- New /settings/tools page: max_risk selector (low/medium/high) + per-tool
override dropdowns (Default / Force include / Force exclude) for all 58 tools
grouped by category with color-coded risk badges; JS updates Auto status live
- get_tools_for_role() + get_openai_tools_for_role() now accept max_risk,
whitelist, blacklist; _apply_risk_policy() handles the filtering logic
- get_risk_policy() helper in auth_utils reads from tool_policy.json
- Risk policy wired through orchestrator.py, openai_orchestrator.py,
orchestrator_engine.py, nextcloud_talk.py, homeassistant.py
- Tools nav link added to settings.html and notifications.html
- CLAUDE.md and ARCH__SYSTEM.md updated: tool count 50→58, risk system docs,
tool access control three-layer model documented
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- inject_mode: per-role toggle (parallel to inject_datetime) gates the
"Current mode: Off The Record" line in the system prompt; wired through
model_registry, context_loader, chat router, orchestrator router, and
local_llm settings UI
- OTR orchestrator fix: OrchestrateRequest now carries off_record;
_finalize_job stores it per message and gates log_turn on it; JS
orchestrate payload sends off_record correctly
- Per-message hover metadata: removed always-visible .model-tag; replaced
with .msg-meta strip in the action bar (hover-only); shows model label,
host, fallback indicator, and OTR badge; stored in session JSON
- Send/stop button tooltip: shows role + model and (when tools on)
separate orchestrator model + engine label; live elapsed timer on stop
button via startRunTimer/stopRunTimer
- OrchestratorResult.backend_label: new field; openai_orchestrator fills
it; finalize_job propagates it to job dict and session messages
- GET /backend: exposes orchestrator_model label so the frontend tooltip
can show both models separately
- TODO: session delete confirmation added
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Mirrors the pattern already in openai_orchestrator.py. The Gemini engine
was still hardcoded to the global orchestrator_max_rounds setting.
- orchestrator_engine.py: max_rounds param on run() and _run_from_contents();
effective_limit = min(per_model_limit, global_limit); stored in checkpoint
so resume() respects it across confirmation gates
- routers/orchestrator.py: passes orch_model.get("max_rounds") to run()
- tools/agents.py: passes model_cfg.get("max_rounds") for gemini_api spawns
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Each role can now disable the current date/time header injected into the
system prompt. Default is true (all existing roles unchanged). Useful for
pure processing roles (summarizer, classifier, translator) where temporal
context is irrelevant or could cause unexpected model behavior.
Changes:
- model_registry: set_role_config/get_role_config gain inject_datetime field
- context_loader: load_context gains inject_datetime param (default True)
- orchestrator router: passes inject_datetime from role_cfg to load_context
- local_llm router: reads inject_datetime from POST body, passes to registry;
role_config_data_js includes the field
- local_llm.html: checkbox in role config panel; populate on open, save on submit
Session logs still timestamp every turn (HH:MM header in YYYY-MM-DD.md files)
regardless of this setting — the toggle only affects the system prompt header.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Each role in model_registry.json can now carry two optional keys:
system_append — injected into the system prompt at position 7 (after
memory, closest to the turn) for the active chat_role
tools — explicit tool allow-list; intersected with the user's
access-level filter so it can only restrict, never elevate
No changes needed for existing users — missing keys fall back to current
behavior. Add keys to a role to give it a specialty focus:
"coder": {
"primary": "claude_cli",
"system_append": "You are in code-specialist mode...",
"tools": ["web_search", "file_read", "shell_exec", "scratch_write"]
}
Changes:
- model_registry.py: get_role_config() returns system_append + tools
- context_loader.py: role_append param appended as "--- Role Context ---"
- tools/__init__.py: get_tools_for_role/get_openai_tools_for_role accept
optional tool_list and intersect with access-level filter
- orchestrator_engine.py: tool_list threaded through run/resume/checkpoint
- openai_orchestrator.py: tool_list threaded through run/resume/checkpoint;
_build_client now calls get_openai_tools_for_role instead of returning
unfiltered OPENAI_TOOL_SCHEMAS
- routers/orchestrator.py: pulls role_cfg for chat_role, passes both
role_append and tool_list to context loader and engine
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Fixes the broken confirmation gate where users had no way to approve
or deny a blocked tool call in the web UI.
Changes:
- orchestrator_engine.py: add OrchestrateCheckpoint dataclass, extract
loop into _run_from_contents(), add resume() function
- openai_orchestrator.py: same treatment — _run_from_messages(), resume()
- routers/orchestrator.py: POST /{job_id}/confirm and /deny endpoints,
separate _checkpoints store, _resume_job() + _finalize_job() helpers,
"awaiting_confirmation" job status with pending_confirmation payload
- auth_utils.py: get_tool_policy() and save_tool_policy() helpers reading
home/{user}/tool_policy.json (allow/deny lists)
- routers/orchestrator.py: loads tool_policy per user and passes
confirm_allow/confirm_deny to both engines
- app.js: poll loop handles awaiting_confirmation — shows Confirm/Deny
buttons inline, resumes polling after user action
- settings.html + settings.py: Tool Permissions section with allow/deny
textareas, POST /settings/tool-policy route
- style.css: .confirm-gate, .confirm-btn, .deny-btn styles
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- auth_utils: get_user_role() reads role from auth.json (admin|user, default user)
- manage_passwords: new `role` command to promote/demote users (admin-only by convention)
- tools/__init__: TOOL_ROLES map, CONFIRM_REQUIRED set, get_tools_for_role(),
get_openai_tools_for_role() — both orchestrators now filter tools by caller's role
- tools/system: cortex_restart (detached subprocess, 5s delay), cortex_logs (admin-only)
- tools/web: http_fetch — direct URL fetch, distinct from web_search
- tools/files: file_list (directory listing), file_write (restricted paths, admin-only)
- tools/notify: nc_talk_send — proactive outbound via notification.py
- orchestrator_engine + openai_orchestrator: user_role param; CONFIRM_REQUIRED tools
return a confirmation-request result instead of executing — loop breaks after Claude
asks user to confirm in a follow-up message
- home/scott/auth.json: role set to admin
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Remove 'agent' from mode dropdown; Chat/Note/OTR remain
- Add ⚡ tools toggle button in input bar (persisted in localStorage)
When on: routes to POST /orchestrate (Gemini tool loop); send btn → "Run"
When off: routes to POST /chat (direct to active role); no change
- Role selector and tools toggle are now fully independent:
active chat_role sent in orchestrate payload → used for final response
- orchestrator_engine.run() accepts response_role param; passes it to
complete(role=...) instead of hardcoded model="claude"
- OrchestrateRequest gains chat_role field (default "chat")
- Migrate stored 'agent' mode/MRU entries to 'chat' on load
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Add shell_exec to orchestrator tool suite (system.py + __init__.py)
Runs arbitrary shell commands on the Cortex host with timeout (1–120s),
combined stdout/stderr output, optional working_dir, and exit code reporting.
Enables system diagnostics (df, ls, ps, journalctl, etc.) from Agent mode.
- Fix orchestrator_engine.run() to use model_name from resolved registry entry
Previously used settings.orchestrator_model (.env hardcode) regardless of
what model was assigned to the orchestrator role. Now accepts model_name param
and falls back to settings value only when registry has no model_name.
- Update ARCH__FUTURE.md: date, running host, local orchestrator status,
model registry V2 progress, added Cortex Mesh concept (section 9)
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Adds a providers section to the per-user model registry for Anthropic and
Google as first-class providers alongside local hosts. Google accounts
(API keys) are now stored as a list so multiple Google accounts can coexist.
Changes:
- model_registry.py: V2 schema, auto migration V1→V2 (pulls gemini_api_key
from auth.json into providers.google.accounts), _resolve_model() merges
account API key for gemini_api type models
- routers/orchestrator.py: uses model-resolved api_key when orchestrator
role resolves to a gemini_api model with account_id
- ANTHROPIC_CATALOG and GOOGLE_CATALOG constants for model picker (Phase 2)
- New functions: get_google_api_key(), save/remove_google_account(), get_catalog()
- Documentation: ARCH__BACKENDS.md updated to V2 schema, DESIGN doc added
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- openai_orchestrator.py — new ReAct tool loop engine for any
OpenAI-compatible endpoint (OpenRouter, Open WebUI, Ollama, LiteLLM);
model handles both tool loop and final response, no Claude handoff needed
- tools/__init__.py — auto-derive OpenAI JSON Schema from existing Gemini
FunctionDeclarations so tool definitions have a single source of truth
- routers/orchestrator.py — route to openai_orchestrator when model registry
"orchestrator" role resolves to a local_openai type host
- routers/chat.py — pass role to _backend_label(); fix fallback_used logic
(only meaningful for explicit backend overrides, not auto-routing)
- static/app.js — add null/"auto" to backend cycle; fetch local model hint
without overriding the auto default on page load
- model_registry.py — _normalize() back-fills host_type on old registry files
- requirements.txt — add openai>=1.0.0
- ARCH__BACKENDS.md — document OpenAI-compat backend and routing logic
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
NameError: name 'user' is not defined in orchestrator._run_job —
user was resolved in the endpoint but not forwarded to the background task.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Users with Google accounts can now sign in without a password.
Auth flow:
- GET /auth/google → Google consent page (CSRF state cookie)
- GET /auth/google/callback → exchange code, lookup user, set JWT
- auth.json gains google_sub + google_email fields
- set_password() no longer overwrites unrelated auth.json fields
Admin setup:
python manage_passwords.py google-add <username> <email>
# add GOOGLE_CLIENT_ID + GOOGLE_CLIENT_SECRET to .env
Per-user Gemini key:
- get_user_gemini_key() reads gemini_api_key from auth.json
- orchestrator_engine.run() accepts gemini_api_key param
- orchestrator router passes user's key, falls back to server key
login.html: "Sign in with Google" button above the password form.
manage_passwords.py list: now shows auth method columns (pw / google).
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Restructures persona storage from a flat personas/{name}/ layout to
home/{username}/persona/{name}/, mirroring Linux home directories.
Changes:
- persona.py: two ContextVars (user + persona), Linux-style name validation,
set_context(), get_user(), get_persona(), validate(), list_users(),
list_user_personas(); persona_path() takes (username, name)
- config.py: replaces personas_dir with home_dir + home_root()
- git mv personas/inara → home/scott/persona/inara (history preserved)
- home/holly/persona/tina/: Holly's persona stub added
- cron_runner.py: all storage functions take (username, persona) params
- tools/cron.py: stamps user + persona on jobs; APScheduler IDs are
{user}:{persona}:{job_id} to prevent collisions across users
- memory_distiller.py: distill_short/mid/long take (username, persona);
added missing Path + settings imports
- scheduler.py: _load_user_crons() iterates home/*/persona/* (two-level)
- routers/chat.py, orchestrator.py: user field added; set_context() called
- tests/conftest.py: home_root fixture with two-level structure;
patches home_dir instead of personas_dir
- tests/test_persona.py: fully rewritten for two-level API
- tests/test_api_files.py: updated fixture name and path
- .env.default: documents HOME_DIR setting; scrubs stale API key
- CLAUDE.md, README.md: directory maps updated for new layout
All 80 tests pass.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Add cortex/persona.py: ContextVar-based per-request routing with
path traversal protection and persona validation
- Migrate inara/ → personas/inara/ (git history preserved via git mv)
- config.py: add personas_root(), inara_path() delegates to personas/inara
- All 14 settings.inara_path() call sites replaced with persona_path()
- ChatRequest + OrchestrateRequest: add persona field (default: "inara")
with validation at request entry before any processing
- memory_distiller: add optional persona param for future per-persona distill
- cron_runner/tools/cron: stamp persona on jobs, prefix APScheduler IDs
(persona:job_id) to prevent collisions across personas
- scheduler: _load_user_crons() iterates all personas at startup
Adding a new persona: create personas/<name>/ with IDENTITY.md + SOUL.md.
Auth: handled at nginx level (inject X-Cortex-Persona header per subdomain).
Future: persona maps to Aether account_id_random for full integration.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Orchestrator now saves turns to session store so history survives page refresh
- UI session_id updated from job result; history controls attached to agent turns
- Cortex migrated from system service to systemd user service (no more sudo)
- Update README.md and CLAUDE.md with correct service commands
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
