write-a-prd
Pass
Audited by Gen Agent Trust Hub on Apr 17, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill exhibits a surface for indirect prompt injection (Category 8).
- Ingestion points: The agent is instructed to "Explore the repo" and ask for a "detailed description" from the user to understand the codebase state.
- Boundary markers: The instructions lack explicit delimiters or warnings to treat codebase content as data rather than instructions, which could lead the agent to follow malicious commands embedded in code comments or documentation.
- Capability inventory: The skill has the capability to write to the external repository by submitting the generated PRD as a GitHub issue.
- Sanitization: There is no evidence of sanitization or validation logic to filter content extracted from the repo before it is processed or submitted.
Audit Metadata