What it does
When tests break, this skill first runs the whole suite, groups the failures by error type and affected module, then fixes them group by group and re-runs everything at the end. It fixes in a deliberate order, public dependencies before leaf cases, so one change does not cascade into five new reds, and it tells real regressions apart from environment problems. Say make the tests pass and the workflow takes over.
Works with
Any mainstream coding agent, text only; it runs the project's own make test or uv run pytest.
Source and safety
From the claude-skills-marketplace repo. Listing-time review 2026-11-05: there are no scripts; it only runs make test and uv run pytest locally, and uv may pull dependencies from PyPI (a legitimate install), flagged as justified. (This review was conducted when the skill was listed, not as ongoing monitoring. Sources are beyond our control — evaluate skills before using them.)
广告

