verification-loop

Automated pre-PR checks run sequential verification phases across build, type checking, linting, tests, security scans, and diff review. This process returns a consolidated pass/fail report with error counts, coverage percentage, and actionable issues to resolve before merging.

To prevent accidental secrets or console leakage, run security greps and diff inspection during your verification workflow. This catches regressions and hidden secrets before a pull request, producing an actionable list of security issues from the scan phase.

The best way to verify code quality after refactoring is running an end-to-end verification loop. This confirms the project builds, types cleanly, meets linting and coverage thresholds, and contains no obvious secret leakage before opening a pull request.

You can enforce quality gates and coverage targets before creating a pull request by applying a verification workflow. It executes sequential checks for build, type, lint, tests, and security, returning a clear pass/fail status with actionable items for any unmet thresholds.

The verification workflow operates with no external dependencies. It performs build, type checking, lint, tests, security scans, and diff review directly within your Claude Code session to produce pass/fail statuses and error counts.

Code passes local tests but fails end-to-end verification phases because the workflow enforces sequential checks beyond testing, including build, type checking, linting, security greps, and diff inspection, catching regressions or threshold misses local tests ignore.