meccha-chameleon-game-cheat-detection

Pass

Audited by Gen Agent Trust Hub on Jul 1, 2026

Risk Level: SAFE
Full Analysis
  • [PROMPT_INJECTION]: The skill uses natural instructional language and does not contain markers designed to bypass safety filters or override system instructions.
  • [EXTERNAL_DOWNLOADS]: The skill identifies 'skydock.netlify.app' as a suspicious indicator for the third-party software it is designed to detect. As Netlify is a well-known service, this reference is documented neutrally. The example code imports the 'requests' library, a standard package for HTTP operations.
  • [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze external data such as repository READMEs and metadata.
  • Ingestion points: 'readme_content' and 'repo_metadata' parameters.
  • Boundary markers: Absent in the warning response template.
  • Capability inventory: The skill logic is limited to text analysis and regex matching without dangerous system operations.
  • Sanitization: No sanitization is performed on analyzed content before interpolation into the response.
  • [SAFE]: The skill's primary purpose is defensive analysis and user warning, and its code does not contain malicious patterns like exfiltration or persistence.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 1, 2026, 06:58 AM
Security Audit — agent-trust-hub — meccha-chameleon-game-cheat-detection