Ln 522 Manual Tester

Generates executable bash test scripts in tests/manual/ that verify Story acceptance criteria against running code. It's the hands-on verification step after implementation tasks are marked Done, creating permanent test scripts (not throwaway checks) that hit APIs with curl or UIs with Puppeteer. The key discipline here is diff-based validation against expected output files, no heuristics or "close enough" checks. Scripts fail fast, save results to a gitignored results folder, and the skill posts structured pass/fail matrices back to your tracker. If you're running test-planning workflows where ln-521 does research and this does execution, it slots in as the L3 worker that produces evidence for quality gates.

Install to Claude Code

npx -y skills add levnikolaevich/claude-code-skills --skill ln-522-manual-tester --agent claude-code