reviewing
Pass
Audited by Gen Agent Trust Hub on Oct 8, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted content from research, plans, and brainstorm documents, which may contain malicious instructions designed to influence the agent's behavior.\n
- Ingestion points: Input documents are read and classified based on their content and filesystem path (e.g.,
/research/,/plans/,/brainstorms/,/qa/) as described in Step 1.\n - Boundary markers: Absent. The instructions do not specify delimiters or markers to isolate document content from agent logic.\n
- Capability inventory: The skill can modify documents (Step 6), write new files to the
thoughts/*/reviews/directory, and spawn acodebase-analyzeragent (Step 4).\n - Sanitization: Absent. No explicit sanitization or validation steps are mentioned for the content being read or for the automated fixes being applied.
Audit Metadata