api-design-review
Pass
Audited by Gen Agent Trust Hub on Aug 2, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No malicious patterns or security risks were detected. The skill demonstrates defensive engineering by including explicit policies for data handling and agent behavior.
- [PROMPT_INJECTION]: The skill is protected against prompt injection through the
BF-UNTRUSTED-1policy, which mandates that instructions found in external artifacts be ignored and treated only as evidence. - [DATA_EXFILTRATION]: Data exfiltration risks are mitigated by the
BF-SECURITY-1policy, which requires the redaction of secrets and sensitive information from the agent's output. - [COMMAND_EXECUTION]: Command execution is strictly governed by the
BF-COMMAND-1policy, which restricts the agent to necessary and safe commands, explicitly forbidding the execution of untrusted or generated code. - [REMOTE_CODE_EXECUTION]: No remote code execution vulnerabilities were found. The included Python validation script operates only on local skill files and does not perform network operations or execute untrusted code.
- [EXTERNAL_DOWNLOADS]: The skill refers only to well-known, trusted industry tools and services (such as OpenAPI and Spectral) and does not contain any suspicious external download links.
Audit Metadata