multi-agent-coordinator
Pass
Audited by Gen Agent Trust Hub on Oct 31, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to coordinate multiple agents and process external data such as worker messages, pull request text, and CI logs, which represents an indirect injection surface.
- Ingestion points: The skill reads from worker prompts, mailbox messages, pull request threads, and CI logs as described in SKILL.md.
- Boundary markers: The instructions explicitly define the use of
<EXTERNAL_DATA>delimiters and provide robust instructions to treat retrieved content as untrusted data rather than instructions. - Capability inventory: The skill is capable of performing file writes (logs and handoffs), managing worktrees, and killing processes by process ID (PID).
- Sanitization: The skill includes comprehensive requirements for schema validation, URL scheme validation (specifically rejecting loopback and private network destinations), and capping content size and parsing depth.
- [EXTERNAL_DOWNLOADS]: The skill references an optional external peer-messaging tool,
claude-peers-mcp, located on the author's GitHub repository. This reference is informational and the workflow is explicitly stated to not depend on it.
Audit Metadata