claude-code-toolkit

Use when a feature branch is ready for merge — orchestrates PR creation, spec/quality check, review-loop, test-completeness, Copilot review, Codex review, doc consistency, conflict check, and CI verification in a strict sequential pipeline

Use when editing a long file (100+ lines of changes) or when an Edit call freezes, times out, or silently fails - requires splitting large edits into multiple smaller sequential Edit calls

Use when code changes need iterative multi-round review until no Critical or Important issues remain, with isolated independent reviewers, verification of findings, and parallel review after fixes

Use when you need to audit whether tests are complete for a project or PR — checks good case, bad case, boundary cases, 100% coverage, unit/integration/e2e test layers, then proposes and generates missing tests