This is a paid remote service that enforces gating and audit trails when Claude Code modifies agent skills. It exposes five tools: check_skill_update_gate to validate changes before they're applied, compare_skill_manifest to diff versions, issue_update_receipt for tamper-proof records, assign_skill_owner for access control, and export_skill_update_log for compliance reporting. Reach for this when you need reviewer-ready evidence of what an AI agent changed in your codebase, or when you're building workflows where skill modifications require approval chains. Authentication is bearer token over streamable HTTP. The repository is documentation only, the actual service runs at agentskillupdategate.clauxel.com.
claude mcp add --transport http clauxel-agent-skill-update-gate-mcp https://agentskillupdategate.clauxel.com/mcp