Caveat verdict

improvement-evaluator

auto-improvement-evaluator

88
🟢 Trusted
No high-risk patterns surfaced by the deep scan — automated capability review, not behavioral proof.

Skill evaluation tool that runs task suites against candidate SKILL.md files using local Python scripts and optionally a local claude CLI; dynamic_eval capability is used for its legitimate purpose of measuring skill execution quality.

⚠ Flagged for review — coarse, uncorroborated signal, not a confirmed exploit. Review the config yourself before installing.

Automated static analysis — not a human review. Caveat flags capabilities, not confirmed intent, and can produce false positives. Disagree with this verdict? Use Dispute below.

45
security
100
transparency
90
maintenance

Findings (5)

Pattern match high

Dynamic __import__('os') — Python OS command execution

interfaces/judges.py · prose · downgraded · __import__("os")

Pattern match high

Accesses sensitive system files

tests/test_judges.py · prose · downgraded · /etc/passwd

Pattern match medium

subprocess execution — runs system commands from Python

interfaces/judges.py · prose · downgraded · subprocess.run(

Pattern match low

Popular HTTP library — network access

interfaces/judges.py · prose · downgraded · got

Pattern match low

Python shutil file operation — copies/moves/deletes files

interfaces/judges.py · prose · downgraded · shutil.rmtree(

Why the tier is capped

Execution sink present in raw bytes (Hard Floor: class D). Final tier capped at Caution — cannot be lifted by any downgrade, example-payload opt-in, or allowlist.

Permissions & capabilities

No declared permissions — minimal attack surface.

dynamic_eval

Is this flag fair?

Check another skill Browse the registry Auditing your own skills or configs? Use the API