writing-beats
Pass
Audited by Gen Agent Trust Hub on Sep 28, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes external markdown files as 'raw material' and re-reads the generated article from disk to determine subsequent 'beats'. This ingestion of untrusted data represents an indirect prompt injection surface.
- Ingestion points: Reads raw material markdown files and the destination article file as specified by the user.
- Boundary markers: The instructions direct the agent to 'mine the pile' to fill beats but lack explicit warnings to ignore or delimit embedded instructions within the source data.
- Capability inventory: The skill possesses file-read and file-write capabilities on the local system.
- Sanitization: No specific sanitization, filtering, or validation of the input raw material is defined to prevent instruction execution from data.
Audit Metadata