agent-collaboration
Pass
Audited by Gen Agent Trust Hub on Aug 23, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill defines behavioral guidelines for routing work to different subagents (Grok, Pi, Herdr) without including any executable code or dangerous tool invocations.
- [SAFE]: No evidence of hardcoded credentials, sensitive file path access, or unauthorized network activity was found.
- [INDIRECT_PROMPT_INJECTION]: While the skill facilitates multi-agent communication which is a potential surface for indirect injection, it includes significant mitigations by instructing the agent to use isolated worktrees for parallel writers, grant read-only authority to reviewers, and ensure the top-level owner verifies all external effects directly before completion.
- [SAFE]: No obfuscation, persistence mechanisms, or privilege escalation techniques were identified in the instructions or metadata.
Audit Metadata