adversarial-review
Pass
Audited by Gen Agent Trust Hub on Jun 13, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill operates as a structural prompt framework for adversarial analysis of local artifacts or GitHub Pull Requests. All external data retrieval is performed via legitimate
gh(GitHub CLI) commands to fetch PR metadata and diffs, which is consistent with the skill's stated purpose. - [SAFE]: The presence of a testing harness (
tests/eval.sh) and structural evaluation rules (tests/README.md) demonstrate best practices in skill development, asserting contract invariants without introducing dangerous execution patterns. - [SAFE]: The skill's instructions specifically warn against speculation without concrete examples and require a validation gate for critical findings, reducing the risk of hallucinated security issues or 'fluff' reports.
- [SAFE]: No hardcoded credentials, sensitive file access, or suspicious network operations were detected. Domain references are restricted to well-known academic or technical documentation sources (e.g., arXiv, GitHub).
Audit Metadata