cursor_agent_run

THE RISK

High Risk

Why cursor_agent_run needs a policy

This tool executes an external CLI-based agent (cursor-agent) with arbitrary prompts, making the effects dependent on what the prompt instructs the agent to do. While the server offers specialized read/write tools (search, edit), this 'run' variant bypasses those guardrails by accepting free-form prompts, creating risk of unintended file modifications, code execution, or other side effects.

From the tool's definition Tool description states 'Run cursor-agent with a prompt and desired output format' — the verb 'Run' indicates execution of an external agent/operation.

Documented attack patterns abuse exactly the kind of access cursor_agent_run gives an agent:

POLICY

How to control cursor_agent_run

PolicyLayer is an MCP gateway — it sits between your AI agents and Cursor Agent MCP Server, and nothing reaches the server without passing your rules. This is the rule we recommend for cursor_agent_run:

policy.json

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

cursor_agent_run 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.

Create a free account and register Cursor Agent MCP Server — nothing to install.
Add this policy — paste it, or build it visually.
Point your MCP client (Claude, Cursor, anything) at your gateway URL.

RATE-LIMIT THIS TOOL →

Free to start. No card required.

EXPLORE

Related tools and policies

More Cursor Agent MCP Server tools

Execute cursor_agent_chat Chat with cursor-agent using a prompt and optional model/force/output_format. Execute cursor_agent_raw Advanced: provide raw argv array to pass after common flags (e.g., [ Write cursor_agent_edit_file Edit a file with an instruction. Prompt-based wrapper; no CLI subcommand required. Read cursor_agent_analyze_files Analyze one or more paths; optional prompt. Prompt-based wrapper. Read cursor_agent_search_repo Search repository code with include/exclude patterns. Prompt-based wrapper. Read cursor_agent_plan_task Generate a plan for a goal with optional constraints. Prompt-based wrapper.

All 7 Cursor Agent MCP Server tools →

Execute tools on other servers

Android Forensics ADB MCP Server adb_shell_command EMBA-MCP run_emba_scan Frida Game Hacking MCP remove_breakpoint Procmon request_elevation

Go deeper

High-risk tools in Cursor Agent MCP Server → Every high-risk tool this server exposes, in one view.
All high-risk MCP tools → Every high-risk tool across the catalogue, and the policy pattern that contains them.
The MCP Attack Database → Documented attack patterns against MCP deployments — and the policies that stop them.
Agent sandbox → Bounding what an agent can execute and where the boundary should sit.
Least privilege for MCP → Granting agents the minimum tool surface the task needs.
Rate Limiting MCP Tool Calls: A Practical Guide →
MCP Security: Why Prompt Guardrails Aren't Enough →

FAQ

Questions about cursor_agent_run

What does the cursor_agent_run tool do? +

Run cursor-agent with a prompt and desired output format (legacy single-shot). It is categorised as a Execute tool in the Cursor Agent MCP Server MCP Server, which means it can trigger actions or run processes. Use rate limits and argument validation.

How do I enforce a policy on cursor_agent_run? +

Register the Cursor Agent MCP Server MCP server in PolicyLayer and add a rule for cursor_agent_run: 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 Cursor Agent MCP Server. Nothing to install.

What risk level is cursor_agent_run? +

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

Can I rate-limit cursor_agent_run? +

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

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

cursor_agent_run is provided by the Cursor Agent MCP Server MCP server (sailay1996/cursor-agent-mcp). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.

Enforce policy on every Cursor Agent MCP Server tool call.

Start from Cursor Agent MCP Server, add the rest of your stack, and see everything your agents can call. Then put policy on all of it.

CHECK YOUR STACK →