Caveat verdict

expertpack-eval

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

Receives external input AND uses eval

An evaluation tool that measures 'esoteric knowledge' ratio by calling OpenRouter API with user-provided credentials; all external API calls are for the skill's stated legitimate evaluation purpose.

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.

23
security
80
transparency
70
maintenance

What it does

These are capability combinations: each listed behavior occurs in the skill, but Caveat detects co-occurrence — it does not verify that one flows into another. Read the code to confirm a live chain.

Capability combination critical

Receives external input AND uses eval — the remote code-injection pattern (data-flow not verified)

LLM01 · LLM05 · ASI01 · ASI05

Findings (5)

Pattern match high

Possible hardcoded credential

scripts/run-eval.py · prose · downgraded · API_KEY='): api_key = line.split(

Pattern match medium

Accesses OpenClaw config/secrets directly

scripts/run-eval.py · prose · downgraded · ~/.openclaw/.env

Pattern match low

Python os.environ.get — reads environment variable

scripts/eval-ek.py · prose · downgraded · os.environ.get(

Pattern match low

Opens WebSocket connection

scripts/run-eval.py · prose · downgraded · WebSocket

Pattern match low

POSTs data to external URL

scripts/run-eval.py · prose · downgraded · .post( "https://

Permissions & capabilities

Requires 1 system binary.

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