Before I connect an agent to an MCP server, I now run through the same short list. I turned it into a repo so anyone can copy it into a pull request or a review ticket.
The problem in one paragraph. An MCP client calls tools/list, and whatever text the server returns, tool descriptions included, goes into the model's context. So a server can hide instructions in a description (tool poisoning), or change a tool after you approved it (a rug pull). The protocol does not make that call authenticated, so you have to check it yourself.
What the checklist covers, in six sections:
- Tool descriptions: read the raw
tools/listoutput, look for instructions aimed at the model and for hidden Unicode. - Changes after approval: hash each approved tool definition and re-check it on every session and on
notifications/tools/list_changed. - Authentication: both listing and calling must require auth. Unauthenticated listing is a finding.
- Supply chain: exact package name, locked dependencies, read the install scripts.
- Execution boundaries: sandbox, filesystem and network allowlists, treat tool output as untrusted input.
- Logging and response: log every call, be able to switch one server off quickly.
Every item is a yes/no question with a quick way to test it. Sections 1 to 3 are blockers, so you end up with a pass or fail instead of a pile of advice. Where a MITRE ATLAS technique exists (tool poisoning, rug pull, supply chain), the item is mapped to it.
Real cases show why patch level matters: CVE-2025-6514 (OS command injection in mcp-remote through an untrusted server, CVSS 9.6) and CVE-2025-32711 (EchoLeak in M365 Copilot, CVSS 9.3).
It is not the first MCP checklist. SlowMist has a good one, and I link it in the README. Mine is built as a review gate rather than a reference guide. If you find an item that is untestable, wrong, or missing, a pull request is welcome.
Repo: https://github.com/sunilgentyala/mcp-security-checklist
Top comments (0)