fix(enrichment): circuit-break exhausted quota calls

[mikemikimike]

Summary

Closes #222.

When the enrichment LLM returns a definitive quota exhaustion error, repeated stores previously retried summarization and classification independently. This adds a shared, fail-soft cooldown circuit across both enrichment paths.

Changes

• Open the circuit on insufficient_quota/quota errors for a configurable ENRICHMENT_CIRCUIT_COOLDOWN_SECONDS period (default 300 seconds).
• Skip additional upstream LLM requests while the circuit is open, while preserving deterministic fallback behavior.
• Allow one probe after cooldown and close the circuit after a successful response.
• Expose circuit-open skips and recoveries through the enrichment status response.
• Add focused regression tests for open, skip, non-quota, and recovery behavior.

The default remains backward compatible for successful requests and non-quota transient failures. The change only suppresses further enrichment requests after a quota-related failure; stores remain fail-soft.

Verification

• python -m compileall -q automem app.py — passed.
• Direct circuit regression assertions — passed.
• python -m black --check app.py automem tests/test_enrichment_circuit.py — passed.
• python -m flake8 app.py automem --count --select=E9,F63,F7,F82 --show-source --statistics — passed.
• python -m pytest -q tests/test_enrichment_circuit.py — not runnable in this environment: pytest collection resolves an unrelated installed tests package (D:\project\tscodex\connectonion-pr\tests) and cannot import this checkout's tests.support; no test assertion was executed.
• Full make test and Docker integration tests were not run because the focused pytest collection issue must be resolved before they can provide meaningful results.

[mikemikimike]

Follow-up validation: reproduced and fixed a probe-state bug where a non-quota failure after cooldown could leave the circuit permanently blocking requests. Added regression coverage for probe recovery, classifier request suppression, summarizer request suppression, and non-quota behavior. Verified with compileall, Black, focused Flake8, and direct classifier/circuit assertions. Full pytest collection remains blocked locally by the unrelated installed tests package collision documented in the PR.

[mikemikimike]

Docker verification completed: docker compose run --rm flask-api sh -c "pip install --no-cache-dir -r requirements-dev.txt && python -m pytest -q tests --ignore=tests/benchmarks" — 664 passed, 12 skipped, 3 warnings in 6:06:41. The warnings are deprecations from huggingface_hub/hf_xet; no test failures.