pr-status-triage

Triage CI failures by parsing job and thread reports from status scripts, applying blocker-first prioritization (build, lint, types, tests) to identify real issues. This approach separates actual failures from flaky tests and provides clear guidance for resolving review threads efficiently.

Blocker-first prioritization ranks CI failures by severity: build errors first, then lint, types, and finally tests. This ensures critical blocking issues are resolved before minor failures, streamlining the triage process and getting PRs merged faster.

Reproduce CI failures locally by aligning your environment variables with CI mode and running the relevant status scripts. Deterministic execution and environment-variable mirroring ensure the local failure matches the CI environment exactly before you attempt a fix.

Separate flaky tests from real failures by comparing job reports against a Known Flaky Tests list. This filtering mechanism isolates genuine regressions requiring immediate attention from intermittent test flakiness, focusing your triage efforts on actual code issues.

Start triage by running node scripts/pr-status.js --wait in the background to generate job and thread reports. These reports provide the necessary data for applying blocker-first prioritization and synthesizing actionable feedback for PR reviews.