prd-quality-gate
Pass
Audited by Gen Agent Trust Hub on Jul 20, 2026
Risk Level: SAFEPROMPT_INJECTIONNO_CODE
Full Analysis
- [PROMPT_INJECTION]: Indirect Prompt Injection Surface. The skill ingests untrusted text from external PRDs or issue bodies to validate their structure. An attacker could embed instructions within these documents (e.g., inside an 'Acceptance Criteria' bullet) to attempt to bypass the validation gate or influence subsequent agents in a planning pipeline.
- Ingestion points: PRD content or issue bodies (SKILL.md).
- Boundary markers: None. The skill treats the document content directly as the target for validation without explicit delimiters for the untrusted data.
- Capability inventory: No subprocess execution, file system access, or network operations are present in the skill definition.
- Sanitization: No sanitization or escaping of the PRD content is specified before the agent processes it.
- [NO_CODE]: The skill contains no executable scripts or binaries. It relies entirely on natural language instructions and regex patterns within the markdown file to perform its validation tasks.
Audit Metadata