poet-analyst
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFE
Full Analysis
- [NO_CODE] (SAFE): The provided skill consists entirely of Markdown documentation (README.md and QUICK_REFERENCE.md). There are no executable scripts (Python, JavaScript, Shell) or configuration files that trigger automated actions.
- [EXTERNAL_DOWNLOADS] (SAFE): The documentation contains links to legitimate educational and literary organizations (Academy of American Poets, Poetry Foundation). These are standard reference links and do not involve automated downloads or runtime dependencies.
- [DATA_EXFILTRATION] (SAFE): There is no logic for network operations, file system access, or credential handling. The skill is designed for internal natural language processing of user-provided text.
- [PROMPT_INJECTION] (SAFE): The instructions use standard natural language for task definition. No bypass markers, role-play jailbreaks, or instructions to ignore safety filters were detected.
- [INDIRECT_PROMPT_INJECTION] (LOW): As an analytical tool, the skill is designed to ingest and process untrusted user data (e.g., speeches, slogans).
- Ingestion points: Text inputs provided by the user for poetic analysis.
- Boundary markers: Not explicitly defined in the documentation; reliance on standard agent boundary handling.
- Capability inventory: No risky capabilities (subprocess, eval, network) are requested or utilized in the files provided.
- Sanitization: None described, but the risk is mitigated by the lack of executable capabilities.
Audit Metadata