security-review
Pass
Audited by Gen Agent Trust Hub on May 18, 2026
Risk Level: SAFE
Full Analysis
- [PROMPT_INJECTION]: No malicious bypass or override patterns detected. The instructions use authoritative language (e.g., "Don't skip them") solely to enforce the security auditing workflow, which is benign.
- [DATA_EXFILTRATION]: No network operations or attempts to access sensitive system files were found. The mention of 'curl' in the output template is illustrative for report writing and does not execute.
- [REMOTE_CODE_EXECUTION]: The skill does not contain executable code, scripts, or commands to download and run external payloads. It references other platform tools (e.g., secrets-scanner, sast-orchestrator) as part of a standard modular workflow.
- [INDIRECT_PROMPT_INJECTION]: While the skill is designed to ingest and process untrusted code from pull requests, the instructions include a 'verification-loop' and multi-phase analysis to ensure the agent maintains objectivity and does not execute instructions found within the audited data. This is a standard surface for a security tool and is handled safely here.
- [SAFE]: References provided in the skill point to well-known, trusted security organizations including OWASP, MITRE (CWE), NIST, and FIRST (CVSS).
Audit Metadata