codex · Codex MCP: Risk & Policy

WHAT IT DOES

What codex does on Codex

AI agents invoke codex to trigger actions in Codex. What it does depends on the arguments the agent supplies, and its effects often reach beyond the immediate call — builds kicked off, notifications sent, workflows started.

WHY IT NEEDS A POLICY

Why codex needs a policy

The server description references CLI execution, tmux persistence, git worktree isolation, and async dispatch — all hallmarks of code execution infrastructure. The tool named 'codex' on such a server is almost certainly the primary entry point for dispatching Codex CLI commands, which can run arbitrary code. However, the tool description is empty, lowering confidence.

From the tool's definition Tool name is 'codex' on a server described as 'MCP server for Codex CLI — tmux persistence, git worktree isolation, async dispatch.' Sibling tools include codex_dispatch, codex_cancel, codex_status, suggesting this is the primary invocation tool for Codex CLI…

Attacks that exploit this kind of access

Recommended rule Ready

Verdict

codex Rate-limited

codex stays usable, but rate-capped — a runaway agent can't fire it dozens of times a minute. Everything else on the server is denied unless you say otherwise.

View as policy code

policy.json

{
  "version": "1",
  "default": "deny",
  "tools": {
    "codex": {
      "limits": [
        {
          "counter": "codex_rate",
          "window": "minute",
          "max": 10,
          "scope": "grant"
        }
      ]
    }
  }
}

RATE-LIMIT THIS TOOL →

Instant setup, no code required.

PolicyLayer is an MCP gateway: every tool call is checked against your rule before it reaches Codex, so you can run agents against it safely.

FAQ

Questions about codex

What does the codex tool do? +

codex. It is categorised as a Execute tool in the Codex MCP Server, which means it can trigger actions or run processes. Use rate limits and argument validation.

How do I enforce a policy on codex? +

Register the Codex MCP server in PolicyLayer and add a rule for codex: allow, deny, rate-limit, or require approval. Point your MCP client at the PolicyLayer proxy URL and the rule is enforced on every call, before it reaches Codex. Nothing to install.

What risk level is codex? +

codex is a Execute tool with high risk. Execute tools should be rate-limited and have argument validation enabled.

Can I rate-limit codex? +

Yes. Add a rate_limit block to the codex rule in your PolicyLayer policy. For example, setting max: 10 and window: 60 limits the tool to 10 calls per minute. Rate limits are tracked per agent session and reset automatically.

How do I block codex completely? +

Set action: deny in the PolicyLayer policy for codex. The AI agent will receive a policy violation error and cannot call the tool. You can also include a reason field to explain why the tool is blocked.

What MCP server provides codex? +

codex is provided by the Codex MCP server (shilong20/codexmcp). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.