icp-refine
Pass
Audited by Gen Agent Trust Hub on Feb 27, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill uses internal tools (e.g., list_events, list_findings) to gather data for analysis, which is consistent with its stated purpose of ICP refinement.
- [SAFE]: Write operations (update_entity, create_entity) are only performed after a human-in-the-loop confirmation step, where the user reviews the recommended updates before they are applied.
- [SAFE]: No external network requests, obfuscation, or unauthorized command execution patterns were identified.
- [SAFE]: The skill manages internal vendor resources (segments, personas) using standard Model Context Protocol (MCP) tool patterns.
Audit Metadata