iterative-retrieval
Pass
Audited by Gen Agent Trust Hub on Oct 8, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill defines a coordination protocol for sub-agent management and does not contain malicious patterns.
- [COMMAND_EXECUTION]: Uses
gh issue listfor issue deduplication. This is a legitimate use of the GitHub CLI for project management within the agent's workspace. - [INDIRECT_PROMPT_INJECTION]: The skill defines a prompt template for spawning sub-agents, creating a structured data flow. 1. Ingestion points: The
TaskandSuccess criteriafields in the spawn template (SKILL.md). 2. Boundary markers: Uses clear markdown headers (e.g.,## Task) as delimiters. 3. Capability inventory: Includes GitHub CLI operations and file system writes for recording decisions in.squad/decisions/. 4. Sanitization: Not explicitly defined in the template, but the risk is mitigated by the mandatory coordinator validation checklist and the 3-cycle protocol.
Audit Metadata