---
title: "Discussion on \"Twenty Security Checks Before Your MCP Server Goes Public\" | Hashnode"
description: "Discussion on \"Twenty Security Checks Before Your MCP Server Goes Public\". An MCP server is effectively an API that agents call for people, and that gives it an unusual set of risks. The caller is a model that text from elsewhere can steer, the human may never see individual"
url: "https://hashnode.com/posts/twenty-security-checks-before-your-mcp-server-goes-public/6ad155a64c89af0468c46f61"
---

[Feed](https://hashnode.com/feed) /

Discussion

# [Twenty Security Checks Before Your MCP Server Goes Public](https://built2win.hashnode.dev/twenty-security-checks-before-your-mcp-server-goes-public?utm_source=hashnode&utm_medium=feed)

An MCP server is effectively an API that agents call for people, and that gives it an unusual set of risks. The caller is a model that text from elsewhere can steer, the human may never see individual

Oct 15 · 2 min read · built2win.hashnode.dev  
Read the article

-   [#mcp](https://hashnode.com/tag/mcp)
-   [#security](https://hashnode.com/tag/security)
-   [#ai-agents](https://hashnode.com/tag/ai-agents)
-   [#api-security](https://hashnode.com/tag/api-security)
-   [#checklist](https://hashnode.com/tag/checklist)

## Written by

[Jacob Campbell](https://hashnode.com/@built2win)

@built2win · Austin, USA

  
[View profile](https://hashnode.com/@built2win)

The Built2Win team wants to teach people how to compete on Google for clicks. Check out our installations!

100% custom PHP code from scratch.

## Responses 3

Markdown supported · ⌘↵ to post

[Cai](https://hashnode.com/@cailab)

[Oct 16](https://hashnode.com/posts/twenty-security-checks-before-your-mcp-server-goes-public/6ad155a64c89af0468c46f61/comment/6ad1da80927ef2e6acd61a45)

Solid list, especially the rate limiting and resource path traversal points. One area that could use more attention: provenance verification. When you discover an MCP server in a directory, how do you verify it was published by who it claims and hasnt been tampered with? For production agent workflows this is a real blind spot code signing plus identity attestation at the discovery layer would close it.

[Mihai\_LeanZero](https://hashnode.com/@mihaileanzero)

[Oct 16](https://hashnode.com/posts/twenty-security-checks-before-your-mcp-server-goes-public/6ad155a64c89af0468c46f61/comment/6ad1bb3a0c5c64ab3d9e8f1a)

Good checklist. On keeping tool descriptions static, I'd also pin them on the client side: hash each tool's name and description when someone approves the server, and re-prompt when the hash changes. Otherwise "static" only protects you from servers that behave. Same logic for annotations like readOnlyHint, the spec treats them as hints, so the real enforcement has to be your own scopes and server-side checks, which you already say in the last group.

[Yunus Emre Vurgun](https://hashnode.com/@yunusemrevurgun)

[Oct 16](https://hashnode.com/posts/twenty-security-checks-before-your-mcp-server-goes-public/6ad155a64c89af0468c46f61/comment/6ad1b4c9fc1b3ef701e5fd1e)

Keeping tool descriptions static is the check I see skipped most, and it's easy to enforce: hash the tool list at release time and have a CI test fail if the served descriptions or schemas ever differ from the pinned hash. On rate limiting, limiting per tool and per IP still leaves room for one agent session to fan out across many cheap read calls, so a per-session budget on total calls helps too. And for "return only what the output schema promises", validating responses against the schema on the server before sending catches accidental leaks of internal fields when the underlying data model grows.
