army-fanin
Pass
Audited by Gen Agent Trust Hub on Sep 20, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to process data produced by other agents (worktrees, diffs, and test results). This represents a surface for indirect prompt injection where a compromised agent could attempt to influence the fan-in process through its output.
- Ingestion points: Parallel worktrees, diffs, and test result logs provided by the fleet.
- Boundary markers: The instructions do not define specific delimiter boundaries for the agent output.
- Capability inventory: Git worktree management (list, reap), branch archiving, and writing manifest files.
- Sanitization: The skill includes inherent security best practices by requiring verifiable test results and known provenance before any change is selected for integration, reducing the risk of accidental or malicious code execution during the fan-in process.
Audit Metadata