collabute-mcp
Pass
Audited by Gen Agent Trust Hub on Feb 23, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill facilitates data retrieval from organizational tools like Slack, Linear, and Vercel through a Model Context Protocol (MCP) server, operating within a defined and controlled environment.- [SAFE]: Authentication follows an OAuth-first approach, and the skill includes instructions to avoid static PAT headers, ensuring secure credential handling.- [SAFE]: The skill enforces a 'proposal-only' security model for write intents (e.g.,
meeting.propose_task_from_meeting), which requires manual user approval before any external state changes occur, significantly mitigating the risk of automated unauthorized actions.- [SAFE]: Data processing instructions include pagination requirements (e.g.,meeting.get_transcript) to prevent excessive data exposure in single responses and ensure stable performance.
Audit Metadata