search-first
Pass
Audited by Gen Agent Trust Hub on Sep 8, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONNO_CODE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to search and evaluate third-party libraries and documentation from public sources like npm, PyPI, and GitHub. This introduces a risk of indirect prompt injection where malicious content in external READMEs or package descriptions could influence the agent's behavior.
- Ingestion points: The 'Workflow' and 'Full Mode' sections in SKILL.md.
- Boundary markers: No delimiters or specific instructions to treat external data as untrusted are provided.
- Capability inventory: The agent is empowered to install packages and modify the codebase based on its research.
- Sanitization: No sanitization methods are specified for processing external data.
- [NO_CODE]: The skill consists solely of descriptive instructions and workflow documentation in Markdown format, without any executable code or scripts.
Audit Metadata