test(coverage): add unit tests for agent-profiles context fetching and hintConfig trailing messages

[aaight]

Summary

• Adds comprehensive unit tests for src/backends/agent-profiles.ts context-fetching functions, covering all 4 orchestration paths (fetchWorkItemContext, fetchReviewContext, fetchCIContext, fetchPRCommentResponseContext) and their helper functions (fetchDirectoryListing, fetchContextFileInjections, fetchSquintOverview, fetchWorkItemInjection)
• Adds new tests for src/config/hintConfig.ts covering the implementation agent trailing message builder, formatIterationStatus urgency levels (no emoji / ⚠️ / 🚨), and formatDiagnosticLoopWarning threshold behavior
• Extends the existing agent-profiles.test.ts mock setup with readPRFileContents, resolveSquintDbPath, and execFileSync mocks needed for context tests
• Total: 57 tests in agent-profiles.test.ts (+29 new context tests) and 29 tests in hintConfig.test.ts (+18 new tests) — 86 tests in these two files

Test plan

• npm test — all 3139 tests pass (184 files)
• npm run lint:fix — no errors (1 pre-existing warning in unrelated file)
• npm run typecheck — zero errors

Card

https://trello.com/c/79qkpRR2/109-find-top-area-that-needs-test-coverage-and-plan-new-tests

🤖 Generated with Claude Code

[nhopeatall]

🤖 On it — checking the PR now

---

Progress: [██░░░░░░░░] 21% (iteration 15/70)

🔍 Code Review Update (1 min)
I've completed an initial review of the PR, focusing on the new unit tests for agent-profiles.ts and hintConfig.ts. I've examined the source files and the tests themselves, noting that params.input is accessed correctly. I'm currently analyzing the test quality and have identified a minor issue with an unused beforeEach import in hintConfig.test.ts.

_{Last updated: iteration 15 · review}

[nhopeatall]

LGTM - Well-structured test additions covering context-fetching orchestration paths in agent-profiles.ts and trailing message logic in hintConfig.ts. Tests correctly exercise happy paths, error handling (squint failures, work item errors), and edge cases (empty outputs, missing cardId). Mock isolation is proper with the top-level beforeEach clearing all mocks between tests. CI is green.