swarm
Pass
Audited by Gen Agent Trust Hub on Oct 25, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill aggregates results from multiple workers to generate a consolidated report, creating a surface for potential injection from worker outputs.
- Ingestion points: Processes results from workers spawned via delegate_task and t3_thread_launch in Phase C.
- Boundary markers: No explicit markers are used to separate worker content from orchestrator instructions.
- Capability inventory: Can spawn further sub-agents, perform git operations (commit, worktree), and write to the local filesystem.
- Sanitization: No sanitization of worker-reported text is performed before processing.
- [DYNAMIC_EXECUTION]: The skill dynamically resolves and follows instructions from files in sibling directories (e.g., ..//SKILL.md) based on names mentioned in the prompt, facilitating multi-skill workflows.
Audit Metadata