
Verify Python Change Tests
Checks that Python changes are caught by tests by running mutants and adding tests until they fail.
- Repository
- sickn33/agentic-awesome-skills
- Stars
- 49,023
- Licence
- MIT
- Files
- 1 file
What it does
When Claude writes or changes Python code, this skill makes it prove the change is tested. It runs a mutation checker that deliberately breaks the changed lines in small ways and confirms the tests catch each break, adding or strengthening tests until they do. The user gets a report of evidence (which checks passed, which mutants were caught, which new tests fail on the old code) rather than having to read the diff.
Great for
- Python developers using pytest who want proof that new code is actually tested
- Teams reviewing AI-written Python changes before merging
- Anyone who wants a verdict on test coverage quality, not just a pass/fail on existing tests
Skip it if
- Non-Python projects, since it targets Python and pytest (wrong tool for other languages)
- Reviewing a whole test suite; the skill's own docs point to a different skill (agentic-test-audit) for that
- Situations needing guaranteed bug-free code: the skill states a runner verdict cannot prove every defect would be caught
Try asking
- “Check whether the change I just made is covered by tests before we call it done”
- “Run verify on my uncommitted work and tell me which mutants survived”
- “Compare my branch against main and show which new tests fail on the old code”
- “Write tests for the edge cases in this function until the mutation check passes”
Install in 3 steps
- 1
Download the skill
The ZIP contains the 1 original files, the MIT licence and an attribution note.
agentic-verify.zip - 2
Add it to your agent
Claude Code (all projects)
mkdir -p ~/.claude/skills && unzip agentic-verify.zip -d ~/.claude/skills/Claude Code / Agent SDK (this project only)
mkdir -p .claude/skills && unzip agentic-verify.zip -d .claude/skills/Claude apps (web/desktop): Settings → Capabilities → Skills → Upload skill, and choose the ZIP.
- 3
Ask for the task
No special command needed — the agent reads the skill's description and loads it when your request matches. Try one of the prompts above.
Get better results
- Pass the project root (where pytest runs), not a sub-folder or single file, or the check will not work correctly.
- Give the verifier a clear success criterion first; the skill asks for one before coding and tests come out stronger.
- Use a larger --budget-minutes for big changes, otherwise the result may be incomplete.
What's inside
- SKILL.md
Redistributed unmodified from sickn33/agentic-awesome-skills under the MIT licence. The assessment was drafted with AI from the skill's files on 2026-10-29. Skills can instruct your agent to run code — read SKILL.md and any scripts before installing.