py

@@ -12,12 +12,24 @@ import os
 import re
 from typing import Any, Dict, List, Optional
 
+from fastapi import HTTPException
 from src.constants import MAX_READ_CHARS, DEEP_RESEARCH_DIR, VAULT_FILE
 from src.tool_utils import get_mcp_manager
 from core.constants import internal_api_base
+from routes._validators import validate_remote_host, validate_ssh_port
 
 logger = logging.getLogger(__name__)
 
+
+def _string_arg(value: Any) -> str:
+    return "" if value is None else str(value).strip()
+
+
+def _validate_cookbook_ssh_target(remote_host: Any, ssh_port: Any = "") -> tuple[str, str]:
+    remote = validate_remote_host(_string_arg(remote_host) or None) or ""
+    sport = validate_ssh_port(_string_arg(ssh_port) or None) or ""
+    return remote, sport
+
 # ---------------------------------------------------------------------------
 # Active email state
 # ---------------------------------------------------------------------------
@@ -645,6 +657,137 @@ async def do_manage_endpoints(content: str, owner: Optional[str] = None) -> Dict
 # MCP server management tool
 # ---------------------------------------------------------------------------
 
+# Parallel to routes/cookbook_helpers._validate_serve_cmd but deliberately the
+# opposite policy: that gate guards an admin-only serve command and allows
+# interpreters (python3/etc) because model-serving needs them, whereas this is
+# the model/prompt-injection-reachable manage_mcp path, so interpreters and
+# runners are denied here.
+#
+# Commands that can execute arbitrary code regardless of their arguments. These
+# are NEVER accepted on the manage_mcp agent path, even if an operator lists one
+# in ODYSSEUS_MCP_ALLOWED_COMMANDS -- a stdio server that genuinely needs an
+# interpreter or package runner must be registered via the trusted admin route.
+_MCP_DENIED_COMMANDS = frozenset({
+    "sh", "bash", "zsh", "fish", "dash", "ksh", "csh", "tcsh", "ash", "busybox",
+    "cmd", "command.com", "powershell", "pwsh",
+    "python", "pypy", "node", "nodejs", "deno", "bun", "ruby", "jruby",
+    "perl", "raku", "php", "lua", "luajit", "tclsh", "wish", "expect", "rscript",
+    "groovy", "scala", "elixir", "erl", "iex", "java", "javac", "jshell", "jbang",
+    "kotlin", "kotlinc", "dotnet", "mono", "swift", "osascript", "tsx", "ts-node",
+    "npx", "bunx", "uvx", "pipx", "npm", "pnpm", "yarn", "pip", "uv",
+    "gem", "cargo", "go", "bundle", "poetry", "conda", "mamba", "brew",
+    "apt", "apt-get", "yum", "dnf", "pacman", "apk",
+    "env", "xargs", "nohup", "setsid", "nice", "ionice", "time", "timeout",
+    "watch", "stdbuf", "unbuffer", "script", "ssh", "scp", "sshpass", "sudo",
+    "doas", "su", "make", "cmake", "docker", "podman", "kubectl", "find",
+    "awk", "gawk", "sed", "vi", "vim", "nvim", "emacs", "ed", "tee", "eval",
+})
+
+# Argv flags that make even an allowlisted binary execute inline code. Matched
+# by prefix so glued forms (-cimport os, --eval=...) are caught, not just the
+# exact-token form.
+_MCP_CODE_EXEC_SHORT_FLAGS = ("-c", "-e", "-m")
+_MCP_CODE_EXEC_LONG_FLAGS = ("--eval", "--exec", "--print", "--module", "--command", "--require")
+
+_MCP_URL_SCHEMES = ("http://", "https://", "ftp://", "ftps://", "file://", "data:", "jar:", "blob:")
+
+# Shell metacharacters refused in command/args. Args are passed as an argv list
+# (no shell), but refusing these keeps the surface narrow and obvious.
+_MCP_SHELL_METACHARS = set(";|&$`><\n\r")
+
+# Env vars that let a child process load attacker-supplied code before main().
+_MCP_DANGEROUS_ENV = frozenset({
+    "LD_PRELOAD", "LD_LIBRARY_PATH", "LD_AUDIT", "DYLD_INSERT_LIBRARIES",
+    "DYLD_LIBRARY_PATH", "DYLD_FRAMEWORK_PATH", "PYTHONPATH", "PYTHONSTARTUP",
+    "PYTHONHOME", "PYTHONEXECUTABLE", "NODE_OPTIONS", "NODE_PATH", "BASH_ENV",
+    "ENV", "SHELLOPTS", "PERL5LIB", "PERL5OPT", "RUBYOPT", "RUBYLIB", "GEM_PATH",
+    "R_PROFILE", "R_HOME", "PATH", "IFS", "PROMPT_COMMAND",
+})
+
+
+def _mcp_allowed_commands() -> set:
+    """Operator-configured allowlist of safe MCP launcher basenames for the agent
+    path. Empty by default; set ODYSSEUS_MCP_ALLOWED_COMMANDS (comma-separated)
+    to opt specific trusted binaries in. Denied commands are rejected even if
+    listed here."""
+    raw = os.environ.get("ODYSSEUS_MCP_ALLOWED_COMMANDS", "")
+    return {c.strip().lower() for c in raw.split(",") if c.strip()}
+
+
+def _validate_mcp_command(command, args, env) -> Optional[str]:
+    """Validate a model-supplied stdio MCP registration. Returns an error string
+    if it must be rejected, else None.
+
+    Closes the RCE where manage_mcp 'add' passed prompt-injection-controlled
+    command/args/env straight to a subprocess spawn (issue #438): a payload
+    smuggled into a skill description, memory entry, fetched page, or email body
+    could register a stdio server running arbitrary code as the app UID.
+    """
+    if not isinstance(command, str) or not command.strip():
+        return "command must be a non-empty string"
+    command = command.strip()
+    if "/" in command or "\\" in command:
+        return "command must be a bare executable name, not a path"
+    if any(ch in _MCP_SHELL_METACHARS for ch in command):
+        return "command contains shell metacharacters"
+    base = command.lower()
+    if base.endswith(".exe") or base.endswith(".cmd") or base.endswith(".bat"):
+        base = base.rsplit(".", 1)[0]
+    # Canonicalize a trailing version suffix so versioned aliases collapse to the
+    # family name (python3.11 -> python, node18 -> node, pip3 -> pip); both the
+    # raw basename and the canonical form are denied, so an operator cannot
+    # accidentally allowlist a runtime alias back into the path.
+    canon = re.sub(r"[-_.]?\d+(?:\.\d+)*$", "", base)
+    if base in _MCP_DENIED_COMMANDS or canon in _MCP_DENIED_COMMANDS:
+        return (
+            f"command '{command}' is not allowed on the agent MCP path: "
+            "interpreters, runtimes, package runners, and shells can execute "
+            "arbitrary code. Register such a server via the admin route instead."
+        )
+    if base not in _mcp_allowed_commands():
+        return (
+            f"command '{command}' is not in the MCP allowlist. Add it to "
+            "ODYSSEUS_MCP_ALLOWED_COMMANDS if you trust it, or register the "
+            "server via the admin route."
+        )
+
+    if args is not None:
+        if isinstance(args, str):
+            try:
+                args = json.loads(args)
+            except Exception:
+                return "args must be a JSON list"
+        if not isinstance(args, list):
+            return "args must be a list"
+        for a in args:
+            if not isinstance(a, str):
+                return "args must all be strings"
+            s = a.strip()
+            low = s.lower()
+            if any(s == f or s.startswith(f) for f in _MCP_CODE_EXEC_SHORT_FLAGS):
+                return f"arg '{a}' is a code-execution flag and is not allowed"
+            if any(low == f or low.startswith(f + "=") for f in _MCP_CODE_EXEC_LONG_FLAGS):
+                return f"arg '{a}' is a code-execution flag and is not allowed"
+            if any(low.startswith(u) for u in _MCP_URL_SCHEMES):
+                return f"arg '{a}' is a remote URL and is not allowed"
+            if any(ch in _MCP_SHELL_METACHARS for ch in a):
+                return f"arg '{a}' contains shell metacharacters"
+
+    if env:
+        if isinstance(env, str):
+            try:
+                env = json.loads(env)
+            except Exception:
+                return "env must be a JSON object"
+        if not isinstance(env, dict):
+            return "env must be an object"
+        for k in env:
+            if str(k).strip().upper() in _MCP_DANGEROUS_ENV:
+                return f"env var '{k}' can inject code into the child process and is not allowed"
+
+    return None
+
+
 async def do_manage_mcp(content: str, owner: Optional[str] = None) -> Dict:
     """Manage MCP servers: list, add, delete, enable, disable, reconnect."""
     try:
@@ -684,6 +827,12 @@ async def do_manage_mcp(content: str, owner: Optional[str] = None) -> Dict:
         env = args.get("env", {})
         if not name or not command:
             return {"error": "name and command are required", "exit_code": 1}
+        # Validate BEFORE any DB write or spawn: a rejected registration must
+        # leave no enabled row (which would otherwise auto-reconnect on restart)
+        # and must not attempt a connection.
+        _mcp_err = _validate_mcp_command(command, cmd_args, env)
+        if _mcp_err:
+            return {"error": f"manage_mcp: refused unsafe server registration: {_mcp_err}", "exit_code": 1}
         sid = str(_uuid.uuid4())[:8]
         db = SessionLocal()
         try:
@@ -2888,6 +3037,10 @@ async def _cookbook_kill_session(session_id: str, *, remote_host: str = "",
             break
 
     if remote:
+        try:
+            remote, sport = _validate_cookbook_ssh_target(remote, sport)
+        except HTTPException as e:
+            return {"error": str(getattr(e, "detail", e)), "exit_code": 1}
         _pf = f"-p {shlex.quote(str(sport))} " if sport and str(sport) != "22" else ""
         cmd = (
             f"ssh -o ConnectTimeout=5 -o StrictHostKeyChecking=no "
@@ -2976,8 +3129,8 @@ async def do_tail_serve_output(content: str, owner: Optional[str] = None) -> Dic
         tail = 400
     tail = max(20, min(tail, 4000))
     headers = _internal_headers()
-    remote = (args.get("remote_host") or args.get("host") or "").strip()
+    remote = _string_arg(args.get("remote_host") or args.get("host"))
-    sport = (args.get("ssh_port") or "").strip()
+    sport = _string_arg(args.get("ssh_port"))
     # Resolve host from cookbook state if caller didn't pass one — same
     # lookup _cookbook_kill_session uses.
     if not remote:
@@ -2995,6 +3148,12 @@ async def do_tail_serve_output(content: str, owner: Optional[str] = None) -> Dic
                     if not sport:
                         sport = t.get("sshPort") or ""
                     break
+    if remote:
+        try:
+            remote, sport = _validate_cookbook_ssh_target(remote, sport)
+        except HTTPException as e:
+            return {"error": str(getattr(e, "detail", e)), "exit_code": 1}
+
     # Prefer the persisted /tmp/odysseus-tmux/SESSION.log file over the
     # live tmux pane. The pane is what the user would see scrolling on
     # their screen — including the post-crash neofetch banner and the
@@ -3172,7 +3331,7 @@ async def do_adopt_served_model(content: str, owner: Optional[str] = None) -> Di
     except ValueError:
         return {"error": "Invalid JSON arguments", "exit_code": 1}
 
-    host = (args.get("host") or args.get("remote_host") or "").strip()
+    host = _string_arg(args.get("host") or args.get("remote_host"))
     sess = (args.get("tmux_session") or args.get("session_id") or "").strip()
     model = (args.get("model") or args.get("repo_id") or "").strip()
     port = args.get("port") or 8000
@@ -3183,6 +3342,12 @@ async def do_adopt_served_model(content: str, owner: Optional[str] = None) -> Di
         return {"error": "tmux_session and model are required", "exit_code": 1}
 
     # Verify tmux session exists on the target host
+    if host:
+        try:
+            host, _ = _validate_cookbook_ssh_target(host)
+        except HTTPException as e:
+            return {"error": str(getattr(e, "detail", e)), "exit_code": 1}
+
     headers = _internal_headers()
     if host:
         check = f"ssh -o ConnectTimeout=5 -o StrictHostKeyChecking=no {shlex.quote(host)} 'tmux has-session -t {shlex.quote(sess)} 2>&1'"

src/tool_index.py

@@ -94,6 +94,7 @@ BUILTIN_TOOL_DESCRIPTIONS: Dict[str, str] = {
     "manage_endpoints": "Endpoint management: list, add, delete, enable, or disable model API endpoints.",
     "manage_mcp": "MCP server management: list, add, delete, reconnect servers, or list available tools.",
     "manage_webhooks": "Webhook management: list, add, delete, enable, or disable webhooks.",
+    "api_call": "Call a configured API integration by name (Home Assistant, Miniflux, Gitea, Linkding, Jellyfin, RSS reader, git forge, bookmark manager, smart home, or any other registered service). Make a GET/POST/PUT/PATCH/DELETE request to the integration's endpoint path, with an optional JSON body. Use whenever the user asks to query or control one of their connected integrations/services.",
     "manage_tokens": "API token management: list, create, or delete API access tokens.",
     "manage_documents": "List, read, delete, or tidy documents in the editor panel. action='list' returns clickable rows (most-recent first) so the user can open any doc by clicking. action='read' (aka view/open/get) with document_id returns the content; supports offset=<N> + limit=<N> to page through large docs (response includes next_offset when more remains, so you can keep calling with offset=next_offset). action='delete' with document_id removes a doc (only way to delete). Use this for ANY 'show/read/list/open my documents/docs/files/notes' request — never shell or curl.",
     "manage_research": "List, read/open, or delete saved DEEP RESEARCH results from the Library. action='list' returns clickable [query](#research-<id>) rows (most-recent first). action='read' (aka open/view/get) with id returns the report + sources. action='delete' with id removes it. Use this for ANY 'open/read/find/delete my research / that report / the research on X' request. NOTE: this is for EXISTING research; to START new research use trigger_research.",
@@ -134,6 +135,7 @@ BUILTIN_TOOL_DESCRIPTIONS: Dict[str, str] = {
     "app_api": "Generic loopback to allowed Odysseus internal endpoints. Use this when the user wants something the UI can do but there's no named tool for it. Covers calendar, gallery, library/documents, memory, notes, tasks, settings, research, compare, cookbook GPUs/state — allowed UI buttons hit /api/* endpoints and you can hit them too. Sensitive auth/user/admin/shell paths and host-control Cookbook mutation routes are blocked; do NOT use app_api for shell commands, package installs, engine rebuilds, or PID signalling. Use named command tooling for shell commands. action='endpoints' with filter=<keyword> lists available endpoints. action='call' takes method+path+body. Hits same routes the UI uses — auth flows free. NOTE: themes are NOT an API endpoint — use the ui_control tool (create_theme / set_theme), not app_api. SESSIONS/CHATS: do NOT use app_api for these — GET /api/sessions returns EMPTY for tool calls (it's owner-filtered and tool calls authenticate as a different identity). EMAIL ACCOUNTS: do NOT use /api/email/accounts via app_api; use list_email_accounts, list_emails, and read_email instead. To list/rename/archive/delete/fork chats use the list_sessions and manage_session tools instead.",
     "edit_image": "Edit an image in the gallery: upscale (increase resolution), remove background (rembg), inpaint (fill selected area), or harmonize (blend edits). Specify image ID and action.",
     "trigger_research": "Start a deep research job on any topic — appears in the Deep Research sidebar, streams progress, produces a detailed report. Use for 'research X', 'look into Y', 'do deep research on Z', 'investigate'. NOT a scheduled task — it runs now and surfaces in the sidebar.",
+    "manage_bg_jobs": "Inspect and control detached background `bash` jobs (the ones started with a `#!bg` marker). action='list' shows this chat's jobs (id/status/age/command); action='output' returns a job's captured output so far (check on a long-running job, or re-read a finished one); action='kill' stops a runaway job by id. Use for 'is the background job done', 'check on that job', 'show the build output', 'kill the background job', 'stop the bg task'. output/kill need a job_id from list.",
 }
 
 
@@ -348,6 +350,12 @@ class ToolIndex:
             {"list_email_accounts", "list_emails", "read_email", "send_email", "reply_to_email", "bulk_email", "delete_email", "archive_email", "mark_email_read", "resolve_contact", "ui_control"},
         frozenset({"calendar", "event", "meeting", "schedule", "appointment"}):
             {"manage_calendar"},
+        # Detached background `bash` jobs (#!bg): check on / read output / kill.
+        frozenset({"background job", "background jobs", "bg job", "bg jobs",
+                   "background task", "is the job done", "check the job",
+                   "check on that job", "job output", "kill the job",
+                   "kill the background", "stop the background", "running job"}):
+            {"manage_bg_jobs"},
         frozenset({"note", "todo", "reminder", "remind", "checklist", "remember to"}):
             {"manage_notes"},
         # Chat/session management. "rename" alone maps to documents below, so a
@@ -414,6 +422,14 @@ class ToolIndex:
                    "my settings", "change setting", "change a setting", "set setting",
                    "preference", "preferences", "configure"}):
             {"manage_settings", "ui_control"},
+        # API-integration intent → the api_call tool. Mirrors the agent-loop
+        # "integrations" domain so api_call still surfaces on the retrieval and
+        # keyword-fallback paths (not just the deterministic domain seed) when a
+        # user names a connected service.
+        frozenset({"api_call", "api call", "integration", "integrations",
+                   "home assistant", "homeassistant", "miniflux", "gitea",
+                   "linkding", "jellyfin"}):
+            {"api_call"},
         # Managing EXISTING research in the Library — open/read/find/delete.
         frozenset({"my research", "the research", "research on", "open research",
                    "read research", "find research", "delete research",

src/tool_parsing.py

@@ -175,6 +175,9 @@ _TOOL_NAME_MAP = {
     "notes": "manage_notes",
     "todo": "manage_notes",
     "todos": "manage_notes",
+    "manage_bg_jobs": "manage_bg_jobs",
+    "bg_jobs": "manage_bg_jobs",
+    "background_jobs": "manage_bg_jobs",
 }
 
 _MISFENCED_WEB_TOOL_NAMES = {

src/tool_schemas.py

@@ -1188,6 +1188,21 @@ FUNCTION_TOOL_SCHEMAS = [
             }
         }
     },
+    {
+        "type": "function",
+        "function": {
+            "name": "manage_bg_jobs",
+            "description": "Inspect and control detached background `bash` jobs (started with the `#!bg` marker). action='list' shows this chat's jobs with id/status/age/command; action='output' returns a job's captured output so far (use for a still-running job, or to re-read a finished one); action='kill' terminates a runaway job's process tree instead of waiting out its max-runtime. output and kill need job_id from list.",
+            "parameters": {
+                "type": "object",
+                "properties": {
+                    "action": {"type": "string", "enum": ["list", "output", "kill"], "description": "list | output | kill (default: list)"},
+                    "job_id": {"type": "string", "description": "Background job id (required for output/kill; from action='list')"},
+                },
+                "required": ["action"]
+            }
+        }
+    },
 ]
 
 
@@ -1206,23 +1221,26 @@ def function_call_to_tool_block(name: str, arguments: str) -> Optional[ToolBlock
         logger.error(f"Failed to parse function call arguments for {name}: {arguments}")
         return None
 
+    tool_type = _TOOL_NAME_MAP.get(name, name)
+    _BUILTIN_EMAIL_TOOLS = {"list_email_accounts", "send_email", "list_emails", "read_email", "reply_to_email",
+                            "archive_email", "delete_email", "mark_email_read", "bulk_email", "download_attachment"}
+
     # Some models emit valid JSON that isn't an object (e.g. a bare array
-    # ["ls -la"], string, or number) as the function arguments. Every branch
+    # ["ls -la"], string, or number) as function arguments. Most local tools keep
-    # below assumes a dict and calls args.get(...), so a non-dict would raise
+    # the legacy empty-object coercion for stream robustness, but email MCP tools
-    # AttributeError and abort the whole agent stream. Coerce to {} instead.
+    # must fail closed so a malformed call cannot read the default mailbox.
     if not isinstance(args, dict):
+        if tool_type.startswith("mcp__email__") or name in _BUILTIN_EMAIL_TOOLS:
+            logger.warning(f"Non-object email function call arguments for {name}: {args!r}; rejecting")
+            return None
         logger.warning(f"Non-object function call arguments for {name}: {args!r}; treating as empty")
         args = {}
 
-    tool_type = _TOOL_NAME_MAP.get(name, name)
-
     # Allow MCP tools through (namespaced as mcp__serverid__toolname)
     if tool_type.startswith("mcp__"):
         content = json.dumps(args) if args else "{}"
         return ToolBlock(tool_type, content)
     # Email tools are implemented as MCP — route them to email
-    _BUILTIN_EMAIL_TOOLS = {"list_email_accounts", "send_email", "list_emails", "read_email", "reply_to_email",
-                            "archive_email", "delete_email", "mark_email_read", "bulk_email", "download_attachment"}
     if name in _BUILTIN_EMAIL_TOOLS:
         return ToolBlock(f"mcp__email__{name}", json.dumps(args) if args else "{}")
     if tool_type not in TOOL_TAGS:

src/tool_security.py

@@ -14,6 +14,7 @@ logger = logging.getLogger(__name__)
 NON_ADMIN_BLOCKED_TOOLS = {
     "bash",
     "python",
+    "manage_bg_jobs",
     "read_file",
     "write_file",
     "edit_file",
@@ -114,6 +115,8 @@ _PLAN_MODE_KNOWN_MUTATORS = {
     # Shell is never read-only-safe; block it explicitly so it stays out of plan
     # mode even if the schema list fails to load.
     "bash", "python",
+    # Controls shell processes (kill); plan mode can't run bash anyway.
+    "manage_bg_jobs",
 }
 
 

static/index.html

@@ -1913,7 +1913,7 @@
             <h2><svg width="14" height="14" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" style="vertical-align:-2px;margin-right:5px;opacity:0.6"><rect x="3" y="11" width="18" height="11" rx="2" ry="2"/><path d="M7 11V7a5 5 0 0 1 10 0v4"/></svg>Change Password</h2>
             <div class="settings-col">
               <input id="settings-pw-current" type="password" placeholder="Current password" autocomplete="current-password" style="padding:6px 8px;background:var(--bg);border:1px solid var(--border);border-radius:4px;color:var(--fg);font-family:inherit;font-size:12px;">
-              <input id="settings-pw-new" type="password" placeholder="New password (min 8)" autocomplete="new-password" style="padding:6px 8px;background:var(--bg);border:1px solid var(--border);border-radius:4px;color:var(--fg);font-family:inherit;font-size:12px;">
+              <input id="settings-pw-new" type="password" placeholder="New password" autocomplete="new-password" style="padding:6px 8px;background:var(--bg);border:1px solid var(--border);border-radius:4px;color:var(--fg);font-family:inherit;font-size:12px;">
               <input id="settings-pw-confirm" type="password" placeholder="Confirm new password" autocomplete="new-password" style="padding:6px 8px;background:var(--bg);border:1px solid var(--border);border-radius:4px;color:var(--fg);font-family:inherit;font-size:12px;">
               <div class="settings-row" style="margin-top:2px;justify-content:flex-end;">
                 <span id="settings-pw-msg" style="font-size:11px;margin-right:auto;"></span>
@@ -2049,7 +2049,7 @@
             <h2><svg width="14" height="14" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" style="vertical-align:-2px;margin-right:5px;opacity:0.6"><path d="M16 21v-2a4 4 0 0 0-4-4H5a4 4 0 0 0-4 4v2"/><circle cx="8.5" cy="7" r="4"/><line x1="20" y1="8" x2="20" y2="14"/><line x1="23" y1="11" x2="17" y2="11"/></svg>Add User</h2>
             <div class="admin-add-form">
               <input id="adm-newUsername" type="text" placeholder="Username">
-              <input id="adm-newPassword" type="password" placeholder="Password (min 8)">
+              <input id="adm-newPassword" type="password" placeholder="Password">
               <div class="admin-switch-inline" title="Grant full admin access"><label class="admin-switch"><input type="checkbox" id="adm-newIsAdmin"><span class="admin-slider"></span></label> Admin</div>
             </div>
             <div class="settings-row" style="margin-top:6px;">

static/js/admin.js

@@ -13,6 +13,7 @@ let modalEl = null;
 // the endpoints list can flash a glow on that row. Cleared once the
 // animation fires.
 let _recentlyAddedEpId = null;
+let _authPolicy = { password_min_length: 8, reserved_usernames: [] };
 
 function el(id) { return document.getElementById(id); }
 function esc(s) { return uiModule.esc(s); }
@@ -343,6 +344,15 @@ function initSignupToggle() {
 }
 
 function initAddUser() {
+  fetch('/api/auth/policy', { credentials: 'same-origin' })
+    .then(r => r.ok ? r.json() : null)
+    .then(policy => {
+      if (!policy) return;
+      _authPolicy = policy;
+      const admPw = el('adm-newPassword');
+      if (admPw) admPw.placeholder = `Password (min ${policy.password_min_length})`;
+    })
+    .catch(() => {});
   el('adm-addBtn').addEventListener('click', async () => {
     const msg = el('adm-addMsg');
     msg.textContent = ''; msg.className = '';
@@ -350,7 +360,8 @@ function initAddUser() {
     const password = el('adm-newPassword').value;
     const is_admin = el('adm-newIsAdmin').checked;
     if (!username) { msg.textContent = 'Username required'; msg.className = 'admin-error'; return; }
-    if (password.length < 8) { msg.textContent = 'Password must be at least 8 characters'; msg.className = 'admin-error'; return; }
+    if (password.length < _authPolicy.password_min_length) { msg.textContent = `Password must be at least ${_authPolicy.password_min_length} characters`; msg.className = 'admin-error'; return; }
+    if (_authPolicy.reserved_usernames.includes(username.toLowerCase())) { msg.textContent = 'This username is reserved'; msg.className = 'admin-error'; return; }
     el('adm-addBtn').disabled = true;
     try {
       const res = await fetch('/api/auth/users', { method: 'POST', credentials: 'same-origin', headers: { 'Content-Type': 'application/json' }, body: JSON.stringify({ username, password, is_admin }) });
@@ -1745,7 +1756,6 @@ const TOOL_META = {
   manage_skills:     { name: 'Skills',           desc: 'Learn and use procedures',        cat: 'Knowledge',  ctx: '~200' },
   manage_rag:        { name: 'RAG / Docs',       desc: 'Query indexed documents',         cat: 'Knowledge',  ctx: '~150' },
   chat_with_model:   { name: 'Chat with Model',  desc: 'Talk to another AI model',        cat: 'Multi-Agent', ctx: '~200' },
-  second_opinion:    { name: 'Second Opinion',   desc: 'Get another model\'s take',       cat: 'Multi-Agent', ctx: '~150' },
   pipeline:          { name: 'Pipeline',         desc: 'Multi-step AI workflows',         cat: 'Multi-Agent', ctx: '~200' },
   ask_teacher:       { name: 'Ask Teacher',      desc: 'Query a more capable model',      cat: 'Multi-Agent', ctx: '~150' },
   send_to_session:   { name: 'Send to Session',  desc: 'Send message to another chat',    cat: 'Sessions',   ctx: '~100' },

static/js/assistant.js

@@ -125,7 +125,7 @@ const TOOL_GROUPS = {
   'Knowledge': ['web_search', 'read_file', 'manage_memory', 'manage_rag', 'search_chats'],
   'Code': ['bash', 'python', 'write_file'],
   'Documents': ['create_document', 'edit_document', 'update_document', 'suggest_document'],
-  'AI & Models': ['chat_with_model', 'second_opinion', 'ask_teacher', 'pipeline', 'list_models', 'generate_image'],
+  'AI & Models': ['chat_with_model', 'ask_teacher', 'pipeline', 'list_models', 'generate_image'],
   'System': ['manage_session', 'manage_endpoints', 'manage_mcp', 'manage_settings', 'manage_skills', 'manage_webhooks', 'manage_tokens', 'manage_documents', 'create_session', 'list_sessions', 'send_to_session', 'ui_control'],
 };
 

static/js/calendar.js

@@ -413,7 +413,7 @@ function _calEventFg(ev) {
 // Returns '' for normal solid-color events.
 function _calItemBgStyle(ev) {
   if (!_isCalBgImage(ev.color)) return '';
-  const url = _calBgImageUrl(ev.color).replace(/'/g, "\\'");
+  const url = _calBgImageUrl(ev.color).replace(/'/g, "\\'").replace(/"/g, "%22");
   return `background-image: linear-gradient(color-mix(in srgb, var(--bg) 70%, transparent), color-mix(in srgb, var(--bg) 70%, transparent)), url('${url}'); background-size: cover; background-position: center;`;
 }
 
@@ -1260,7 +1260,7 @@ async function _renderWeek() {
       // events keep the original tinted treatment.
       let bgDecl;
       if (_isCalBgImage(ev.color)) {
-        const _url = _calBgImageUrl(ev.color).replace(/'/g, "\\'");
+        const _url = _calBgImageUrl(ev.color).replace(/'/g, "\\'").replace(/"/g, "%22");
         bgDecl = `background-image: linear-gradient(color-mix(in srgb, var(--bg) 55%, transparent), color-mix(in srgb, var(--bg) 55%, transparent)), url('${_url}'); background-size: cover; background-position: center;`;
       } else {
         bgDecl = `background:color-mix(in srgb, ${_calColor(ev)} 18%, var(--bg));`;

static/js/chatRenderer.js

@@ -635,8 +635,8 @@ export function applyModelColor(roleEl, modelName) {
       popup.className = 'ctx-popup';
       let html = '<div style="font-weight:600;margin-bottom:6px;color:var(--fg);display:flex;align-items:center;gap:6px;">';
       if (logoHtml) html += '<span class="role-provider-logo" style="opacity:0.7">' + logoHtml + '</span>';
-      html += short + '</div>';
+      html += uiModule.esc(short) + '</div>';
-      html += '<div><span class="ctx-label">Model</span> ' + modelName.split('/').pop() + '</div>';
+      html += '<div><span class="ctx-label">Model</span> ' + uiModule.esc(modelName.split('/').pop()) + '</div>';
       // Provider = the serving endpoint, distinct from the model vendor/logo
       // (e.g. the same model via OpenRouter vs Copilot vs Anthropic direct).
       const _epUrl = (window.sessionModule && window.sessionModule.getCurrentEndpointUrl)

static/js/cookbook-deps-recipes.js

@@ -56,7 +56,7 @@ const _RECIPES = [
     match: () => true,
     variants: {
       pip:    { commands: ['CMAKE_ARGS="-DGGML_CUDA=on" uv pip install -U "llama-cpp-python[server]"'] },
-      docker: { commands: ['docker pull ghcr.io/ggerganov/llama.cpp:server-cuda'] },
+      docker: { commands: ['docker pull ghcr.io/ggml-org/llama.cpp:server-cuda'] },
     },
   },
 ];

static/js/cookbook-diagnosis.js

@@ -578,24 +578,50 @@ export const ERROR_PATTERNS = [
     ],
   },
   {
-    // Tail-only + healthy-server suppression. tmux capture-pane returns the
+    // Dependency-install (pip) build failure — a required package failed to
-    // entire scrollback every poll, so a one-shot startup traceback would
+    // build its wheel (common when an old sdist's setup.py breaks on a newer
-    // otherwise stick on the panel forever even while the server happily
+    // Python, e.g. basicsr on 3.13). This is an install problem, NOT a serve
-    // serves /v1/models. Only fire if the traceback is in recent output AND
+    // problem, so it must never suggest killing vLLM.
-    // the server isn't currently logging healthy traffic.
+    match: (text) => {
+      const TAIL = text.slice(-6000);
+      // A serve script can run a fallback build and then start serving fine —
+      // don't flag a stale build error once the server is up.
+      if (/Application startup complete|"(?:GET|POST)\s+\/v1\/[^"]+ HTTP\/[\d.]+"\s*2\d\d|Uvicorn running on|server is listening on https?:\/\//i.test(TAIL)) return false;
+      return /Failed to build\b|subprocess-exited-with-error|Could not build wheels|metadata-generation-failed/i.test(TAIL);
+    },
+    message: 'A dependency failed to build during install — usually an older package whose build breaks on this Python version, not a server problem. The install did not finish.',
+    suggestion: 'Suggested action: check the captured output for the package that failed to build; it may need a newer release or a patch to install on this Python version.',
+    fixes: [],
+  },
+  {
+    // vLLM-specific traceback: only offer the kill-processes recovery when the
+    // output is actually about vLLM. Tail-only + healthy-server suppression so
+    // a one-shot startup traceback doesn't stick on the panel forever while
+    // the server happily serves /v1/models.
     match: (text) => {
       const TAIL = text.slice(-4096);
       if (!/Traceback \(most recent call last\)/i.test(TAIL)) return false;
-      // Healthy markers in the tail mean whatever blew up has been recovered
-      // from — the server is up and answering requests.
       if (/Application startup complete|"GET \/v1\/[^"]+ HTTP\/[\d.]+" 2\d\d|Uvicorn running on/i.test(TAIL)) return false;
-      return true;
+      return /vllm/i.test(TAIL);
     },
-    message: 'Python traceback detected — may be a handled error, check logs.',
+    message: 'A vLLM process hit a Python traceback and may be wedged.',
     fixes: [
       { label: 'Kill vLLM processes', action: (panel) => _runQuickCmd(panel, 'pkill -f vllm') },
     ],
   },
+  {
+    // Generic traceback (not vLLM, not a pip build): surface it without
+    // suggesting an unrelated vLLM kill. Same tail-only + healthy suppression.
+    match: (text) => {
+      const TAIL = text.slice(-4096);
+      if (!/Traceback \(most recent call last\)/i.test(TAIL)) return false;
+      if (/Application startup complete|"GET \/v1\/[^"]+ HTTP\/[\d.]+" 2\d\d|Uvicorn running on/i.test(TAIL)) return false;
+      return true;
+    },
+    message: 'Python traceback detected — check the captured output below for the underlying error.',
+    suggestion: 'Suggested action: read the captured output for the failing step; copy the troubleshooting bundle if you need help.',
+    fixes: [],
+  },
 ];
 
 export function _diagnose(text) {
@@ -757,7 +783,7 @@ export function _showDiagnosis(panel, diagnosis, sourceText) {
       });
       row.appendChild(btn);
     }
-    body.appendChild(row);
+    diag.appendChild(row);
   }
 }
 

static/js/cookbook.js

@@ -2462,10 +2462,13 @@ export async function open(opts) {
   // returned before hydration — and since close/reopen doesn't reset the page,
   // only a full reload recovered it. Re-rendering is cheap and the in-progress
   // Running tab is rendered separately just below.
-  _renderRecipes();
+  // Guard the render passes: a single broken task card must not throw out of
+  // open() and leave the modal stuck hidden (it has no catch, so the panel
+  // would silently never appear). Show the window regardless; log and move on.
+  try { _renderRecipes(); } catch (e) { console.error('[cookbook] renderRecipes failed', e); }
   _rendered = true;
   _clearCookbookNotif();
-  _renderRunningTab();
+  try { _renderRunningTab(); } catch (e) { console.error('[cookbook] renderRunningTab failed', e); }
   // Self-heal: revive any download tasks whose tmux session is still alive
   // but were persisted as done/error (covers the "restarted server while a
   // big multi-shard download was in flight" case — the task survived in

static/js/cookbookRunning.js

@@ -784,40 +784,47 @@ function _winSessionCmd(task, tmuxArgs) {
     const ps = host
       ? `Get-Content '${sd}\\${sid}.log' -Tail ${lines} -ErrorAction SilentlyContinue`
       : `Get-Content (Join-Path $env:TEMP 'odysseus-tmux\\${sid}.log') -Tail ${lines} -ErrorAction SilentlyContinue`;
-    return host ? `ssh ${pf}${host} "powershell -Command \\"${ps}\\""` : `powershell -Command "${ps}"`;
+    return _winPowerShellCmd(task, ps);
   }
   if (tmuxArgs.includes('has-session')) {
     const ps = host
       ? `$p = Get-Content '${sd}\\${sid}.pid' -ErrorAction SilentlyContinue; if ($p) { Get-Process -Id $p -ErrorAction SilentlyContinue | Out-Null; if ($?) { exit 0 } else { exit 1 } } else { exit 1 }`
       : `$p = Get-Content (Join-Path $env:TEMP 'odysseus-tmux\\${sid}.pid') -ErrorAction SilentlyContinue; if ($p) { Get-Process -Id $p -ErrorAction SilentlyContinue | Out-Null; if ($?) { exit 0 } else { exit 1 } } else { exit 1 }`;
-    return host ? `ssh ${pf}${host} "powershell -Command \\"${ps}\\""` : `powershell -Command "${ps}"`;
+    return _winPowerShellCmd(task, ps);
   }
   if (tmuxArgs.includes('kill-session')) {
-    const stopTree = `function Stop-Tree([int]$Id) { Get-CimInstance Win32_Process -Filter "ParentProcessId = $Id" -ErrorAction SilentlyContinue | ForEach-Object { Stop-Tree ([int]$_.ProcessId) }; Stop-Process -Id $Id -Force -ErrorAction SilentlyContinue }`;
+    const ps = _winSessionStopTreePs(task);
-    const ps = host
+    return _winPowerShellCmd(task, ps);
-      ? `${stopTree}; $p = Get-Content '${sd}\\${sid}.pid' -ErrorAction SilentlyContinue; if ($p -match '^\\d+$') { Stop-Tree ([int]$p) }; Remove-Item '${sd}\\${sid}.*' -Force -ErrorAction SilentlyContinue`
-      : `${stopTree}; $p = Get-Content (Join-Path $env:TEMP 'odysseus-tmux\\${sid}.pid') -ErrorAction SilentlyContinue; if ($p -match '^\\d+$') { Stop-Tree ([int]$p) }; Remove-Item (Join-Path $env:TEMP 'odysseus-tmux\\${sid}.*') -Force -ErrorAction SilentlyContinue`;
-    return host ? `ssh ${pf}${host} "powershell -Command \\"${ps}\\""` : `powershell -Command "${ps}"`;
   }
   if (tmuxArgs.includes('send-keys') && tmuxArgs.includes('C-c')) {
     const ps = host
       ? `$p = Get-Content '${sd}\\${sid}.pid' -ErrorAction SilentlyContinue; if ($p) { Stop-Process -Id $p -ErrorAction SilentlyContinue }`
       : `$p = Get-Content (Join-Path $env:TEMP 'odysseus-tmux\\${sid}.pid') -ErrorAction SilentlyContinue; if ($p) { Stop-Process -Id $p -ErrorAction SilentlyContinue }`;
-    return host ? `ssh ${pf}${host} "powershell -Command \\"${ps}\\""` : `powershell -Command "${ps}"`;
+    return _winPowerShellCmd(task, ps);
   }
   return host ? `ssh ${pf}${host} 'tmux ${tmuxArgs}' 2>/dev/null` : `tmux ${tmuxArgs} 2>/dev/null`;
 }
 
-export function _tmuxGracefulKill(task) {
+function _winPowerShellCmd(task, ps) {
-  if (_isWindows(task)) {
+  const command = `powershell -Command "${ps}"`;
+  if (!task.remoteHost) return command;
+  return `ssh ${_sshPrefix(_getPort(task))}${task.remoteHost} ${_shQuote(command)}`;
+}
+
+function _winSessionStopTreePs(task) {
   const host = task.remoteHost;
   const sd = host ? '$env:TEMP\\odysseus-sessions' : '$env:TEMP\\odysseus-tmux';
   const sid = task.sessionId;
-    const pf = _sshPrefix(_getPort(task));
+  const stopTree = `function Stop-Tree([int]$Id) { Get-CimInstance Win32_Process -Filter ('ParentProcessId = ' + $Id) -ErrorAction SilentlyContinue | ForEach-Object { Stop-Tree ([int]$_.ProcessId) }; Stop-Process -Id $Id -Force -ErrorAction SilentlyContinue }`;
-    const ps = host
+  return host
-      ? `$p = Get-Content '${sd}\\${sid}.pid' -ErrorAction SilentlyContinue; if ($p) { Stop-Process -Id $p -Force -ErrorAction SilentlyContinue }; Remove-Item '${sd}\\${sid}.*' -Force -ErrorAction SilentlyContinue`
+    ? `${stopTree}; $p = Get-Content '${sd}\\${sid}.pid' -ErrorAction SilentlyContinue; if ($p -match '^\\d+$') { Stop-Tree ([int]$p) }; Remove-Item '${sd}\\${sid}.*' -Force -ErrorAction SilentlyContinue`
-      : `$p = Get-Content (Join-Path $env:TEMP 'odysseus-tmux\\${sid}.pid') -ErrorAction SilentlyContinue; if ($p) { Stop-Process -Id $p -Force -ErrorAction SilentlyContinue }; Remove-Item (Join-Path $env:TEMP 'odysseus-tmux\\${sid}.*') -Force -ErrorAction SilentlyContinue`;
+    : `${stopTree}; $p = Get-Content (Join-Path $env:TEMP 'odysseus-tmux\\${sid}.pid') -ErrorAction SilentlyContinue; if ($p -match '^\\d+$') { Stop-Tree ([int]$p) }; Remove-Item (Join-Path $env:TEMP 'odysseus-tmux\\${sid}.*') -Force -ErrorAction SilentlyContinue`;
-    return host ? `ssh ${pf}${host} "powershell -Command \\"${ps}\\""` : `powershell -Command "${ps}"`;
+}
+
+export function _tmuxGracefulKill(task) {
+  if (_isWindows(task)) {
+    const ps = _winSessionStopTreePs(task);
+    return _winPowerShellCmd(task, ps);
   }
   if (task.remoteHost) {
     return `ssh ${_sshPrefix(_getPort(task))}${task.remoteHost} 'tmux send-keys -t ${task.sessionId} C-c 2>/dev/null; sleep 2; tmux kill-session -t ${task.sessionId} 2>/dev/null'`;

static/js/cookbookServe.js

@@ -116,13 +116,28 @@ function _selectedServeTarget(panel) {
     : (server?.name || 'local server');
   return {
     host,
-    port: host ? (_getPort(host) || server?.port || '') : '',
+    port: host ? (server?.port || _getPort(host) || '') : '',
+    env: server?.env || '',
     venv,
     platform: server?.platform || _envState.platform || '',
     label,
   };
 }
 
+function _remoteWindowsDiffusersUnsupported(target) {
+  return !!(target?.host && target?.platform === 'windows');
+}
+
+function _backendChoicesForTarget(target) {
+  if (target?.platform === 'windows') {
+    if (_remoteWindowsDiffusersUnsupported(target)) return [['llamacpp','llama.cpp']];
+    return [['llamacpp','llama.cpp'],['diffusers','Diffusers']];
+  }
+  return _isMetal()
+    ? [['llamacpp','llama.cpp'],['ollama','Ollama']]
+    : [['vllm','vLLM'],['sglang','SGLang'],['llamacpp','llama.cpp'],['ollama','Ollama'],['diffusers','Diffusers']];
+}
+
 async function _fetchServeRuntimePackage(panel, backend) {
   const packageByBackend = {
     vllm: 'vllm',
@@ -529,13 +544,14 @@ function _rerenderCachedModels() {
       const ss = (_byRepo[repo] && typeof _byRepo[repo] === 'object')
         ? _byRepo[repo]
         : (_lastUsed || (_isLegacyFlat ? _allSs : {}));
+      const _serveTarget = _selectedServeTarget();
+      const _backendChoices = _backendChoicesForTarget(_serveTarget);
+      const _allowedBackends = new Set(_backendChoices.map(([v]) => v));
       const detectedBackend = _detectBackend(m).backend;
-      const _allowedBackends = new Set(_isWindows()
+      let defaultBackend = (ss._forceBackend && ss.backend && _allowedBackends.has(ss.backend))
-        ? ['llamacpp', 'diffusers']
-        : (_isMetal() ? ['llamacpp', 'ollama'] : ['vllm', 'sglang', 'llamacpp', 'ollama', 'diffusers']));
-      const defaultBackend = (ss._forceBackend && ss.backend && _allowedBackends.has(ss.backend))
         ? ss.backend
         : detectedBackend;
+      if (!_allowedBackends.has(defaultBackend)) defaultBackend = _backendChoices[0]?.[0] || detectedBackend;
       const savedMatchesBackend = !!ss._forceBackend || (ss.backend || 'vllm') === detectedBackend;
       const sv = (k, def) => (ss[k] !== undefined && savedMatchesBackend) ? ss[k] : def;
       const defaultTp = defaultBackend === 'llamacpp' ? '1' : sv('tp', '1');
@@ -607,12 +623,6 @@ function _rerenderCachedModels() {
       }
       // Row 1: Backend + Server + Env
       panelHtml += `<div class="hwfit-serve-row">`;
-      const _backendChoices = _isWindows()
-        ? [['llamacpp','llama.cpp'],['diffusers','Diffusers']]
-        : _isMetal()
-        // Diffusers (diffusion_server.py) is CUDA-only — omit it on Metal.
-        ? [['llamacpp','llama.cpp'],['ollama','Ollama']]
-        : [['vllm','vLLM'],['sglang','SGLang'],['llamacpp','llama.cpp'],['ollama','Ollama'],['diffusers','Diffusers']];
       const backendOpts = _backendChoices.map(([v,l]) => `<option value="${v}"${defaultBackend===v?' selected':''}>${l}</option>`).join('');
       // Custom Backend picker — native <select> can't host SVG inside
       // options, so we render a button + menu that show the backend logo
@@ -1971,6 +1981,12 @@ function _rerenderCachedModels() {
           else serveState[el.dataset.field] = el.value;
         });
         serveState.backend = serveState.backend || (_detectBackend(m).backend) || 'vllm';
+        const launchTarget = _selectedServeTarget(panel);
+        if (serveState.backend === 'diffusers' && _remoteWindowsDiffusersUnsupported(launchTarget)) {
+          _restoreLaunchBtn();
+          uiModule.showToast('Diffusers serving is not supported on remote Windows servers yet. Use local Windows or a Linux server.', 9000);
+          return;
+        }
 
         // Pre-launch: check our own task list for a serve already running
         // on this host. Offer to stop+launch as the default action — the
@@ -1979,7 +1995,7 @@ function _rerenderCachedModels() {
         // common case instantly without waiting for a network round-trip.
         try {
           const _runningMod = await import('./cookbookRunning.js');
-          const _hostStr = _envState.remoteHost || '';
+          const _hostStr = launchTarget.host || '';
           const _active = (_runningMod._loadTasks ? _runningMod._loadTasks() : []).filter(t =>
             t && t.type === 'serve'
             && (t.remoteHost || '') === _hostStr
@@ -2033,12 +2049,11 @@ function _rerenderCachedModels() {
           || (serveState.backend === 'diffusers');
         if (_needsGpu) {
           try {
-            const _probeHost = (_envState.remoteHost || '').trim();
+            const _probeHost = (launchTarget.host || '').trim();
             const _probeParams = new URLSearchParams();
             if (_probeHost) {
               _probeParams.set('host', _probeHost);
-              const _sp = (_serverByVal?.(_envState.remoteServerKey || _probeHost) || {}).port;
+              if (launchTarget.port) _probeParams.set('ssh_port', launchTarget.port);
-              if (_sp) _probeParams.set('ssh_port', _sp);
             }
             const _probeRes = await fetch('/api/cookbook/gpus' + (_probeParams.toString() ? '?' + _probeParams : ''), { credentials: 'same-origin' });
             const _probeData = await _probeRes.json();
@@ -2071,10 +2086,10 @@ function _rerenderCachedModels() {
             || launchCmd.match(/OLLAMA_HOST=[^:\s]+:(\d{2,5})\b/);
           const _port = _portMatch ? _portMatch[1] : '';
           if (_port) {
-            const _portHost = (_envState.remoteHost || '').trim();
+            const _portHost = (launchTarget.host || '').trim();
             const _checkInner = `ss -tlnp 2>/dev/null | awk '$4 ~ /:${_port}$/ {print; exit}' || netstat -tlnp 2>/dev/null | awk '$4 ~ /:${_port}$/ {print; exit}'`;
             const _cmd = _portHost
-              ? `ss h ${_portHost} <<<"" 2>/dev/null; ssh -o ConnectTimeout=4 -o StrictHostKeyChecking=no ${_portHost} ${JSON.stringify(_checkInner)}`
+              ? `ssh -o ConnectTimeout=4 -o StrictHostKeyChecking=no ${_sshPrefix(launchTarget.port)}${_portHost} ${JSON.stringify(_checkInner)}`
               : _checkInner;
             const _res = await fetch('/api/shell/exec', {
               method: 'POST', credentials: 'same-origin',
@@ -2131,20 +2146,8 @@ function _rerenderCachedModels() {
         // Resolve the target host from the visible Server dropdown — the reliable
         // source. Relying on _envState.remoteHost silently sent serves to Local
         // when that value was stale/empty. Pass it explicitly to the launcher.
-        let serveHost = _envState.remoteHost || '';
+        let serveHost = launchTarget.host || '';
-        let _srvEnv = '', _srvEnvPath = '';
+        let _srvEnv = launchTarget.env || '', _srvEnvPath = launchTarget.venv || '';
-        const _ssEl = document.getElementById('hwfit-server-select') || document.getElementById('hwfit-dl-server');
-        if (_ssEl && _ssEl.value != null) {
-          if (_ssEl.value === 'local') serveHost = '';
-          else {
-            const _srv = _serverByVal?.(_ssEl.value) || _envState.servers[parseInt(_ssEl.value)];
-            if (_srv) {
-              serveHost = _srv.host;
-              _srvEnv = _srv.env || '';
-              _srvEnvPath = _srv.envPath || '';
-            }
-          }
-        }
         // The venv field wins; otherwise fall back to the env configured for the
         // selected server in Settings, so the activation isn't silently dropped
         // when the field is left blank (the per-server venv wasn't being applied).

static/js/document.js

@@ -284,8 +284,8 @@ import * as Modals from './modalManager.js';
         ? langIcon(doc.language, 12, { style: 'opacity:0.65;flex-shrink:0;color:currentColor;margin-right:4px;' })
         : '';
       const langChip = `<span class="doc-tab-lang">${lic}</span>`;
-      html += `<div class="doc-tab${isActive ? ' active' : ''}" draggable="true" data-doc-id="${id}" title="${title}">
+      html += `<div class="doc-tab${isActive ? ' active' : ''}" draggable="true" data-doc-id="${id}" title="${_esc(title)}">
-        ${verChip}${langChip}<span class="doc-tab-title">${shortTitle}</span>
+        ${verChip}${langChip}<span class="doc-tab-title">${_esc(shortTitle)}</span>
         <button class="doc-tab-close" data-doc-id="${id}" title="Unlink from chat (kept in the Library)">&times;</button>
       </div>`;
     }

static/js/editor/canvas-coords.js

@@ -12,8 +12,8 @@ export function canvasCoords(e, canvas) {
   const rect = canvas.getBoundingClientRect();
   const scaleX = canvas.width / rect.width;
   const scaleY = canvas.height / rect.height;
-  const clientX = e.touches ? e.touches[0].clientX : e.clientX;
+  const clientX = e.touches && e.touches.length ? e.touches[0].clientX : e.clientX;
-  const clientY = e.touches ? e.touches[0].clientY : e.clientY;
+  const clientY = e.touches && e.touches.length ? e.touches[0].clientY : e.clientY;
   return {
     x: (clientX - rect.left) * scaleX,
     y: (clientY - rect.top) * scaleY,

static/js/modalManager.js

@@ -28,6 +28,7 @@
 import { previewZoneAt, clearPreview, snapModalToZone } from './tileManager.js';
 import { suspendDock, resumeDock, clearRightDock, applyEdgeDock } from './modalSnap.js';
 import { dismissOrRemove } from './escMenuStack.js';
+import { nextToolWindowZ } from './toolWindowZOrder.js';
 
 const _state = new Map(); // id -> { restoreFn, closeFn, railBtnId, isMinimized, restoreMinHeight }
 
@@ -63,7 +64,14 @@ function _applyRememberedDock(id) {
 // those statics and bump on every bring-to-front.
 let _modalTopZ = 300;
 function _bringToFront(modal) {
-  if (modal) modal.style.setProperty('z-index', String(++_modalTopZ), 'important');
+  if (!modal) return;
+  const z = nextToolWindowZ({
+    exclude: modal,
+    current: getComputedStyle(modal).zIndex,
+    floor: _modalTopZ,
+  });
+  _modalTopZ = Math.max(_modalTopZ, z);
+  modal.style.setProperty('z-index', String(z), 'important');
 }
 
 function _emitModalOpened(id, modal) {

static/js/notes.js

@@ -10,6 +10,7 @@ import { attachColorPicker } from './colorPicker.js';
 import { makeWindowDraggable } from './windowDrag.js';
 import { snapModalToZone } from './tileManager.js';
 import { applyEdgeDock, clearDockSide } from './modalSnap.js';
+import { topToolWindowZ } from './toolWindowZOrder.js';
 
 const API_BASE = window.location.origin;
 let _open = false;
@@ -200,6 +201,23 @@ function _restoreNotesSidebarDock(pane) {
   applyEdgeDock(pane, 'right');
 }
 
+// Notes is not a `.modal`; its backdrop is the top-level stacking surface.
+function _topToolWindowZ(exclude = null) {
+  return topToolWindowZ({ exclude });
+}
+
+function _bringNotesToFront(pane = document.getElementById('notes-pane')) {
+  if (!pane) return;
+  const backdrop = document.getElementById('notes-pane-backdrop') || pane.parentElement;
+  const z = _topToolWindowZ(backdrop) + 1;
+  if (backdrop) backdrop.style.setProperty('z-index', String(z), 'important');
+  try {
+    window.dispatchEvent(new CustomEvent('odysseus:modal-opened', {
+      detail: { id: 'notes-panel', modal: pane },
+    }));
+  } catch (_) {}
+}
+
 function _loadPendingHighlights() {
   try { return new Set(JSON.parse(localStorage.getItem(REMINDER_PENDING_HIGHLIGHT_KEY) || '[]')); }
   catch { return new Set(); }
@@ -1096,7 +1114,10 @@ export async function refreshDueBadge(opts = {}) {
 // ---- Panel ----
 
 export function openPanel() {
-  if (_open) return;
+  if (_open) {
+    _bringNotesToFront();
+    return;
+  }
   _open = true;
   _editingId = null;
   // Reset the search filter — the rebuilt pane's search input renders empty, so a
@@ -1192,6 +1213,7 @@ export function openPanel() {
   document.body.appendChild(backdrop);
   _wireNotesWindow(pane);
   _restoreNotesSidebarDock(pane);
+  _bringNotesToFront(pane);
 
   // Events
   // (Close chevron removed — swipe down on mobile, tool-rail toggle on desktop.)
@@ -1202,6 +1224,9 @@ export function openPanel() {
   _wireNotesSwipeDismiss(pane.querySelector('.notes-mobile-grabber'), pane);
   _wireNotesSwipeDismiss(pane.querySelector('.notes-pane-header'), pane);
 
+  pane.addEventListener('pointerdown', () => _bringNotesToFront(pane), true);
+  pane.addEventListener('focusin', () => _bringNotesToFront(pane), true);
+
   const minBtn = document.getElementById('notes-minimize-btn');
   if (minBtn) minBtn.addEventListener('click', (e) => {
     e.preventDefault();

static/js/settings.js

@@ -11,6 +11,7 @@ import { isAltGrEvent } from './platform.js';
 
 let initialized = false;
 let modalEl = null;
+let _authPolicy = { password_min_length: 8 };
 
 function el(id) { return document.getElementById(id); }
 function esc(s) { return uiModule.esc(s); }
@@ -2160,6 +2161,16 @@ function initAccount() {
       }
     }).catch(() => {});
 
+  // Update password placeholder and policy from server
+  fetch('/api/auth/policy', { credentials: 'same-origin' })
+    .then(r => r.ok ? r.json() : null)
+    .then(policy => {
+      if (!policy) return;
+      _authPolicy = policy;
+      const pwNew = el('settings-pw-new');
+      if (pwNew) pwNew.placeholder = `New password (min ${policy.password_min_length})`;
+    }).catch(() => {});
+
   // Change password
   const saveBtn = el('settings-pw-save');
   const msgEl = el('settings-pw-msg');
@@ -2170,7 +2181,7 @@ function initAccount() {
       const conf = el('settings-pw-confirm').value;
       msgEl.style.color = '';
       if (!cur || !nw) { msgEl.textContent = 'Fill in all fields'; msgEl.style.color = 'var(--red)'; return; }
-      if (nw.length < 8) { msgEl.textContent = 'Min 8 characters'; msgEl.style.color = 'var(--red)'; return; }
+      if (nw.length < _authPolicy.password_min_length) { msgEl.textContent = `Min ${_authPolicy.password_min_length} characters`; msgEl.style.color = 'var(--red)'; return; }
       if (nw !== conf) { msgEl.textContent = 'Passwords don\'t match'; msgEl.style.color = 'var(--red)'; return; }
       saveBtn.disabled = true;
       try {

static/js/toolWindowZOrder.js

@@ -0,0 +1,29 @@
+export const TOOL_WINDOW_SELECTOR = 'body > .modal, body > .research-overlay, body > .notes-pane-backdrop';
+
+export function topToolWindowZ(options = {}) {
+  const {
+    exclude = null,
+    root = globalThis.document,
+    getStyle = globalThis.getComputedStyle,
+    floor = 250,
+  } = options;
+  let top = floor;
+  if (!root || typeof root.querySelectorAll !== 'function' || typeof getStyle !== 'function') return top;
+  root.querySelectorAll(TOOL_WINDOW_SELECTOR).forEach(el => {
+    if (!el || el === exclude) return;
+    if (el.classList?.contains('hidden') || el.classList?.contains('modal-minimized')) return;
+    const cs = getStyle(el);
+    if (cs.display === 'none' || cs.visibility === 'hidden') return;
+    const z = parseInt(cs.zIndex, 10);
+    if (Number.isFinite(z)) top = Math.max(top, z);
+  });
+  return top;
+}
+
+export function nextToolWindowZ(options = {}) {
+  const { current = null } = options;
+  const top = topToolWindowZ(options);
+  const currentZ = parseInt(current, 10);
+  if (Number.isFinite(currentZ) && currentZ > top) return currentZ;
+  return top + 1;
+}

static/js/ui.js

@@ -8,6 +8,7 @@ import themeModule from './theme.js';
 import * as Modals from './modalManager.js';
 import spinnerModule from './spinner.js';
 import { registerMenuDismiss, dismissTopMenu, dismissOrRemove } from './escMenuStack.js';
+import { nextToolWindowZ, topToolWindowZ } from './toolWindowZOrder.js';
 
 let toastEl = null;
 let autoScrollEnabled = true;
@@ -1088,14 +1089,22 @@ if ('ontouchstart' in window) {
 
 // ---- Bring modal to front on click ----
 {
-  let topModalZ = 250;
+  const raiseModalToFront = (modal, floor = 250) => {
+    const z = nextToolWindowZ({
+      exclude: modal,
+      current: getComputedStyle(modal).zIndex,
+      floor,
+    });
+    modal.style.setProperty('z-index', String(z), 'important');
+    return z;
+  };
+
   document.addEventListener('mousedown', (e) => {
     const modalContent = e.target.closest('.modal-content');
     if (!modalContent) return;
     const modal = modalContent.closest('.modal');
     if (!modal) return;
-    topModalZ += 1;
+    raiseModalToFront(modal);
-    modal.style.zIndex = topModalZ;
   });
 
   // Backdrop tap to close — delegated for all modals
@@ -1190,9 +1199,15 @@ if (!window._odyEscExpandGuard) {
     // Re-entry guard: setting style.zIndex itself fires the observer that
     // calls us back. Skip if this element is already pinned to the top
     // (matches the current counter) so we don't spin into an infinite loop.
-    const cur = parseInt(m.style.zIndex, 10) || 0;
+    const cur = parseInt(getComputedStyle(m).zIndex, 10) || 0;
-    if (cur === _zCounter) return;
+    if (cur === _zCounter && cur > topToolWindowZ({ exclude: m })) return;
-    m.style.zIndex = String(++_zCounter);
+    const z = nextToolWindowZ({
+      exclude: m,
+      current: cur,
+      floor: _zCounter,
+    });
+    _zCounter = Math.max(_zCounter, z);
+    if (z !== cur) m.style.setProperty('z-index', String(z), 'important');
   };
   new MutationObserver((muts) => {
     for (const m of muts) {

static/login.html

@@ -328,6 +328,7 @@
 
   let mode = 'login'; // 'login' | 'signup' | 'setup'
   let signupAllowed = false;
+  let policy = { password_min_length: 8, reserved_usernames: [] };
 
   const rememberToggle = document.getElementById('rememberToggle');
 
@@ -360,10 +361,12 @@
     }
   }
 
-  // Check auth status
+  // Check auth status and fetch policy in parallel, but don't block the
+  // authenticated redirect on the policy response.
+  const policyPromise = fetch('/api/auth/policy', { credentials: 'same-origin' }).catch(() => null);
   try {
-    const res = await fetch('/api/auth/status', { credentials: 'same-origin' });
+    const statusRes = await fetch('/api/auth/status', { credentials: 'same-origin' });
-    const data = await res.json();
+    const data = await statusRes.json();
     if (data.authenticated) {
       window.location.replace('/');
       return;
@@ -374,6 +377,10 @@
     } else {
       setMode('login');
     }
+    const policyRes = await policyPromise;
+    if (policyRes && policyRes.ok) {
+      policy = await policyRes.json();
+    }
   } catch (e) {
     setMode('login');
   }
@@ -426,8 +433,14 @@
         submitBtn.disabled = false;
         return;
       }
-      if (password.length < 8) {
+      if (password.length < policy.password_min_length) {
-        errEl.textContent = 'Password must be at least 8 characters';
+        errEl.textContent = `Password must be at least ${policy.password_min_length} characters`;
+        errEl.style.display = 'block';
+        submitBtn.disabled = false;
+        return;
+      }
+      if (policy.reserved_usernames.includes(username.toLowerCase())) {
+        errEl.textContent = 'This username is reserved';
         errEl.style.display = 'block';
         submitBtn.disabled = false;
         return;

tests/LAYOUT_INVENTORY.md

@@ -119,7 +119,7 @@ Read-only checks, run from the repo root on this branch. Note the real API is
 ```bash
 # Compute the area_cli set and confirm test_backup_cli_security.py is
 # area_security. Expected: 28 files, then "security".
-.venv/bin/python - <<'PY'
+./venv/bin/python - <<'PY'
 from pathlib import Path
 from tests._taxonomy import classify_test_path
 
@@ -138,7 +138,7 @@ rg -n "TestClient|FastAPI|create_app|SessionLocal|sqlite|dependency_overrides" \
   tests/test_*cli*.py tests/test_sessions_cli.py
 
 # Hard-coded flat paths to the exact CLI files outside tests/. Expected: no matches.
-.venv/bin/python - <<'PY2' > /tmp/area_cli_paths.txt
+./venv/bin/python - <<'PY2' > /tmp/area_cli_paths.txt
 from pathlib import Path
 from tests._taxonomy import classify_test_path
 
@@ -158,26 +158,26 @@ tokens only (plus the `tests/helpers/` directory rule), so the markers of the
 
 ## Validation for the future move PR
 
-Run with the project venv (`.venv/bin/python`); system `python3` may miss
+Run with the project venv (`./venv/bin/python`); system `python3` may miss
 pinned deps. Before the move, record the baseline; after, compare:
 
 ```bash
 # Selection must match the 28 files before and after the move.
-.venv/bin/python tests/run_focus.py --dry-run --area cli
+./venv/bin/python tests/run_focus.py --dry-run --area cli
-.venv/bin/python -m pytest -m area_cli -q
+./venv/bin/python -m pytest -m area_cli -q
 
 # Moved files pass when targeted directly.
-.venv/bin/python -m pytest tests/cli/ -q
+./venv/bin/python -m pytest tests/cli/ -q
 
 # Whole-suite collection still succeeds (catches import/path breakage).
-.venv/bin/python -m pytest --collect-only -q
+./venv/bin/python -m pytest --collect-only -q
 
 # Taxonomy/runner infrastructure is unaffected.
-.venv/bin/python -m pytest tests/test_taxonomy.py tests/test_run_focus.py -q
+./venv/bin/python -m pytest tests/test_taxonomy.py tests/test_run_focus.py -q
 
 # No stale flat-path references to the moved files. Expected: no matches
 # outside tests/cli/ itself.
-.venv/bin/python - <<'PY2' > /tmp/area_cli_paths.txt
+./venv/bin/python - <<'PY2' > /tmp/area_cli_paths.txt
 from pathlib import Path
 from tests._taxonomy import classify_test_path
 

tests/OVERSIZED_TEST_SPLIT_PLAN.md

@@ -0,0 +1,326 @@
+# Oversized Test File Split Plan
+
+## Purpose
+
+This document plans future oversized test-file splits using current repo data.
+It does not move files, rewrite assertions, extract helpers, or change CI.
+
+## Roadmap context
+
+- Issue: #3983
+- Parent tracker: #2523
+- Follows #3973 / #3982, the report-only order-sensitivity diagnostics slice.
+
+## Methodology
+
+Metrics were generated from the current test tree using:
+
+- physical line counts for every recursive `test_*.py` file under `tests/`;
+- AST counts for `test_*` functions and `Test*` classes;
+- one `pytest --collect-only -q tests` run to count collected items per file;
+- current taxonomy classification from `tests._taxonomy.classify_test_path`; and
+- static setup-signal scans for route/API, DB/session, import-state, security, filesystem, subprocess/script, async/threading, and UI/static indicators.
+
+Static signals are not proof of risk. They are review prompts.
+Future split PRs must still inspect each file manually before editing.
+
+## Current summary
+
+- test files scanned: 583
+- collected pytest items counted: 3586
+- large-file threshold: 300 lines
+- large-collected threshold: 20 collected items
+
+Area distribution:
+
+| Value | Files |
+|---|---:|
+| cli | 28 |
+| helpers | 1 |
+| js | 39 |
+| routes | 23 |
+| security | 77 |
+| services | 144 |
+| uncategorized | 234 |
+| unit | 37 |
+
+Sub-area distribution:
+
+| Value | Files |
+|---|---:|
+| api | 6 |
+| atomic | 3 |
+| auth | 9 |
+| calendar | 10 |
+| cli | 28 |
+| confinement | 7 |
+| cookbook | 13 |
+| document | 11 |
+| email | 12 |
+| embedding | 3 |
+| gallery | 5 |
+| history | 3 |
+| js | 39 |
+| llm | 16 |
+| mcp | 8 |
+| memory | 15 |
+| nondict | 7 |
+| nonstring | 22 |
+| owner | 14 |
+| owner_scope | 23 |
+| parse | 4 |
+| provider | 6 |
+| research | 16 |
+| route | 6 |
+| routes | 9 |
+| scheduler | 3 |
+| scope | 5 |
+| security | 9 |
+| session | 16 |
+| ssrf | 3 |
+| webhook | 3 |
+| xss | 5 |
+
+Values below 2 files: 244 values covering 244 files.
+
+## Top files by collected pytest items
+
+| File | Lines | Collected tests | Test defs | Test classes | Area | Sub-area | Signals |
+|---|---:|---:|---:|---:|---|---|---|
+| `tests/test_model_routes.py` | 1778 | 139 | 116 | 10 | routes | routes | route/api, db/session, import-state, async/threading |
+| `tests/test_security_regressions.py` | 1224 | 92 | 68 | 0 | security | security | route/api, db/session, import-state, security, filesystem, async/threading, ui/static |
+| `tests/test_provider_classification.py` | 188 | 67 | 21 | 4 | services | provider | - |
+| `tests/test_cookbook_helpers.py` | 912 | 65 | 65 | 0 | services | cookbook | route/api, filesystem, subprocess/script, async/threading, ui/static |
+| `tests/test_shell_routes.py` | 481 | 63 | 48 | 8 | routes | routes | route/api, import-state, filesystem |
+| `tests/test_pr_blocker_audit.py` | 964 | 58 | 58 | 0 | uncategorized | pr_blocker_audit | import-state, security, filesystem |
+| `tests/test_provider_endpoints.py` | 241 | 58 | 18 | 1 | services | provider | subprocess/script |
+| `tests/test_agent_loop.py` | 469 | 52 | 52 | 5 | uncategorized | agent_loop | db/session, import-state |
+| `tests/test_service_health.py` | 472 | 47 | 42 | 0 | uncategorized | service_health | async/threading |
+| `tests/test_run_focus.py` | 399 | 47 | 44 | 0 | uncategorized | run_focus | security, filesystem, subprocess/script, ui/static |
+| `tests/test_llm_core_temperature.py` | 196 | 41 | 17 | 0 | services | llm | - |
+| `tests/test_endpoint_probing.py` | 411 | 34 | 30 | 6 | uncategorized | endpoint_probing | route/api, db/session, import-state |
+| `tests/test_llm_core_anthropic_temp_omit.py` | 94 | 32 | 6 | 0 | services | llm | db/session |
+| `tests/test_chat_helpers.py` | 264 | 31 | 18 | 0 | uncategorized | chat_helpers | route/api |
+| `tests/test_provider_detection.py` | 148 | 31 | 31 | 5 | services | provider | - |
+| `tests/test_model_context.py` | 251 | 30 | 30 | 4 | uncategorized | model_context | db/session, import-state |
+| `tests/test_endpoint_resolver.py` | 148 | 30 | 30 | 6 | uncategorized | endpoint_resolver | - |
+| `tests/test_embedding_lanes.py` | 1104 | 29 | 29 | 0 | services | embedding | filesystem |
+| `tests/test_upload_limits_centralized.py` | 110 | 29 | 5 | 0 | uncategorized | upload_limits_centralized | import-state, filesystem |
+| `tests/test_email_oauth.py` | 580 | 28 | 25 | 0 | services | email | route/api, db/session, security, async/threading |
+| `tests/test_review_regressions.py` | 930 | 26 | 26 | 0 | uncategorized | review_regressions | route/api, db/session, import-state, filesystem, async/threading |
+| `tests/test_rename_user_owner_sync.py` | 686 | 26 | 26 | 0 | security | owner | route/api, db/session, import-state, filesystem, async/threading |
+| `tests/test_helpers_import_state.py` | 426 | 26 | 26 | 0 | helpers | helpers | route/api, db/session, import-state |
+| `tests/test_taxonomy.py` | 145 | 26 | 16 | 0 | uncategorized | taxonomy | security, ui/static |
+| `tests/test_tool_path_confinement.py` | 282 | 24 | 24 | 0 | security | confinement | import-state, filesystem, async/threading |
+| `tests/test_copilot.py` | 170 | 23 | 16 | 0 | uncategorized | copilot | - |
+| `tests/test_research_utils.py` | 97 | 23 | 23 | 2 | services | research | - |
+| `tests/test_api_chat_security.py` | 401 | 22 | 8 | 0 | security | security | route/api, db/session, import-state, filesystem, async/threading |
+| `tests/test_tool_support_heuristic.py` | 166 | 22 | 22 | 3 | uncategorized | tool_support_heuristic | - |
+| `tests/test_platform_compat.py` | 318 | 21 | 21 | 0 | uncategorized | platform_compat | import-state, filesystem, subprocess/script |
+
+## Top files by physical line count
+
+| File | Lines | Collected tests | Test defs | Test classes | Area | Sub-area | Signals |
+|---|---:|---:|---:|---:|---|---|---|
+| `tests/test_model_routes.py` | 1778 | 139 | 116 | 10 | routes | routes | route/api, db/session, import-state, async/threading |
+| `tests/test_security_regressions.py` | 1224 | 92 | 68 | 0 | security | security | route/api, db/session, import-state, security, filesystem, async/threading, ui/static |
+| `tests/test_embedding_lanes.py` | 1104 | 29 | 29 | 0 | services | embedding | filesystem |
+| `tests/test_pr_blocker_audit.py` | 964 | 58 | 58 | 0 | uncategorized | pr_blocker_audit | import-state, security, filesystem |
+| `tests/test_review_regressions.py` | 930 | 26 | 26 | 0 | uncategorized | review_regressions | route/api, db/session, import-state, filesystem, async/threading |
+| `tests/test_cookbook_helpers.py` | 912 | 65 | 65 | 0 | services | cookbook | route/api, filesystem, subprocess/script, async/threading, ui/static |
+| `tests/test_rename_user_owner_sync.py` | 686 | 26 | 26 | 0 | security | owner | route/api, db/session, import-state, filesystem, async/threading |
+| `tests/test_email_oauth.py` | 580 | 28 | 25 | 0 | services | email | route/api, db/session, security, async/threading |
+| `tests/test_api_token_routes.py` | 578 | 17 | 17 | 0 | routes | api_routes | route/api, db/session, import-state, async/threading |
+| `tests/test_shell_routes.py` | 481 | 63 | 48 | 8 | routes | routes | route/api, import-state, filesystem |
+| `tests/test_email_owner_scope.py` | 474 | 9 | 9 | 0 | security | owner_scope | route/api, db/session, filesystem, async/threading |
+| `tests/test_service_health.py` | 472 | 47 | 42 | 0 | uncategorized | service_health | async/threading |
+| `tests/test_agent_loop.py` | 469 | 52 | 52 | 5 | uncategorized | agent_loop | db/session, import-state |
+| `tests/test_kv_cache_invalidation_2927.py` | 463 | 8 | 8 | 0 | uncategorized | kv_cache_invalidation_2927 | route/api, db/session, import-state, async/threading |
+| `tests/test_helpers_import_state.py` | 426 | 26 | 26 | 0 | helpers | helpers | route/api, db/session, import-state |
+| `tests/test_endpoint_owner_scope_followup.py` | 414 | 11 | 11 | 0 | security | owner_scope | route/api, db/session, filesystem |
+| `tests/test_endpoint_probing.py` | 411 | 34 | 30 | 6 | uncategorized | endpoint_probing | route/api, db/session, import-state |
+| `tests/test_imap_leak_fixes.py` | 404 | 15 | 15 | 0 | uncategorized | imap_leak_fixes | route/api, db/session, security, filesystem |
+| `tests/test_companion_readonly.py` | 402 | 17 | 17 | 0 | uncategorized | companion_readonly | db/session, import-state |
+| `tests/test_api_chat_security.py` | 401 | 22 | 8 | 0 | security | security | route/api, db/session, import-state, filesystem, async/threading |
+| `tests/test_upload_handler_atomicity.py` | 401 | 9 | 9 | 0 | uncategorized | upload_handler_atomicity | filesystem, async/threading |
+| `tests/test_run_focus.py` | 399 | 47 | 44 | 0 | uncategorized | run_focus | security, filesystem, subprocess/script, ui/static |
+| `tests/test_auth_regressions.py` | 375 | 15 | 15 | 0 | security | auth | route/api, db/session, import-state, async/threading |
+| `tests/test_calendar_owner_scope.py` | 345 | 7 | 7 | 0 | security | owner_scope | route/api, db/session, import-state, filesystem, async/threading, ui/static |
+| `tests/test_null_owner_gates.py` | 342 | 20 | 20 | 0 | security | owner | route/api, db/session, import-state |
+| `tests/test_agent_migration_manifest.py` | 340 | 15 | 15 | 0 | uncategorized | agent_migration_manifest | import-state, filesystem |
+| `tests/test_calendar_recurrence.py` | 338 | 19 | 19 | 0 | services | calendar | - |
+| `tests/test_tool_policy.py` | 330 | 13 | 13 | 0 | uncategorized | tool_policy | import-state, async/threading |
+| `tests/test_workspace_confine.py` | 328 | 18 | 18 | 0 | uncategorized | workspace_confine | route/api, filesystem, subprocess/script, async/threading |
+| `tests/test_diffusion_server_security.py` | 325 | 14 | 14 | 0 | security | security | route/api, import-state, security, filesystem, async/threading, ui/static |
+
+## Split planning candidates
+
+This section is generated from metrics, not from manual judgement.
+Files are included when they meet at least one threshold:
+
+- at least 300 physical lines; or
+- at least 20 collected pytest items.
+
+These are planning candidates only. A later split PR still needs a focused manual review of each file before moving tests.
+
+| File | Why included | Setup/risk signals | Suggested handling |
+|---|---|---|---|
+| `tests/test_model_routes.py` | 1778 lines, 139 collected tests | route/api, db/session, import-state, async/threading | Defer mechanical split until setup/risk boundaries are mapped. |
+| `tests/test_security_regressions.py` | 1224 lines, 92 collected tests | route/api, db/session, import-state, security, filesystem, async/threading, ui/static | Defer mechanical split until setup/risk boundaries are mapped. |
+| `tests/test_provider_classification.py` | 67 collected tests | No obvious setup signals from static scan. | Good first manual-review candidate if test themes are cohesive. |
+| `tests/test_cookbook_helpers.py` | 912 lines, 65 collected tests | route/api, filesystem, subprocess/script, async/threading, ui/static | Defer mechanical split until setup/risk boundaries are mapped. |
+| `tests/test_shell_routes.py` | 481 lines, 63 collected tests | route/api, import-state, filesystem | Defer mechanical split until setup/risk boundaries are mapped. |
+| `tests/test_pr_blocker_audit.py` | 964 lines, 58 collected tests | import-state, security, filesystem | Defer mechanical split until setup/risk boundaries are mapped. |
+| `tests/test_provider_endpoints.py` | 58 collected tests | subprocess/script | Good first manual-review candidate if test themes are cohesive. |
+| `tests/test_agent_loop.py` | 469 lines, 52 collected tests | db/session, import-state | Defer mechanical split until setup/risk boundaries are mapped. |
+| `tests/test_service_health.py` | 472 lines, 47 collected tests | async/threading | Good first manual-review candidate if test themes are cohesive. |
+| `tests/test_run_focus.py` | 399 lines, 47 collected tests | security, filesystem, subprocess/script, ui/static | Defer mechanical split until setup/risk boundaries are mapped. |
+| `tests/test_llm_core_temperature.py` | 41 collected tests | No obvious setup signals from static scan. | Good first manual-review candidate if test themes are cohesive. |
+| `tests/test_endpoint_probing.py` | 411 lines, 34 collected tests | route/api, db/session, import-state | Defer mechanical split until setup/risk boundaries are mapped. |
+| `tests/test_llm_core_anthropic_temp_omit.py` | 32 collected tests | db/session | Defer mechanical split until setup/risk boundaries are mapped. |
+| `tests/test_chat_helpers.py` | 31 collected tests | route/api | Defer mechanical split until setup/risk boundaries are mapped. |
+| `tests/test_provider_detection.py` | 31 collected tests | No obvious setup signals from static scan. | Good first manual-review candidate if test themes are cohesive. |
+| `tests/test_model_context.py` | 30 collected tests | db/session, import-state | Defer mechanical split until setup/risk boundaries are mapped. |
+| `tests/test_endpoint_resolver.py` | 30 collected tests | No obvious setup signals from static scan. | Good first manual-review candidate if test themes are cohesive. |
+| `tests/test_embedding_lanes.py` | 1104 lines, 29 collected tests | filesystem | Good first manual-review candidate if test themes are cohesive. |
+| `tests/test_upload_limits_centralized.py` | 29 collected tests | import-state, filesystem | Defer mechanical split until setup/risk boundaries are mapped. |
+| `tests/test_email_oauth.py` | 580 lines, 28 collected tests | route/api, db/session, security, async/threading | Defer mechanical split until setup/risk boundaries are mapped. |
+| `tests/test_review_regressions.py` | 930 lines, 26 collected tests | route/api, db/session, import-state, filesystem, async/threading | Defer mechanical split until setup/risk boundaries are mapped. |
+| `tests/test_rename_user_owner_sync.py` | 686 lines, 26 collected tests | route/api, db/session, import-state, filesystem, async/threading | Defer mechanical split until setup/risk boundaries are mapped. |
+| `tests/test_helpers_import_state.py` | 426 lines, 26 collected tests | route/api, db/session, import-state | Defer mechanical split until setup/risk boundaries are mapped. |
+| `tests/test_taxonomy.py` | 26 collected tests | security, ui/static | Defer mechanical split until setup/risk boundaries are mapped. |
+| `tests/test_tool_path_confinement.py` | 24 collected tests | import-state, filesystem, async/threading | Defer mechanical split until setup/risk boundaries are mapped. |
+| `tests/test_copilot.py` | 23 collected tests | No obvious setup signals from static scan. | Good first manual-review candidate if test themes are cohesive. |
+| `tests/test_research_utils.py` | 23 collected tests | No obvious setup signals from static scan. | Good first manual-review candidate if test themes are cohesive. |
+| `tests/test_api_chat_security.py` | 401 lines, 22 collected tests | route/api, db/session, import-state, filesystem, async/threading | Defer mechanical split until setup/risk boundaries are mapped. |
+| `tests/test_tool_support_heuristic.py` | 22 collected tests | No obvious setup signals from static scan. | Good first manual-review candidate if test themes are cohesive. |
+| `tests/test_platform_compat.py` | 318 lines, 21 collected tests | import-state, filesystem, subprocess/script | Defer mechanical split until setup/risk boundaries are mapped. |
+| `tests/test_context_compactor.py` | 21 collected tests | db/session, import-state, async/threading | Defer mechanical split until setup/risk boundaries are mapped. |
+| `tests/test_prompt_security.py` | 21 collected tests | No obvious setup signals from static scan. | Good first manual-review candidate if test themes are cohesive. |
+| `tests/test_null_owner_gates.py` | 342 lines, 20 collected tests | route/api, db/session, import-state | Defer mechanical split until setup/risk boundaries are mapped. |
+| `tests/test_youtube_handler_consolidation.py` | 20 collected tests | route/api, import-state | Defer mechanical split until setup/risk boundaries are mapped. |
+| `tests/test_calendar_recurrence.py` | 338 lines | No obvious setup signals from static scan. | Plan split boundaries before editing. |
+| `tests/test_workspace_confine.py` | 328 lines | route/api, filesystem, subprocess/script, async/threading | Defer mechanical split until setup/risk boundaries are mapped. |
+| `tests/test_api_token_routes.py` | 578 lines | route/api, db/session, import-state, async/threading | Defer mechanical split until setup/risk boundaries are mapped. |
+| `tests/test_companion_readonly.py` | 402 lines | db/session, import-state | Defer mechanical split until setup/risk boundaries are mapped. |
+| `tests/test_set_admin.py` | 317 lines | route/api, import-state, filesystem, async/threading | Defer mechanical split until setup/risk boundaries are mapped. |
+| `tests/test_imap_leak_fixes.py` | 404 lines | route/api, db/session, security, filesystem | Defer mechanical split until setup/risk boundaries are mapped. |
+| `tests/test_auth_regressions.py` | 375 lines | route/api, db/session, import-state, async/threading | Defer mechanical split until setup/risk boundaries are mapped. |
+| `tests/test_agent_migration_manifest.py` | 340 lines | import-state, filesystem | Defer mechanical split until setup/risk boundaries are mapped. |
+| `tests/test_diffusion_server_security.py` | 325 lines | route/api, import-state, security, filesystem, async/threading, ui/static | Defer mechanical split until setup/risk boundaries are mapped. |
+| `tests/test_tool_policy.py` | 330 lines | import-state, async/threading | Defer mechanical split until setup/risk boundaries are mapped. |
+| `tests/test_endpoint_owner_scope_followup.py` | 414 lines | route/api, db/session, filesystem | Defer mechanical split until setup/risk boundaries are mapped. |
+| `tests/test_upload_routes_owner_scope.py` | 315 lines | route/api, filesystem, async/threading | Defer mechanical split until setup/risk boundaries are mapped. |
+| `tests/test_email_owner_scope.py` | 474 lines | route/api, db/session, filesystem, async/threading | Defer mechanical split until setup/risk boundaries are mapped. |
+| `tests/test_upload_handler_atomicity.py` | 401 lines | filesystem, async/threading | Plan split boundaries before editing. |
+| `tests/test_kv_cache_invalidation_2927.py` | 463 lines | route/api, db/session, import-state, async/threading | Defer mechanical split until setup/risk boundaries are mapped. |
+| `tests/test_calendar_owner_scope.py` | 345 lines | route/api, db/session, import-state, filesystem, async/threading, ui/static | Defer mechanical split until setup/risk boundaries are mapped. |
+| `tests/test_skills_manager_owner_isolation.py` | 306 lines | import-state, filesystem | Defer mechanical split until setup/risk boundaries are mapped. |
+
+## Taxonomy coverage gaps among split candidates
+
+`uncategorized` is a current taxonomy area, not a builder failure.
+This plan does not reclassify tests because taxonomy changes should be reviewed separately from oversized-file split planning.
+
+Before using any of these files as a split target, first decide whether the taxonomy should be refined in a separate focused issue/PR.
+
+| File | Lines | Collected tests | Sub-area | Signals | Suggested follow-up |
+|---|---:|---:|---|---|---|
+| `tests/test_pr_blocker_audit.py` | 964 | 58 | pr_blocker_audit | import-state, security, filesystem | Review taxonomy and setup/risk boundaries before any split. |
+| `tests/test_agent_loop.py` | 469 | 52 | agent_loop | db/session, import-state | Review taxonomy and setup/risk boundaries before any split. |
+| `tests/test_service_health.py` | 472 | 47 | service_health | async/threading | Review taxonomy mapping before using as a split target. |
+| `tests/test_run_focus.py` | 399 | 47 | run_focus | security, filesystem, subprocess/script, ui/static | Review taxonomy and setup/risk boundaries before any split. |
+| `tests/test_endpoint_probing.py` | 411 | 34 | endpoint_probing | route/api, db/session, import-state | Review taxonomy and setup/risk boundaries before any split. |
+| `tests/test_chat_helpers.py` | 264 | 31 | chat_helpers | route/api | Review taxonomy and setup/risk boundaries before any split. |
+| `tests/test_model_context.py` | 251 | 30 | model_context | db/session, import-state | Review taxonomy and setup/risk boundaries before any split. |
+| `tests/test_endpoint_resolver.py` | 148 | 30 | endpoint_resolver | - | Review taxonomy mapping before using as a split target. |
+| `tests/test_upload_limits_centralized.py` | 110 | 29 | upload_limits_centralized | import-state, filesystem | Review taxonomy and setup/risk boundaries before any split. |
+| `tests/test_review_regressions.py` | 930 | 26 | review_regressions | route/api, db/session, import-state, filesystem, async/threading | Review taxonomy and setup/risk boundaries before any split. |
+| `tests/test_taxonomy.py` | 145 | 26 | taxonomy | security, ui/static | Review taxonomy and setup/risk boundaries before any split. |
+| `tests/test_copilot.py` | 170 | 23 | copilot | - | Review taxonomy mapping before using as a split target. |
+| `tests/test_tool_support_heuristic.py` | 166 | 22 | tool_support_heuristic | - | Review taxonomy mapping before using as a split target. |
+| `tests/test_platform_compat.py` | 318 | 21 | platform_compat | import-state, filesystem, subprocess/script | Review taxonomy and setup/risk boundaries before any split. |
+| `tests/test_context_compactor.py` | 233 | 21 | context_compactor | db/session, import-state, async/threading | Review taxonomy and setup/risk boundaries before any split. |
+| `tests/test_youtube_handler_consolidation.py` | 104 | 20 | youtube_handler_consolidation | route/api, import-state | Review taxonomy and setup/risk boundaries before any split. |
+| `tests/test_workspace_confine.py` | 328 | 18 | workspace_confine | route/api, filesystem, subprocess/script, async/threading | Review taxonomy and setup/risk boundaries before any split. |
+| `tests/test_companion_readonly.py` | 402 | 17 | companion_readonly | db/session, import-state | Review taxonomy and setup/risk boundaries before any split. |
+| `tests/test_set_admin.py` | 317 | 17 | set_admin | route/api, import-state, filesystem, async/threading | Review taxonomy and setup/risk boundaries before any split. |
+| `tests/test_imap_leak_fixes.py` | 404 | 15 | imap_leak_fixes | route/api, db/session, security, filesystem | Review taxonomy and setup/risk boundaries before any split. |
+| `tests/test_agent_migration_manifest.py` | 340 | 15 | agent_migration_manifest | import-state, filesystem | Review taxonomy and setup/risk boundaries before any split. |
+| `tests/test_tool_policy.py` | 330 | 13 | tool_policy | import-state, async/threading | Review taxonomy and setup/risk boundaries before any split. |
+| `tests/test_upload_handler_atomicity.py` | 401 | 9 | upload_handler_atomicity | filesystem, async/threading | Review taxonomy mapping before using as a split target. |
+| `tests/test_kv_cache_invalidation_2927.py` | 463 | 8 | kv_cache_invalidation_2927 | route/api, db/session, import-state, async/threading | Review taxonomy and setup/risk boundaries before any split. |
+
+## Suggested first manual-review candidates
+
+These are not automatic split approvals. They are categorized candidates with enough size/collection value and no route/API, DB/session, import-state, or security signal from the static scan.
+
+Files still in the `uncategorized` taxonomy area are listed separately below so taxonomy review does not get mixed into the first split decision.
+
+| File | Lines | Collected tests | Area | Sub-area | Signals | Why this is a candidate |
+|---|---:|---:|---|---|---|---|
+| `tests/test_provider_classification.py` | 188 | 67 | services | provider | - | 67 collected tests |
+| `tests/test_provider_endpoints.py` | 241 | 58 | services | provider | subprocess/script | 58 collected tests |
+| `tests/test_llm_core_temperature.py` | 196 | 41 | services | llm | - | 41 collected tests |
+| `tests/test_provider_detection.py` | 148 | 31 | services | provider | - | 31 collected tests |
+| `tests/test_embedding_lanes.py` | 1104 | 29 | services | embedding | filesystem | 1104 lines, 29 collected tests |
+| `tests/test_research_utils.py` | 97 | 23 | services | research | - | 23 collected tests |
+| `tests/test_prompt_security.py` | 203 | 21 | security | security | - | 21 collected tests |
+| `tests/test_calendar_recurrence.py` | 338 | 19 | services | calendar | - | 338 lines |
+
+## High-risk candidates to defer first
+
+These files may still be split later, but not as the first implementation slice without a separate manual boundary review.
+
+| File | Lines | Collected tests | High-risk signals |
+|---|---:|---:|---|
+| `tests/test_model_routes.py` | 1778 | 139 | db/session, import-state, route/api |
+| `tests/test_security_regressions.py` | 1224 | 92 | db/session, import-state, route/api, security |
+| `tests/test_cookbook_helpers.py` | 912 | 65 | route/api |
+| `tests/test_shell_routes.py` | 481 | 63 | import-state, route/api |
+| `tests/test_pr_blocker_audit.py` | 964 | 58 | import-state, security |
+| `tests/test_agent_loop.py` | 469 | 52 | db/session, import-state |
+| `tests/test_run_focus.py` | 399 | 47 | security |
+| `tests/test_endpoint_probing.py` | 411 | 34 | db/session, import-state, route/api |
+| `tests/test_llm_core_anthropic_temp_omit.py` | 94 | 32 | db/session |
+| `tests/test_chat_helpers.py` | 264 | 31 | route/api |
+| `tests/test_model_context.py` | 251 | 30 | db/session, import-state |
+| `tests/test_upload_limits_centralized.py` | 110 | 29 | import-state |
+| `tests/test_email_oauth.py` | 580 | 28 | db/session, route/api, security |
+| `tests/test_review_regressions.py` | 930 | 26 | db/session, import-state, route/api |
+| `tests/test_rename_user_owner_sync.py` | 686 | 26 | db/session, import-state, route/api |
+
+## Rules for future split PRs
+
+- One file or one coherent file-family per PR.
+- No assertion rewrites mixed with file moves.
+- No helper extraction mixed with file moves.
+- No production code changes.
+- No CI workflow changes.
+- Preserve existing markers and taxonomy unless the split issue explicitly says otherwise.
+- Validate the original file's collected tests before and after the split.
+- Validate any neighboring taxonomy/focused-runner behavior if paths change.
+- Treat files with route/API, DB/session, import-state, or security signals as higher-risk until manually reviewed.
+
+## Suggested next step
+
+Use this plan to choose the first actual oversized-file split issue.
+The first split should prefer a file with high review value and low setup risk.
+Do not start a split PR from this planning issue alone if the file's boundaries are still ambiguous.
+
+## Reproduction command
+
+This document was generated with:
+
+```bash
+.venv/bin/python tests/tools/build_oversized_test_split_plan.py
+```
+
+## Freshness check
+
+After editing the builder or rebasing the branch, regenerate the plan and confirm no unexpected plan drift:
+
+```bash
+.venv/bin/python tests/tools/build_oversized_test_split_plan.py
+git diff --exit-code -- tests/OVERSIZED_TEST_SPLIT_PLAN.md
+```

tests/README.md

@@ -22,8 +22,8 @@ markers only - it moves no files and changes no test behavior. Use them to run a
 focused slice:
 
 ```bash
-python3 -m pytest -m area_security
+./venv/bin/python -m pytest -m area_security
-python3 -m pytest -m "area_services and sub_cookbook"
+./venv/bin/python -m pytest -m "area_services and sub_cookbook"
 ```
 
 Areas are `security`, `routes`, `services`, `cli`, `js`, `helpers`, `unit`, and
@@ -38,13 +38,13 @@ sub-area names, accepts sub-areas with or without the `sub_` prefix, and passes
 extra pytest arguments after `--`:
 
 ```bash
-python3 tests/run_focus.py --area security
+./venv/bin/python tests/run_focus.py --area security
-python3 tests/run_focus.py --area services --sub-area cookbook
+./venv/bin/python tests/run_focus.py --area services --sub-area cookbook
-python3 tests/run_focus.py --sub-area sub_cookbook
+./venv/bin/python tests/run_focus.py --sub-area sub_cookbook
-python3 tests/run_focus.py --keyword taxonomy
+./venv/bin/python tests/run_focus.py --keyword taxonomy
-python3 tests/run_focus.py --last-failed
+./venv/bin/python tests/run_focus.py --last-failed
-python3 tests/run_focus.py --dry-run --area services --sub-area cookbook
+./venv/bin/python tests/run_focus.py --dry-run --area services --sub-area cookbook
-python3 tests/run_focus.py --area services -- --maxfail=1 -q
+./venv/bin/python tests/run_focus.py --area services -- --maxfail=1 -q
 ```
 
 ### Fast lane and duration visibility
@@ -61,15 +61,15 @@ so you can see where time goes. They are reporting only and do not count as a
 focus selector, so `--durations` must be combined with a real selector
 (`--area`, `--sub-area`, `--keyword`, `--last-failed`, or `--fast`).
 
-Activate or otherwise use the project Python environment before running these
+Use the project Python environment before running these commands. The examples
-commands. The examples use `python3` intentionally to avoid hard-coding a local
+use the repo's documented `./venv/bin/python` path so they do not accidentally
-venv path.
+fall back to system Python.
 
 ```bash
-python3 tests/run_focus.py --fast
+./venv/bin/python tests/run_focus.py --fast
-python3 tests/run_focus.py --area services --fast
+./venv/bin/python tests/run_focus.py --area services --fast
-python3 tests/run_focus.py --area services --durations 25
+./venv/bin/python tests/run_focus.py --area services --durations 25
-python3 tests/run_focus.py --area services --fast --durations 25 --durations-min 0.05
+./venv/bin/python tests/run_focus.py --area services --fast --durations 25 --durations-min 0.05
 ```
 
 The `slow` marker is opt-in. Mark a test `slow` only with duration evidence
@@ -79,8 +79,8 @@ replace the full suite before merge. A `slow` mark only excludes a test from the
 fast lane; the test stays runnable directly, e.g.:
 
 ```bash
-python3 -m pytest tests/test_auth_config_lock_concurrency.py
+./venv/bin/python -m pytest tests/test_auth_config_lock_concurrency.py
-python3 -m pytest -m slow
+./venv/bin/python -m pytest -m slow
 ```
 
 ## Order-sensitivity reporting (report-only)
@@ -93,8 +93,8 @@ ordering - the shuffle exists only inside this runner. The seed is always
 printed, and pytest targets/options go after a literal `--`:
 
 ```bash
-python3 tests/run_order_report.py --seed 123 -- tests/cli/ -q
+./venv/bin/python tests/run_order_report.py --seed 123 -- tests/cli/ -q
-python3 tests/run_order_report.py -- tests/cli/ -q   # generates and prints a seed
+./venv/bin/python tests/run_order_report.py -- tests/cli/ -q   # generates and prints a seed
 ```
 
 The same seed reproduces the same order when the reported working directory,
@@ -108,7 +108,7 @@ A generated-seed run starts with output like:
 [order-report] working directory: /path/to/odysseus
 [order-report] shuffling test order with seed 284734921
 [order-report] reproduce from this working directory with the same test environment:
-[order-report] reproduce with: /path/to/odysseus/.venv/bin/python /path/to/odysseus/tests/run_order_report.py --seed 284734921 -- tests/cli/ -q
+[order-report] reproduce with: /path/to/odysseus/venv/bin/python /path/to/odysseus/tests/run_order_report.py --seed 284734921 -- tests/cli/ -q
 ```
 
 Run the printed command from the reported working directory to reproduce the
@@ -118,7 +118,7 @@ same fixed-seed order:
 [order-report] working directory: /path/to/odysseus
 [order-report] shuffling test order with seed 284734921
 [order-report] reproduce from this working directory with the same test environment:
-[order-report] reproduce with: /path/to/odysseus/.venv/bin/python /path/to/odysseus/tests/run_order_report.py --seed 284734921 -- tests/cli/ -q
+[order-report] reproduce with: /path/to/odysseus/venv/bin/python /path/to/odysseus/tests/run_order_report.py --seed 284734921 -- tests/cli/ -q
 ```
 
 Pytest output remains visible between the report header and footer. A failing
@@ -237,10 +237,10 @@ helpers:
 Run validation locally before opening or approving a PR. Practical checks:
 
 - `git diff --check` - catch whitespace and conflict-marker errors.
-- `python3 -m py_compile <changed files>` - confirm changed files compile.
+- `./venv/bin/python -m py_compile <changed files>` - confirm changed files compile.
-- Focused `pytest` on the changed test files.
+- Focused `./venv/bin/python -m pytest` on the changed test files.
-- `pytest` on neighboring or order-sensitive test groups that share import
+- `./venv/bin/python -m pytest` on neighboring or order-sensitive test groups
-  state with the changed files.
+  that share import state with the changed files.
 - `grep` for the old boilerplate when replacing it, to confirm no stragglers
   remain.
 - A fresh audit worktree when changing the helpers themselves, so stale

tests/TESTING_STANDARD.md

@@ -24,7 +24,7 @@ The goal is not only to reorganize `tests/`. The goal is for the suite to be a
 reliable foundation for future development: deterministic, modular, informative,
 behavior-focused, and complete enough to replace manual QA wherever practical.
 
-Run tests with the project virtualenv interpreter (`.venv/bin/python -m pytest`).
+Run tests with the project virtualenv interpreter (`./venv/bin/python -m pytest`).
 The system `python3` may be missing pinned dependencies (e.g. `nh3`), which
 shows up as import/collection errors that are environmental, not real failures.
 
@@ -172,10 +172,10 @@ Prefer tests that exercise real behavior over tests that inspect source code.
 Run locally before opening or approving a refactor PR:
 
 - `git diff --check` - whitespace and conflict-marker errors.
-- `python3 -m py_compile <changed .py files>` - changed files compile.
+- `./venv/bin/python -m py_compile <changed .py files>` - changed files compile.
-- Focused `pytest` on the changed files (use `.venv/bin/python -m pytest`).
+- Focused `./venv/bin/python -m pytest` on the changed files.
-- `pytest` on neighboring / order-sensitive groups that share import state with
+- `./venv/bin/python -m pytest` on neighboring / order-sensitive groups that
-  the changed files.
+  share import state with the changed files.
 - When replacing boilerplate, `grep` for the old pattern to confirm no stragglers.
 - When changing a helper itself, validate in a fresh worktree so stale
   `__pycache__` or import state cannot mask a regression.

tests/test_ai_interaction_owner_scope.py

@@ -3,6 +3,7 @@ import inspect
 import pytest
 
 from src import ai_interaction
+from src.agent_tools import model_interaction_tools
 
 
 def _source(fn) -> str:
@@ -18,7 +19,8 @@ def test_model_resolver_applies_owner_filter():
 
 
 def test_model_listing_and_image_fallback_are_owner_scoped():
-    list_body = _source(ai_interaction.do_list_models)
+    # list_models moved to agent_tools.model_interaction_tools (#3629).
+    list_body = _source(model_interaction_tools.list_models)
     image_body = _source(ai_interaction.do_generate_image)
 
     assert "owner: Optional[str] = None" in list_body
@@ -28,12 +30,13 @@ def test_model_listing_and_image_fallback_are_owner_scoped():
     assert "_resolve_model(model_spec, owner=owner)" in image_body
 
 
+# chat_with_model, list_models and ask_teacher moved to the registry (#3629)
+# and no longer route through dispatch_ai_tool; their owner threading is covered
+# by tests/test_model_interaction_registry.py. The remaining model-ish tools
+# still dispatched here:
 @pytest.mark.parametrize("tool,content", [
-    ("chat_with_model", "gpt-test\nhello"),
     ("pipeline", "gpt-test | summarize this"),
-    ("list_models", ""),
     ("ui_control", "switch_model gpt-test"),
-    ("ask_teacher", "gpt-test\nhelp me"),
 ])
 async def test_dispatch_passes_owner_to_model_tools(monkeypatch, tool, content):
     seen = {}
@@ -42,31 +45,16 @@ async def test_dispatch_passes_owner_to_model_tools(monkeypatch, tool, content):
         seen[name] = {"content": content, "session_id": session_id, "owner": owner}
         return {"ok": True}
 
-    monkeypatch.setattr(
-        ai_interaction,
-        "do_chat_with_model",
-        lambda content, session_id=None, owner=None: capture("chat_with_model", content, session_id, owner),
-    )
     monkeypatch.setattr(
         ai_interaction,
         "do_pipeline",
         lambda content, session_id=None, owner=None: capture("pipeline", content, session_id, owner),
     )
-    monkeypatch.setattr(
-        ai_interaction,
-        "do_list_models",
-        lambda content, session_id=None, owner=None: capture("list_models", content, session_id, owner),
-    )
     monkeypatch.setattr(
         ai_interaction,
         "do_ui_control",
         lambda content, session_id=None, owner=None: capture("ui_control", content, session_id, owner),
     )
-    monkeypatch.setattr(
-        ai_interaction,
-        "do_ask_teacher",
-        lambda content, session_id=None, owner=None: capture("ask_teacher", content, session_id, owner),
-    )
 
     _desc, result = await ai_interaction.dispatch_ai_tool(tool, content, session_id="sid1", owner="alice")
 

tests/test_api_call_integration_routing.py

@@ -0,0 +1,78 @@
+"""Regression: api_call reaches the model for API-integration intent (#3794).
+
+The repro prompt — "Use the api_call tool to call Home Assistant GET
+/api/states" — matched no domain in ``_classify_agent_request``, so it was
+treated as low-signal. The agent loop then skipped retrieval and the function
+schema filter sent only the always-available tools (manage_memory / ask_user /
+update_plan); ``api_call`` was never advertised to the model even though the
+ToolIndex description existed. Adding the registry description alone did not fix
+runtime selection.
+
+These tests drive the real path the agent uses — classifier -> domain tool map
+(relevant tools) -> FUNCTION_TOOL_SCHEMAS filter — using the actual functions and
+constants, so they would fail on the pre-fix code (empty domains -> low-signal ->
+no api_call). They skip locally when the agent's heavy deps (httpx/embeddings)
+are absent, and run in CI where they are installed.
+"""
+import pytest
+
+agent_loop = pytest.importorskip("src.agent_loop")
+
+REPRO = "Use the api_call tool to call Home Assistant GET /api/states"
+
+
+def _selected_tools(domains):
+    """Mirror agent_loop's deterministic domain seeding (see the loop over
+    `_intent['domains']` that updates `_relevant_tools` from `_DOMAIN_TOOL_MAP`)."""
+    tools = set()
+    for domain in domains:
+        tools |= agent_loop._DOMAIN_TOOL_MAP.get(domain, set())
+    return tools
+
+
+def _schema_names_sent(tools):
+    """Mirror the api-model schema filter that keeps only selected tools."""
+    return {
+        s.get("function", {}).get("name")
+        for s in agent_loop.FUNCTION_TOOL_SCHEMAS
+        if s.get("function", {}).get("name") in tools
+    }
+
+
+@pytest.mark.parametrize(
+    "prompt",
+    [
+        REPRO,
+        "check my home assistant lights",
+        "fetch the latest unread from miniflux via the api_call tool",
+        "call my gitea integration to list repos",
+    ],
+)
+def test_integration_prompts_are_not_low_signal(prompt):
+    intent = agent_loop._classify_agent_request([], prompt)
+    assert intent["low_signal"] is False, intent
+    assert "integrations" in intent["domains"], intent
+
+
+def test_repro_selects_and_sends_api_call_schema():
+    intent = agent_loop._classify_agent_request([], REPRO)
+    selected = _selected_tools(intent["domains"])
+    assert "api_call" in selected, selected
+    # The schema filter must actually advertise api_call to the model.
+    assert "api_call" in _schema_names_sent(selected), "api_call schema must reach the model"
+
+
+def test_integrations_domain_has_a_rule_pack():
+    # _domain_rules_for_tools indexes _DOMAIN_RULES[domain] directly, so a domain
+    # present in _DOMAIN_TOOL_MAP without a _DOMAIN_RULES entry would KeyError the
+    # moment api_call is selected.
+    rules = agent_loop._domain_rules_for_tools({"api_call"})
+    assert any("api_call" in r for r in rules), rules
+
+
+def test_plain_greeting_does_not_pull_api_call():
+    # Guard against over-matching: an unrelated message stays low-signal and must
+    # not drag the integration tool into context.
+    intent = agent_loop._classify_agent_request([], "hey there, how are you")
+    assert "integrations" not in intent["domains"], intent
+    assert "api_call" not in _selected_tools(intent["domains"])

tests/test_api_chat_security.py

@@ -219,6 +219,9 @@ class _WebhookManager:
     async def fire(self, event, payload):
         return None
 
+    def fire_and_forget(self, event, payload):
+        return None
+
 
 def _install_sync_chat_stubs(monkeypatch):
     # FastAPI checks for python_multipart at import time when Form is used;

tests/test_auth_policy.py

@@ -0,0 +1,272 @@
+"""Tests for auth policy endpoint and password length validation."""
+
+import asyncio
+import importlib
+import sys
+import types
+from pathlib import Path
+from types import SimpleNamespace
+from unittest.mock import MagicMock
+
+import pytest
+from fastapi import HTTPException
+
+from tests.helpers.import_state import clear_module
+
+
+def _real_core_package():
+    root = Path(__file__).resolve().parent.parent
+    core_path = str(root / "core")
+    core = sys.modules.get("core")
+    if core is None:
+        core = types.ModuleType("core")
+        sys.modules["core"] = core
+    core.__path__ = [core_path]
+    clear_module("core.auth")
+    return core
+
+
+def _auth_module():
+    _real_core_package()
+    return importlib.import_module("core.auth")
+
+
+def _make_manager(tmp_path):
+    auth_mod = _auth_module()
+    auth_mod._hash_password = lambda password: f"hash:{password}"
+    auth_mod._verify_password = lambda password, hashed: hashed == f"hash:{password}"
+    auth_path = tmp_path / "auth.json"
+    mgr = auth_mod.AuthManager(str(auth_path))
+    return mgr
+
+
+async def _immediate_to_thread(fn, *args, **kwargs):
+    return fn(*args, **kwargs)
+
+
+# ── AuthManager.policy() ───────────────────────────────────────────────
+
+
+def test_policy_returns_password_min_length(tmp_path):
+    mgr = _make_manager(tmp_path)
+    policy = mgr.policy()
+    assert policy["password_min_length"] == 8
+
+
+def test_policy_returns_reserved_usernames(tmp_path):
+    mgr = _make_manager(tmp_path)
+    policy = mgr.policy()
+    assert "internal-tool" in policy["reserved_usernames"]
+    assert "api" in policy["reserved_usernames"]
+    assert "demo" in policy["reserved_usernames"]
+    assert "system" in policy["reserved_usernames"]
+    assert isinstance(policy["reserved_usernames"], list)
+
+
+def test_policy_returns_signup_enabled(tmp_path):
+    mgr = _make_manager(tmp_path)
+    policy = mgr.policy()
+    assert policy["signup_enabled"] is False  # default
+
+
+def test_policy_returns_session_days(tmp_path):
+    mgr = _make_manager(tmp_path)
+    policy = mgr.policy()
+    assert policy["session_days"] == 7
+
+
+# ── GET /api/auth/policy endpoint ──────────────────────────────────────
+
+
+def _policy_endpoint(auth_manager):
+    sys.modules.pop("routes.auth_routes", None)
+    _real_core_package()
+    from routes.auth_routes import setup_auth_routes
+
+    router = setup_auth_routes(auth_manager)
+    for route in router.routes:
+        if getattr(route, "path", None) == "/api/auth/policy":
+            return route.endpoint
+    raise AssertionError("policy route not found")
+
+
+def test_policy_endpoint_returns_dict(tmp_path):
+    mgr = _make_manager(tmp_path)
+    endpoint = _policy_endpoint(mgr)
+    result = asyncio.run(endpoint())
+    assert isinstance(result, dict)
+    assert "password_min_length" in result
+    assert "reserved_usernames" in result
+    assert "signup_enabled" in result
+    assert "session_days" in result
+
+
+def test_policy_endpoint_values_match_manager(tmp_path):
+    mgr = _make_manager(tmp_path)
+    endpoint = _policy_endpoint(mgr)
+    result = asyncio.run(endpoint())
+    assert result == mgr.policy()
+
+
+# ── Password length validation ─────────────────────────────────────────
+
+
+def _setup_endpoint(auth_manager):
+    sys.modules.pop("routes.auth_routes", None)
+    _real_core_package()
+    from routes.auth_routes import SetupRequest, setup_auth_routes
+
+    router = setup_auth_routes(auth_manager)
+    for route in router.routes:
+        if getattr(route, "path", None) == "/api/auth/setup":
+            return route.endpoint, SetupRequest
+    raise AssertionError("setup route not found")
+
+
+def _signup_endpoint(auth_manager):
+    sys.modules.pop("routes.auth_routes", None)
+    _real_core_package()
+    from routes.auth_routes import SignupRequest, setup_auth_routes
+
+    router = setup_auth_routes(auth_manager)
+    for route in router.routes:
+        if getattr(route, "path", None) == "/api/auth/signup":
+            return route.endpoint, SignupRequest
+    raise AssertionError("signup route not found")
+
+
+def _change_password_endpoint(auth_manager):
+    sys.modules.pop("routes.auth_routes", None)
+    _real_core_package()
+    from routes.auth_routes import ChangePasswordRequest, setup_auth_routes
+
+    router = setup_auth_routes(auth_manager)
+    for route in router.routes:
+        if getattr(route, "path", None) == "/api/auth/change-password":
+            return route.endpoint, ChangePasswordRequest
+    raise AssertionError("change-password route not found")
+
+
+def test_setup_rejects_short_password(tmp_path):
+    mgr = _make_manager(tmp_path)
+    endpoint, SetupRequest = _setup_endpoint(mgr)
+    request = SimpleNamespace(client=SimpleNamespace(host="127.0.0.1"))
+    body = SetupRequest(username="admin", password="short")
+
+    with pytest.raises(HTTPException) as exc:
+        asyncio.run(endpoint(body=body, request=request))
+
+    assert exc.value.status_code == 400
+    assert "8 characters" in exc.value.detail
+
+
+def test_signup_rejects_short_password(tmp_path):
+    mgr = _make_manager(tmp_path)
+    mgr.create_user("admin", "admin-password", is_admin=True)
+    mgr.signup_enabled = True
+    endpoint, SignupRequest = _signup_endpoint(mgr)
+    request = SimpleNamespace(client=SimpleNamespace(host="127.0.0.1"))
+    body = SignupRequest(username="newuser", password="short")
+
+    with pytest.raises(HTTPException) as exc:
+        asyncio.run(endpoint(body=body, request=request))
+
+    assert exc.value.status_code == 400
+    assert "8 characters" in exc.value.detail
+
+
+def test_change_password_rejects_short_password(tmp_path):
+    mgr = _make_manager(tmp_path)
+    mgr.create_user("alice", "old-password", is_admin=False)
+    endpoint, ChangePasswordRequest = _change_password_endpoint(mgr)
+    request = SimpleNamespace(
+        cookies={"odysseus_session": "current-token"},
+        client=SimpleNamespace(host="127.0.0.1"),
+    )
+    # Mock get_username_for_token to return alice
+    mgr.get_username_for_token = MagicMock(return_value="alice")
+    body = ChangePasswordRequest(current_password="old-password", new_password="short")
+
+    with pytest.raises(HTTPException) as exc:
+        asyncio.run(endpoint(body=body, request=request))
+
+    assert exc.value.status_code == 400
+    assert "8 characters" in exc.value.detail
+
+
+def test_setup_accepts_exactly_min_length_password(tmp_path):
+    mgr = _make_manager(tmp_path)
+    endpoint, SetupRequest = _setup_endpoint(mgr)
+    request = SimpleNamespace(client=SimpleNamespace(host="127.0.0.1"))
+    body = SetupRequest(username="admin", password="12345678")
+
+    result = asyncio.run(endpoint(body=body, request=request))
+
+    assert result == {"ok": True, "message": "Admin account created"}
+
+
+def test_setup_rejects_seven_char_password(tmp_path):
+    mgr = _make_manager(tmp_path)
+    endpoint, SetupRequest = _setup_endpoint(mgr)
+    request = SimpleNamespace(client=SimpleNamespace(host="127.0.0.1"))
+    body = SetupRequest(username="admin", password="1234567")
+
+    with pytest.raises(HTTPException) as exc:
+        asyncio.run(endpoint(body=body, request=request))
+
+    assert exc.value.status_code == 400
+
+
+# ── Login "remember me" cookie lifetime ────────────────────────────────
+
+
+class _CapturingResponse:
+    """Stand-in for fastapi.Response that records set_cookie kwargs."""
+
+    def __init__(self):
+        self.cookie_kwargs = None
+
+    def set_cookie(self, **kwargs):
+        self.cookie_kwargs = kwargs
+
+
+def _login_endpoint(auth_manager):
+    sys.modules.pop("routes.auth_routes", None)
+    _real_core_package()
+    from routes.auth_routes import LoginRequest, setup_auth_routes
+
+    router = setup_auth_routes(auth_manager)
+    for route in router.routes:
+        if getattr(route, "path", None) == "/api/auth/login":
+            return route.endpoint, LoginRequest
+    raise AssertionError("login route not found")
+
+
+def test_remember_cookie_max_age_matches_token_ttl(tmp_path):
+    auth_mod = _auth_module()
+    mgr = _make_manager(tmp_path)
+    mgr.create_user("alice", "alice-password", is_admin=False)
+    endpoint, LoginRequest = _login_endpoint(mgr)
+    request = SimpleNamespace(client=SimpleNamespace(host="127.0.0.1"))
+    response = _CapturingResponse()
+    body = LoginRequest(username="alice", password="alice-password", remember=True)
+
+    result = asyncio.run(endpoint(body=body, request=request, response=response))
+
+    assert result == {"ok": True, "username": "alice"}
+    # The persistent cookie must outlive neither more nor less than the token.
+    assert response.cookie_kwargs["max_age"] == auth_mod.TOKEN_TTL
+
+
+def test_no_remember_omits_cookie_max_age(tmp_path):
+    mgr = _make_manager(tmp_path)
+    mgr.create_user("bob", "bob-password", is_admin=False)
+    endpoint, LoginRequest = _login_endpoint(mgr)
+    request = SimpleNamespace(client=SimpleNamespace(host="127.0.0.1"))
+    response = _CapturingResponse()
+    body = LoginRequest(username="bob", password="bob-password", remember=False)
+
+    asyncio.run(endpoint(body=body, request=request, response=response))
+
+    # Without "remember", the cookie is a session cookie (no max_age).
+    assert "max_age" not in response.cookie_kwargs

tests/test_auth_session_revocation.py

@@ -80,6 +80,16 @@ def test_password_change_allows_new_password_and_blocks_old_password(tmp_path):
     assert mgr.create_session("alice", "new-password") is not None
 
 
+def test_create_session_trusted_rejects_username_renamed_after_verification(tmp_path):
+    mgr = _make_manager(tmp_path)
+    assert mgr.create_user("admin", "admin-password", is_admin=True)
+
+    assert mgr.verify_password("alice", "old-password") is True
+    assert mgr.rename_user("alice", "alice2", "admin") is True
+
+    assert mgr.create_session_trusted("alice") is None
+
+
 def _change_password_endpoint(auth_manager):
     sys.modules.pop("routes.auth_routes", None)
     _real_core_package()
@@ -92,6 +102,39 @@ def _change_password_endpoint(auth_manager):
     raise AssertionError("change-password route not found")
 
 
+def _login_endpoint(auth_manager):
+    sys.modules.pop("routes.auth_routes", None)
+    _real_core_package()
+    from routes.auth_routes import LoginRequest, setup_auth_routes
+
+    router = setup_auth_routes(auth_manager)
+    for route in router.routes:
+        if getattr(route, "path", None) == "/api/auth/login":
+            return route.endpoint, LoginRequest
+    raise AssertionError("login route not found")
+
+
+def test_login_route_does_not_set_cookie_when_trusted_session_rejects_stale_user(monkeypatch):
+    auth = MagicMock()
+    auth.verify_password.return_value = True
+    auth.totp_enabled.return_value = False
+    auth.create_session_trusted.return_value = None
+    endpoint, LoginRequest = _login_endpoint(auth)
+    monkeypatch.setattr(
+        "routes.auth_routes.asyncio.to_thread",
+        lambda fn, *args, **kwargs: _immediate_to_thread(fn, *args, **kwargs),
+    )
+    request = SimpleNamespace(client=SimpleNamespace(host="127.0.0.1"))
+    response = MagicMock()
+    body = LoginRequest(username="alice", password="old-password")
+
+    with pytest.raises(HTTPException) as exc:
+        asyncio.run(endpoint(body=body, request=request, response=response))
+
+    assert exc.value.status_code == 401
+    response.set_cookie.assert_not_called()
+
+
 def test_change_password_route_revokes_other_sessions_after_success(monkeypatch):
     auth = MagicMock()
     auth.get_username_for_token.return_value = "alice"

tests/test_bg_job_tools.py

@@ -0,0 +1,174 @@
+"""Tests for bg_jobs.kill and the manage_bg_jobs agent tool.
+
+Process-free: the store/dir are redirected to tmp, _pid_alive is forced True so
+seeded "running" jobs stay running through refresh(), and _kill is stubbed so no
+real signal is sent. Jobs are scoped to a chat (session_id), which is the main
+invariant under test.
+"""
+import asyncio
+import json
+import time
+
+import pytest
+
+from src import bg_jobs
+from src.agent_tools.bg_job_tools import ManageBgJobsTool
+
+
+@pytest.fixture
+def store(tmp_path, monkeypatch):
+    jobs_dir = tmp_path / "bg_jobs"
+    jobs_dir.mkdir()
+    monkeypatch.setattr(bg_jobs, "_STORE", tmp_path / "bg_jobs.json")
+    monkeypatch.setattr(bg_jobs, "_JOBS_DIR", jobs_dir)
+    monkeypatch.setattr(bg_jobs, "_pid_alive", lambda pid: True)
+    killed: list = []
+    monkeypatch.setattr(bg_jobs, "_kill", lambda pid: killed.append(pid))
+    return {"dir": jobs_dir, "killed": killed}
+
+
+def _seed(session_id="sess-a", status="running", job_id="job0001", output="", pid=4321):
+    rec = {
+        "id": job_id, "session_id": session_id, "command": "sleep 60",
+        "status": status, "pid": pid, "started_at": time.time(),
+        "ended_at": None if status == "running" else time.time(),
+        "exit_code": None if status == "running" else 0,
+        "max_runtime_s": 3600, "followed_up": False,
+        "log_path": str(bg_jobs._JOBS_DIR / f"{job_id}.log"),
+        "exit_path": str(bg_jobs._JOBS_DIR / f"{job_id}.exit"),
+    }
+    if output:
+        (bg_jobs._JOBS_DIR / f"{job_id}.log").write_text(output, encoding="utf-8")
+    jobs = bg_jobs._load()
+    jobs[job_id] = rec
+    bg_jobs._save(jobs)
+    return rec
+
+
+def _run(args, session_id="sess-a"):
+    return asyncio.run(ManageBgJobsTool().execute(json.dumps(args), {"session_id": session_id, "owner": None}))
+
+
+# ── bg_jobs.kill ────────────────────────────────────────────────────────────
+
+def test_kill_marks_killed_and_suppresses_followup(store):
+    _seed(job_id="job0001", pid=4321)
+    rec = bg_jobs.kill("job0001")
+    assert rec["status"] == "failed"
+    assert rec["killed"] is True
+    assert rec["exit_code"] == -1
+    # followed_up True so the monitor won't ALSO auto-continue a deliberate kill.
+    assert rec["followed_up"] is True
+    assert store["killed"] == [4321]
+
+
+def test_kill_unknown_job_returns_none(store):
+    assert bg_jobs.kill("nope") is None
+
+
+def test_kill_finished_job_is_noop(store):
+    _seed(job_id="done01", status="done")
+    rec = bg_jobs.kill("done01")
+    assert rec["status"] == "done"
+    assert store["killed"] == []  # no signal sent to an already-finished job
+
+
+def test_result_text_reports_killed(store):
+    rec = _seed(job_id="job0001")
+    bg_jobs.kill("job0001")
+    assert "killed" in bg_jobs.result_text(bg_jobs.get("job0001")).lower()
+
+
+# ── manage_bg_jobs tool ─────────────────────────────────────────────────────
+
+def test_no_session_is_rejected(store):
+    out = asyncio.run(ManageBgJobsTool().execute('{"action":"list"}', {"session_id": None}))
+    assert "error" in out
+
+
+def test_list_empty(store):
+    assert "No background jobs" in _run({"action": "list"})["output"]
+
+
+def test_list_scoped_to_session(store):
+    _seed(session_id="sess-a", job_id="aaaa")
+    _seed(session_id="sess-b", job_id="bbbb")
+    out = _run({"action": "list"}, session_id="sess-a")["output"]
+    assert "aaaa" in out and "bbbb" not in out
+
+
+def test_output_returns_captured_log(store):
+    _seed(job_id="job0001", output="hello from the job\n")
+    out = _run({"action": "output", "job_id": "job0001"})["output"]
+    assert "hello from the job" in out
+
+
+def test_output_cross_session_denied(store):
+    _seed(session_id="sess-a", job_id="job0001", output="secret")
+    out = _run({"action": "output", "job_id": "job0001"}, session_id="sess-b")
+    assert "error" in out and "secret" not in out.get("error", "")
+
+
+def test_kill_via_tool(store):
+    _seed(job_id="job0001", pid=999)
+    out = _run({"action": "kill", "job_id": "job0001"})
+    assert "Killed" in out["output"]
+    assert store["killed"] == [999]
+    assert bg_jobs.get("job0001")["killed"] is True
+
+
+def test_kill_cross_session_denied(store):
+    _seed(session_id="sess-a", job_id="job0001")
+    out = _run({"action": "kill", "job_id": "job0001"}, session_id="sess-b")
+    assert "error" in out
+    assert store["killed"] == []  # never touched another chat's job
+
+
+def test_kill_requires_job_id(store):
+    assert "error" in _run({"action": "kill"})
+
+
+def test_unknown_action(store):
+    assert "error" in _run({"action": "frobnicate"})
+
+
+def test_action_aliases(store):
+    _seed(job_id="job0001", output="aliased")
+    # 'read' aliases to output, 'jobs' to list, 'stop' to kill
+    assert "aliased" in _run({"action": "read", "job_id": "job0001"})["output"]
+    assert "job0001" in _run({"action": "jobs"})["output"]
+    assert "Killed" in _run({"action": "stop", "job_id": "job0001"})["output"]
+
+
+# ── intent classifier: short bg-job commands must not be dropped as low-signal ─
+# A short imperative ("kill that job") otherwise trips the low-signal gate, which
+# skips tool retrieval entirely and never surfaces manage_bg_jobs (the live bug
+# this feature hit). These lock in that bg-job control reaches the files domain.
+
+
+@pytest.mark.parametrize("msg", [
+    "stop the job",
+    "kill that job",
+    "Now kill that background job.",
+    "is the job done?",
+    "check the job output",
+    "list my jobs",
+    "kill the bg task",
+])
+def test_bg_job_commands_are_not_low_signal(msg):
+    from src.agent_loop import _classify_agent_request, _DOMAIN_TOOL_MAP
+    r = _classify_agent_request([{"role": "user", "content": msg}], msg)
+    assert r["low_signal"] is False
+    assert "files" in r["domains"]
+    # files domain seeds manage_bg_jobs, so it gets offered to the model.
+    assert "manage_bg_jobs" in _DOMAIN_TOOL_MAP["files"]
+
+
+@pytest.mark.parametrize("msg", [
+    "run this in the background",   # launching, not managing
+    "find me a job listing",        # unrelated use of "job"
+])
+def test_non_bg_messages_do_not_trip_files_domain(msg):
+    from src.agent_loop import _classify_agent_request
+    r = _classify_agent_request([{"role": "user", "content": msg}], msg)
+    assert "files" not in r["domains"]

tests/test_builtin_actions_owner_scope.py

@@ -1,5 +1,6 @@
 """Regression tests for owner-scoped model resolution in scheduled actions."""
 
+import sqlite3
 from datetime import datetime
 from types import SimpleNamespace
 
@@ -138,6 +139,108 @@ async def test_learn_sender_signatures_resolves_llm_for_task_owner(monkeypatch):
     assert imap_owners == ["alice"]
 
 
+@pytest.mark.asyncio
+async def test_learn_sender_signatures_writes_owner_scoped_cache(monkeypatch, tmp_path):
+    from routes import email_helpers
+    from src import endpoint_resolver, llm_core
+    from src.builtin_actions import action_learn_sender_signatures
+
+    db_path = tmp_path / "scheduled_emails.db"
+    monkeypatch.setattr(email_helpers, "SCHEDULED_DB", db_path)
+    email_helpers._init_scheduled_db()
+
+    conn = sqlite3.connect(db_path)
+    try:
+        conn.execute(
+            """
+            INSERT INTO sender_signatures
+            (from_address, owner, signature_text, sample_count, last_built_at, model_used, source)
+            VALUES (?, ?, ?, ?, ?, ?, ?)
+            """,
+            (
+                "writer@example.com",
+                "bob",
+                "bob cached signature",
+                3,
+                "2999-01-01T00:00:00",
+                "old-model",
+                "llm",
+            ),
+        )
+        conn.commit()
+    finally:
+        conn.close()
+
+    class FakeImap:
+        def select(self, *_args, **_kwargs):
+            return "OK", []
+
+        def search(self, *_args, **_kwargs):
+            return "OK", [b"1 2 3"]
+
+        def fetch(self, uid, query):
+            if "HEADER.FIELDS" in query:
+                return "OK", [(None, b"From: Writer <writer@example.com>\r\n\r\n")]
+            return "OK", [
+                (
+                    None,
+                    (
+                        b"Thanks for the update.\r\n\r\n"
+                        b"Regards,\r\n"
+                        b"Writer Example\r\n"
+                        b"Example Co.\r\n"
+                        + str(uid).encode()
+                    ),
+                )
+            ]
+
+        def logout(self):
+            return None
+
+    imap_owners = []
+
+    def fake_imap_connect(_account_id=None, owner=""):
+        imap_owners.append(owner)
+        return FakeImap()
+
+    monkeypatch.setattr(email_helpers, "_imap_connect", fake_imap_connect)
+    monkeypatch.setattr(
+        endpoint_resolver,
+        "resolve_endpoint",
+        lambda kind, *args, **kwargs: ("http://llm", "alice-model", {}),
+    )
+
+    async def fake_llm_call_async(**_kwargs):
+        return "Writer Example\nExample Co.\nwriter@example.com"
+
+    monkeypatch.setattr(llm_core, "llm_call_async", fake_llm_call_async)
+
+    message, ok = await action_learn_sender_signatures("alice")
+
+    assert ok is True
+    assert message.startswith("Learned sigs: 1 found")
+    assert imap_owners == ["alice", "alice"]
+
+    conn = sqlite3.connect(db_path)
+    try:
+        rows = conn.execute(
+            """
+            SELECT owner, signature_text, model_used
+            FROM sender_signatures
+            WHERE from_address = ?
+            ORDER BY owner
+            """,
+            ("writer@example.com",),
+        ).fetchall()
+    finally:
+        conn.close()
+
+    assert rows == [
+        ("alice", "Writer Example\nExample Co.\nwriter@example.com", "alice-model"),
+        ("bob", "bob cached signature", "old-model"),
+    ]
+
+
 @pytest.mark.asyncio
 async def test_check_email_urgency_resolves_llm_candidates_for_task_owner(monkeypatch, tmp_path):
     from core import database

tests/test_cache_affinity_local_only.py

@@ -38,6 +38,16 @@ def test_unknown_public_host_gets_no_affinity_fields(monkeypatch):
     assert payload == {}
 
 
+@pytest.mark.parametrize("url", [
+    "https://10.example-cloud.com/v1",
+    "https://172.16.example-cloud.com/v1",
+    "https://192.168.example-cloud.com/v1",
+])
+def test_private_prefix_dns_host_gets_no_affinity_fields(monkeypatch, url):
+    payload = _affinity_fields(url, monkeypatch)
+    assert payload == {}
+
+
 def test_localhost_server_gets_affinity_fields(monkeypatch):
     payload = _affinity_fields("http://localhost:8080/v1", monkeypatch)
     assert payload == {"session_id": "sess-123", "cache_prompt": True}