gm-orient
Pass
Audited by Gen Agent Trust Hub on Oct 19, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests data from external PRD rows, which serves as a primary ingestion point for potentially untrusted content.
- Ingestion points: Step 1 reads data via the
prd-listtool as specified inSKILL.md. - Boundary markers: The instructions do not use explicit delimiters or specific warnings to ignore embedded instructions when the agent processes the retrieved PRD content or the derived acceptance text.
- Capability inventory: The skill possesses significant capabilities, including codebase searching (
codesearch), gathering symbol insights (codeinsight), writing files to the.gm/receipts/directory, and loading the subsequent skillgm-unknowns. - Sanitization: There is no evidence of validation or sanitization of the content retrieved from the PRD before it influences the search queries or analysis results.
Audit Metadata