moai-ref-cross-model-audit
Pass
Audited by Gen Agent Trust Hub on Aug 14, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill is a documentation reference for coordinating multiple AI backends (Claude, Codex, GLM) through a single MCP tool called
mcp__moai__audit_multi. - [SAFE]: The tool is defined as part of the 'moai' vendor's own MCP server, which aligns with the provided author context (modu-ai/moai).
- [SAFE]: No remote code execution, dangerous command patterns, or unauthorized data exfiltration were found. The tool simply passes structured verdict data between internal AI agents and enforces an 'Independence rule' to ensure separate models do not influence each other's reasoning.
- [SAFE]: References to external backends like Codex and GLM (z.ai) are treated as internal system integrations for the auditor's functionality and do not involve untrusted code execution or exfiltration to third-party domains outside the vendor's scope.
Audit Metadata