run_commands · Claude Code Control MCP: Risk & Policy

WHAT IT DOES

What run_commands does on Claude Code Control MCP

AI agents invoke run_commands to trigger actions in Claude Code Control MCP. 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_commands needs a policy

Shell command execution is a classic Execute category action because effects are entirely argument-dependent and can range from benign queries to system compromise. Given this is part of an autonomous coding agent system ('enables programmatic execution...autonomous file operations'), the blast radius is critical: an attacker or misdirected agent could exfiltrate credentials, install malware, modify system state, or…

From the tool's definition 'Execute one or more shell commands in the working directory' — this tool runs arbitrary shell commands with no restriction mentioned on what operations are permitted.

Attacks that exploit this kind of access

Recommended rule Ready

Verdict

run_commands Rate-limited

run_commands 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_commands": {
      "limits": [
        {
          "counter": "run_commands_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 Claude Code Control MCP, so you can run agents against it safely.

FAQ

Questions about run_commands

What does the run_commands tool do? +

Execute one or more shell commands in the working directory. It is categorised as a Execute tool in the Claude Code Control MCP 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_commands? +

Register the Claude Code Control MCP server in PolicyLayer and add a rule for run_commands: 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 Claude Code Control MCP. Nothing to install.

What risk level is run_commands? +

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

Can I rate-limit run_commands? +

Yes. Add a rate_limit block to the run_commands 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_commands completely? +

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

run_commands is provided by the Claude Code Control MCP server (marc-shade/claude-code-control-mcp). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.