feat(ai-service): add per-claim verification cooldown

[omoh25]

Enforce a configurable minimum interval between verification attempts for the same claim identity on both the humanitarian and fraud endpoints.

• services/verification_cooldown.py: in-memory, thread-safe, LRU-bounded cooldown service. check() returns (in_cooldown, prior_result); record() stores the fresh outcome. cooldown_seconds=0 disables it entirely. Process-wide singleton via get_cooldown_service(), injectable via app.state.cooldown_service for tests.
• schemas/common.py: add from_cooldown: bool field to ResultEnvelope so callers can distinguish a replayed response from a fresh computation.
• config.py: add verification_cooldown_seconds (default 60 s, env var VERIFICATION_COOLDOWN_SECONDS).
• api/v1/humanitarian.py: cooldown check runs after all access-control gates; hits return prior result with from_cooldown=True; fresh results are recorded in the store before returning.
• api/v1/fraud.py: same pattern, keyed on anchor claim_id; batches without an anchor always process fresh.
• .env.example: document VERIFICATION_COOLDOWN_SECONDS=60.
• tests/test_verification_cooldown.py: 32 tests covering inside/outside window, disabled cooldown, LRU eviction, key builders, and endpoint integration for both humanitarian and fraud.
  closes Add a Configurable Cooldown Between Repeated Verification Attempts for the Same Claim #1426

[drips-wave]

@omoh25 Great news! 🎉 Based on an automated assessment of this PR, the linked Wave issue(s) no longer count against your application limits.

You can now already apply to more issues while waiting for a review of this PR. Keep up the great work! 🚀

Learn more about application limits