ptlam-creating-prd
Pass
Audited by Gen Agent Trust Hub on Aug 29, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from external artifacts such as 'grilling records' and 'product briefs' as described in
SKILL.md. These ingestion points provide the source material for the agent to synthesize requirements and write new Markdown files to the local filesystem (e.g.,<project-root>/docs/prd/<slug>.md). This read-then-write capability creates an attack surface where malicious instructions embedded in the input documents could potentially influence the agent's output or file creation activities. The skill attempts to mitigate this through instructions that mandate using only 'confirmed' decisions and evidence, and it includes a readiness check schema to ensure all claims trace back to verified sources. However, it lacks programmatic sanitization or strict input validation beyond these natural language boundaries.
Audit Metadata