* fix: resolve portal startup hangs with non-blocking init
- Add socket_connect_timeout/socket_timeout (3s) to Redis connection in
SwarmComms to prevent infinite hangs when Redis is unreachable
- Defer reconcile_on_startup() from SwarmCoordinator.__init__() to an
explicit initialize() call during app lifespan, unblocking the
module-level singleton creation
- Make Ollama health checks non-blocking via asyncio.to_thread() so they
don't freeze the event loop for 2s per call
- Fix _check_redis() to reuse coordinator's SwarmComms singleton instead
of creating a new connection on every health check
- Move discord bot platform registration from lifespan critical path
into background task to avoid heavy import before yield
- Increase Docker healthcheck start_period from 10s/15s to 30s to give
the app adequate time to complete startup
https://claude.ai/code/session_016t5jNBYsUAQuyoR7sXe7Ux
* fix: disable commit signing in git_tools test fixture
The git_repo fixture inherits global gpgsign config, causing git_commit
to fail when the signing server rejects unsigned source context.
Disable signing in the temp repo's local config.
https://claude.ai/code/session_016t5jNBYsUAQuyoR7sXe7Ux
* fix: add dev extras for pip-based CI install
The CI workflow runs `pip install -e ".[dev]"` but after the Poetry
migration there was no `dev` extra defined — only a Poetry dev group.
This caused pytest to not be installed, resulting in exit code 127
(command not found) on every CI run.
Add a pip-compatible `dev` extra that mirrors the Poetry dev group
so both `pip install -e ".[dev]"` and `poetry install` work.
https://claude.ai/code/session_016t5jNBYsUAQuyoR7sXe7Ux
---------
Co-authored-by: Claude <noreply@anthropic.com>
* CI/CD Optimization: Guard Rails, Black Linting, and Pre-commit Hooks
- Fixed all test collection errors (Selenium imports, fixture paths, syntax)
- Implemented pre-commit hooks with Black formatting and isort
- Created comprehensive Makefile with test targets (unit, integration, functional, e2e)
- Added pytest.ini with marker definitions for test categorization
- Established guard rails to prevent future collection errors
- Wrapped optional dependencies (Selenium, MoviePy) in try-except blocks
- Added conftest_markers for automatic test categorization
This ensures a smooth development stream with:
- Fast feedback loops (pre-commit checks before push)
- Consistent code formatting (Black)
- Reliable CI/CD (no collection errors, proper test isolation)
- Clear test organization (unit, integration, functional, E2E)
* Fix CI/CD test failures:
- Export templates from dashboard.app
- Fix model name assertion in test_agent.py
- Fix platform-agnostic path resolution in test_path_resolution.py
- Skip Docker tests in test_docker_deployment.py if docker not available
- Fix test_model_fallback_chain logic in test_ollama_integration.py
* Add preventative pre-commit checks and Docker test skipif decorators:
- Create pre_commit_checks.py script for common CI failures
- Add skipif decorators to Docker tests
- Improve test robustness for CI environments
The task queue was completely stuck: 82 tasks trapped in pending_approval,
4 zombie tasks frozen in running, and the worker loop unable to process
anything. This removes the approval gate as the default and adds startup
recovery for orphaned tasks.
- Auto-approve all tasks by default; only task_type="escalation" requires
human review (and escalations never block the processor)
- Add reconcile_zombie_tasks() to reset RUNNING→APPROVED on startup
- Use in-memory _current_task for concurrency check instead of DB status
so stale RUNNING rows from a crash can't block new work
- Update get_next_pending_task to only query APPROVED tasks
- Update all callsites (chat route, API, form) to match new defaults
Co-authored-by: Alexander Payne <apayne@MM.local>
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
* feat: add task queue system for Timmy - all work goes through the queue
- Add queue position tracking to task_queue models with task_type field
- Add TaskProcessor class that consumes tasks from queue one at a time
- Modify chat route to queue all messages for async processing
- Chat responses get 'high' priority to jump ahead of thought tasks
- Add queue status API endpoints for position polling
- Update UI to show queue position (x/y) and current task banner
- Replace thinking loop with task-based approach - thoughts are queued tasks
- Push responses to user via WebSocket instead of immediate HTTP response
- Add database migrations for existing tables
* feat: Timmy drains task queue on startup, backlogs unhandleable tasks
On spin-up, Timmy now iterates through all pending/approved tasks
immediately instead of waiting for the polling loop. Tasks without a
registered handler or with permanent errors are moved to a new
BACKLOGGED status with a reason, keeping the queue clear for work
Timmy can actually do.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
---------
Co-authored-by: Alexander Payne <apayne@MM.local>
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
Inspired by OpenClaw-RL's multi-model orchestration, this adds four
features for custom model management:
1. Custom model registry (infrastructure/models/registry.py) — SQLite-backed
registry for GGUF, safetensors, HF checkpoint, and Ollama models with
role-based lookups (general, reward, teacher, judge).
2. Per-agent model assignment — each swarm persona can use a different model
instead of sharing the global default. Resolved via registry assignment >
persona default > global default.
3. Runtime model management API (/api/v1/models) — REST endpoints to register,
list, assign, enable/disable, and remove custom models without restart.
Includes a dashboard page at /models.
4. Reward model scoring (PRM-style) — majority-vote quality evaluation of
agent outputs using a configurable reward model. Scores persist in SQLite
and feed into the swarm learner.
New config settings: custom_weights_dir, reward_model_enabled,
reward_model_name, reward_model_votes.
54 new tests covering registry CRUD, API endpoints, agent assignments,
role lookups, and reward scoring.
https://claude.ai/code/session_01V4iTozMwcE2gjfnCJdCugC
