fix(evals): stabilize nightly evaluation suite#494

Merged

cocosheng-g merged 11 commits intomainfrom

fix/nightly-eval-failures-new

Apr 1, 2026

Commits on Mar 31, 2026

Fix nightly eval failures and remove flaky test
cocosheng-g
committed

Commits on Apr 1, 2026

Fix pr-review timeout, issue-fixer timeout, and gemini-scheduled-triage ReferenceError
cocosheng-g
committed
feat(evals): support skill activation via folder-based mocking in pr-review
cocosheng-g
committed
fix(evals): avoid prompt modifications by using specific test data and expanding keywords
cocosheng-g
committed
fix(evals): make expected findings assertion conditional and clear for empty-diff
cocosheng-g
committed
fix(evals): reinforce command execution in triage prompt
cocosheng-g
committed
fix(evals): add realistic content to test file to prevent timeouts
cocosheng-g
committed
fix(ci): remove trailing spaces in evals-nightly.yml
cocosheng-g
committed
chore(evals): delete unused mock-mcp-server.ts
cocosheng-g
committed
fix(evals): address code review feedback locally
cocosheng-g
committed
fix(evals): expand expected findings for architectural-violation
cocosheng-g
committed