Eval harness (#62 P4), status frames, and agentic runs (#63) (#71)

## Summary
- Closes Phase 4 of [#62](#62): `evals/suite.json` (≥40 graded questions), `run_evals` management command, and manually-triggered `.gitea/workflows/run-evals.yml`.
- Emits versioned WS `status` frames during grounded chat (evaluating / searching / reading_sources / refining / writing) for [chat_web_app#96](ai_ml_operations/chat_web_app#96).
- Implements [#63](#63): Redis/Celery optional infra, `AgentRun`/`AgentStep`, tool registry (SSRF-safe `fetch_url`, tenant-scoped docs), LangGraph orchestrator, progress frames, REST `GET/POST /api/agent_runs/…`, gated by `ALLOW_AGENTIC_TASKS` (default off).

## Test plan
- [x] `SKIP_RAG_INIT=1 uv run python manage.py test` for evals, ws frames, agent tools, consumers, grounding
- [ ] Manual: with `ALLOW_AGENTIC_TASKS=false`, chat identical to today
- [ ] Manual: status frames visible in FE with #96 branch
- [ ] Manual (GPU): `python manage.py run_evals --runs 3`
- [ ] Manual: `ALLOW_AGENTIC_TASKS=true` multi-step research prompt creates AgentRun + framesReviewed-on: #71

docker-compose.prod.yml

@@ -12,5 +12,17 @@ services:
       # Chroma vector index only (uploaded file blobs live in Postgres).
       - chroma_data:/app/llm_be/chroma_db
 
+  # Celery worker for long-running agent tasks (#63). Only started when the
+  # `agentic` profile is enabled and REDIS_URL/CELERY_BROKER_URL are set in
+  # .env (control-node secret) — points at a shared Redis instance, no
+  # bundled `redis` service here (mirrors the "no bundled Postgres" policy).
+  worker:
+    build: .
+    profiles: ["agentic"]
+    restart: unless-stopped
+    command: ["uv", "run", "celery", "-A", "llm_be", "worker", "--loglevel=info"]
+    env_file:
+      - .env
+
 volumes:
   chroma_data: