do-create-prd
Pass
Audited by Gen Agent Trust Hub on Oct 27, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill is strictly limited to document generation and explicitly prohibits the execution of application code, tests, or servers.
- [INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied feature requests and web search results to generate a structured PRD. While this creates a potential attack surface for indirect prompt injection, the risk is minimized by the skill's narrow scope and its restriction to outputting markdown documents.
- Ingestion points: User input (Step 1), clarification answers in prd-answers.md (Step 2), and web search results for business rules (Step 3).
- Boundary markers: Absent.
- Capability inventory: File system read/write access (template reading and PRD writing) and tool invocations (AskUserQuestion, WebSearch, TaskUpdate).
- Sanitization: Absent; the skill relies on the LLM to process and format external data without specific sanitization functions.
Audit Metadata