26 files changed
- .claude/rules/testing.md+6−0
- .github/workflows/ai-eval.yml+64−0
- DECISIONS.md+26−0
- apps/web/src/components/runs/ai-trace.test.tsx+66−0
- apps/web/src/components/runs/ai-trace.tsx+54−0
- apps/web/src/components/runs/rule-evidence.tsx+17−0
- apps/worker/src/ai/generate.ts+7−5
- apps/worker/src/jobs/ai-review-delivery.test.ts+93−0
- apps/worker/src/jobs/fail-closed-floor.test.ts+81−0
- apps/worker/src/jobs/pr-surface.ts+6−5
- apps/worker/src/jobs/run-workflows.ts+18−4
- package.json+3−0
- packages/contracts/src/review.test.ts+101−0
- packages/contracts/src/review.ts+106−0
- packages/core/src/rules/ai-review/rule.test.ts+43−1
- scripts/e2e/harness.ts+17−4
- scripts/e2e/lib/interrupt.ts+18−0
- scripts/e2e/lib/runner.ts+29−14
- scripts/eval/config.ts+62−0
- scripts/eval/context.ts+91−0
- scripts/eval/fixtures.ts+270−0
- scripts/eval/prod-smoke.ts+151−0
- scripts/eval/run.ts+284−0
- scripts/eval/scorecards/2026-07-17T08-28-36-930Z_claude-haiku-4.5.json+218−0
- scripts/eval/scorecards/2026-07-17T09-05-00-166Z_grok-4.5.json+218−0
- scripts/eval/scorecards/2026-07-17T09-12-26-582Z_grok-4.5.json+218−0