concept-prd-cn
Pass
Audited by Gen Agent Trust Hub on Nov 1, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest untrusted data in the form of 'confirmed models' and 'requirement sources' to generate project documentation. This ingestion process lacks explicit boundary markers or sanitization steps, creating a potential vector for malicious instructions embedded in the input models to influence the generated output.\n
- Ingestion points: The skill reads external models, source documents, and existing PRD files to inform the transcription process (SKILL.md).\n
- Boundary markers: The instructions do not define specific delimiters or warnings to isolate untrusted input content from the agent's core instructions.\n
- Capability inventory: The skill enables the agent to perform file-reading and file-writing operations to generate project documentation (PRDs, CONCEPT.md, and SYNCS.md).\n
- Sanitization: The skill lacks procedures for validating or escaping content derived from external sources before it is transcribed into the final specifications.
Audit Metadata