// RESEARCH
MCP security research
Original MCP security research, built from PolicyLayer's continuously-updated catalogue of 31,000 classified tools across 2,031 public MCP servers. Each audit measures what Model Context Protocol servers can actually do to the systems behind them — which can destroy data, execute commands, or move money — and names the servers carrying the most risk. Updated monthly.
Past editions
In progress
- State of MCP Approval Gates — how much destructive surface sits behind any approval logic.
- State of MCP OAuth — coverage of OAuth and RFC 8707 Resource Indicators across the ecosystem.
Let agents act without letting them run wild.
Deterministic policy on every MCP tool call. Per-identity grants. Full audit log.
Free to start. No card required.
4,600+ MCP servers and 31,000+ tools scanned and risk-classified.