product-requirements
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill reads external project files (README, package.json, pyproject.toml) to gather context. This is a standard surface for indirect prompt injection if the skill processes an untrusted repository.
- Ingestion points: Interaction Step 1 involves reading project metadata files.
- Boundary markers: None present to delimit ingested content.
- Capability inventory: Writing Product Requirements Documents to the docs/ directory.
- Sanitization: None present.
Audit Metadata