to-prd
Pass
Audited by Gen Agent Trust Hub on Sep 24, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes local bash scripts
scripts/next-prd-index.shandscripts/publish-prd.sh. The latter script interacts with thegh(GitHub) CLI to create issues. These operations are restricted to project management tasks and use proper argument quoting to mitigate shell injection risks. - [INDIRECT_PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface as it ingests and processes untrusted conversation history and local documentation.
- Ingestion points: Reads repository source code, Architectural Decision Records (ADRs) from
docs/adr/, and design notes fromdocs/intent/. - Boundary markers: Synthesizes output using a structured markdown template (
PRD-TEMPLATE.md). - Capability inventory: Can write files to the local repository (
docs/prds/) and create GitHub issues using the authenticated user's GitHub CLI. - Sanitization: No explicit sanitization of input data is performed, though the skill optionally utilizes a
humanizerskill to refine the generated text body.
Audit Metadata