puppeteer_evaluate · Steel Puppeteer MCP: Risk & Policy

WHAT IT DOES

What puppeteer_evaluate does on Steel Puppeteer

AI agents invoke puppeteer_evaluate to trigger actions in Steel Puppeteer. 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 puppeteer_evaluate needs a policy

JavaScript execution in a browser is a classic Execute category risk. The tool's purpose is to run code whose effects are entirely dependent on the JavaScript passed as arguments.

From the tool's definition The tool description explicitly states "Execute JavaScript in the browser console", which permits arbitrary code execution within a real browser environment.

Attacks that exploit this kind of access

Recommended rule Ready

Verdict

puppeteer_evaluate Rate-limited

puppeteer_evaluate 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": {
    "puppeteer_evaluate": {
      "limits": [
        {
          "counter": "puppeteer_evaluate_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 Steel Puppeteer, so you can run agents against it safely.

FAQ

Questions about puppeteer_evaluate

What does the puppeteer_evaluate tool do? +

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

How do I enforce a policy on puppeteer_evaluate? +

Register the Steel Puppeteer MCP server in PolicyLayer and add a rule for puppeteer_evaluate: 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 Steel Puppeteer. Nothing to install.

What risk level is puppeteer_evaluate? +

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

Can I rate-limit puppeteer_evaluate? +

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

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

puppeteer_evaluate is provided by the Steel Puppeteer MCP server (rdvo/mcp-server). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.