AI agents are notorious for writing tests that look thorough but verify little: source restatements, defensive permutations, and mocks that distort production code.
I built test-value-audit for Codex and Claude Code to trim the noise and find real gaps.
github.com/samuelhe52/tes…
