the-catalog-auditor
Pass
Audited by Gen Agent Trust Hub on Aug 14, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill incorporates a comprehensive security framework via
references/agent-security.md, which provides explicit rules for treating all processed data as untrusted. It mandates that any instructions found within fetched or pasted content must be reported as findings rather than obeyed, effectively mitigating indirect prompt injection risks. - [SAFE]: It implements best practices for secret management by instructing the agent to identify potential credentials (like API keys or tokens) in catalog exports and recommend their rotation without reproducing or persisting the sensitive values.
- [SAFE]: The skill includes robust output sanitization requirements, such as HTML-escaping interpolated values and validating URLs to prevent cross-site scripting (XSS) in generated deliverables.
- [SAFE]: The implementation logic uses defensive processing methods, such as chunked data handling (150 rows at a time), to ensure full catalog coverage and prevent the agent from providing unreliable summaries of large datasets.
Audit Metadata