multi-agent-coordinator
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill manages a fleet of agents by maintaining an internal state in a dedicated directory (
~/.openclaw/skill-state/). File access is restricted to this specific path, and no sensitive system directories or credentials are accessed. - [SAFE]: The Python implementation uses the
yaml.safe_loadmethod for parsing the agent registry and coordination state, which effectively prevents YAML-based deserialization attacks. - [SAFE]: Analysis of the execution logic in
run.pyconfirms that the skill does not use subprocesses, network operations, or dynamic code execution. All actions are confined to managing the local state file and providing status reports. - [SAFE]: The documentation and metadata do not contain any prompt injection patterns or attempts to override agent safety guidelines.
Audit Metadata