AI agents invoke cmux_browser_eval to trigger actions in Cmux Agent. 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.
This tool allows arbitrary JavaScript execution in a browser context, which is an Execute operation by definition. The severity is critical because JavaScript can: (1) steal sensitive data from the page, (2) exfiltrate credentials or tokens, (3) modify page content, (4) trigger unwanted navigation, (5) interact with other browser features, or (6) call APIs on behalf of the user.
From the tool's definition Tool name and description explicitly state 'Execute JavaScript in the browser page' — direct code execution capability.
Documented attack patterns abuse exactly the kind of access cmux_browser_eval gives an agent:
PolicyLayer is an MCP gateway — it sits between your AI agents and Cmux Agent, and nothing reaches the server without passing your rules. This is the rule we recommend for cmux_browser_eval:
{
"version": "1",
"default": "deny",
"tools": {
"cmux_browser_eval": {
"limits": [
{
"counter": "cmux_browser_eval_rate",
"window": "minute",
"max": 10,
"scope": "grant"
}
]
}
}
}cmux_browser_eval 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.
Free to start. No card required.
Execute JavaScript in the browser page. It is categorised as a Execute tool in the Cmux Agent MCP Server, which means it can trigger actions or run processes. Use rate limits and argument validation.
Register the Cmux Agent MCP server in PolicyLayer and add a rule for cmux_browser_eval: 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 Cmux Agent. Nothing to install.
cmux_browser_eval is a Execute tool with high risk. Execute tools should be rate-limited and have argument validation enabled.
Yes. Add a rate_limit block to the cmux_browser_eval 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.
Set action: deny in the PolicyLayer policy for cmux_browser_eval. 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.
cmux_browser_eval is provided by the Cmux Agent MCP server (multiagentcognition/cmux-agent-mcp). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
Start from Cmux Agent, add the rest of your stack, and see everything your agents can call. Then put policy on all of it.
Free to start. No card required.
63 Cmux Agent tools catalogued and risk-classified — across an index of 43,000+ MCP servers.