run_code · Vending Machine MCP: Risk & Policy

WHAT IT DOES

What run_code does on Vending Machine

AI agents invoke run_code to trigger actions in Vending Machine. 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 run_code needs a policy

Code execution tools are classified as Execute because their effects are contingent on the code arguments provided. Although sandboxed, this does not eliminate risk; a malicious prompt could direct the agent to run code that exfiltrates data, probes the network, or performs reconnaissance.

From the tool's definition Tool description states it 'Run[s] code in a secure E2B sandbox' — the verb 'run' directly indicates arbitrary code execution. The return of stdout, stderr, and errors confirms the tool executes user-supplied or agent-directed code and surfaces its output.

Attacks that exploit this kind of access

Recommended rule Ready

Verdict

run_code Rate-limited

run_code 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": {
    "run_code": {
      "limits": [
        {
          "counter": "run_code_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 Vending Machine, so you can run agents against it safely.

FAQ

Questions about run_code

What does the run_code tool do? +

Run code in a secure E2B sandbox. Returns stdout, stderr, and any errors. It is categorised as a Execute tool in the Vending Machine MCP Server, which means it can trigger actions or run processes. Use rate limits and argument validation.

How do I enforce a policy on run_code? +

Register the Vending Machine MCP server in PolicyLayer and add a rule for run_code: 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 Vending Machine. Nothing to install.

What risk level is run_code? +

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

Can I rate-limit run_code? +

Yes. Add a rate_limit block to the run_code 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 run_code completely? +

Set action: deny in the PolicyLayer policy for run_code. 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 run_code? +

run_code is provided by the Vending Machine MCP server (yokiidesu/vending-machine-mcp). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.