write-a-prd
Pass
Audited by Gen Agent Trust Hub on Oct 31, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill exhibits an attack surface for indirect prompt injection by ingesting and processing untrusted data from both the user and the local codebase to generate a document that is then submitted to an external platform (GitHub).
- Ingestion points: User-provided problem descriptions and interview responses (Steps 1 and 3 in
SKILL.md) and codebase content discovered during exploration (Step 2 inSKILL.md). - Boundary markers: The skill lacks explicit boundary markers or instructions to the agent to disregard potentially malicious instructions embedded within the user's description or the codebase files.
- Capability inventory: The agent is authorized to explore the local repository and submit content to GitHub as an issue.
- Sanitization: There are no instructions for escaping, validating, or filtering the content ingested from the user or the codebase before it is interpolated into the final PRD template.
Audit Metadata