MCP security
MCP is how AI clients connect to tools. The protocol leaves four gaps open: nothing checks the inputs, nothing requires an audit log, nothing limits call rates, and each server's author decides how auth works. We scanned the ecosystem and wrote up each risk and what reduces it.
We classified 8,216 MCP servers by auth appropriateness (a static scan of three registries, March 2026 snapshot). We also regex-scanned 7,522 agent skills, every skill on one registry (ClawHub), for credential leaks and injection patterns. About 61% of the raw hits were estimated false positives, and 16 threats were confirmed.
The six structural risks are in the MCP Security series. The checklist post below is the practical starting point. A gateway in front of your MCP servers can log, rate-limit and authenticate every call, and scan arguments for PII. What an MCP control plane is explains how.
Posts on MCP security
-
The MCP Security Checklist for Enterprise Teams
A 10-point security checklist for teams running MCP servers in production: identity, access control, PII, rate limits, audit logs, approvals and more, with what ACP covers for each.
-
MCP Security Vulnerabilities: 4 Tool Schema Patterns Behind Real CVEs
Real MCP server CVEs trace back to tool schemas that accept any string. The four patterns behind path traversal, SSRF and injection, and the schema fix for each.
-
MCP Server Audit Logging: I Checked 8,216 Servers. 74.9% Don't Mention It.
In a March 2026 scan, 74.9% of 8,216 MCP servers had no mention of audit logging, and the MCP spec defines no audit format. Here's why that matters for compliance, and where the log can live instead.
-
MCP Gateway Comparison (2026): Composio vs ACP vs DIY
Three ways to put access control, logging and PII checks in front of MCP servers: build it yourself, use Composio, or use ACP. What each gives you and what it leaves to you.
-
I Audited 7,522 AI Agent Skills. Here's What I Found.
A static scan of every skill on ClawHub in March 2026: how many leak credentials, how many carry prompt injection, and what the registry's own moderation catches.
-
MCP Security: Who Decides What an Agent Can Call?
MCP defines how agents call tools. It doesn't decide which calls are allowed or keep an audit log. What the protocol covers, what it leaves to you, and what a control plane adds.
-
How to Rate-Limit an MCP Server (Per-User, Per-Tool, Per-Agent)
An MCP server sees the LLM runtime's shared credential, not the person behind it. How to limit calls per user, per tool and per agent, and what ACP does for you today.
-
MCP Re-Auth: What ChatGPT Needs When Tokens Expire
As of March 2026, ChatGPT does not restart OAuth on an HTTP 401 or a JSON-RPC error. It needs a JSON-RPC success response with _meta. The signal, and how a gateway sends it.
Related: tool permissions & policy · the MCP Security series · all posts