run_colab_shell · Colab MCP: Risk & Policy

WHAT IT DOES

What run_colab_shell does on Colab MCP

AI agents invoke run_colab_shell to trigger actions in Colab 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_colab_shell needs a policy

Executing shell commands on a remote system (Google Colab) is an Execute-category action because it triggers external operations whose effects entirely depend on the arguments provided. Shell execution has critical severity due to the unrestricted blast radius—an AI agent could invoke destructive commands (rm -rf /), exfiltrate data, install malware, or compromise the Colab environment.

From the tool's definition Server description states the tool 'supports executing shell commands' on Google Colab instances. The tool name 'run_colab_shell' combined with the server's documented capability to execute shell commands indicates this tool runs arbitrary shell code on a…

Attacks that exploit this kind of access

Recommended rule Ready

Verdict

run_colab_shell Rate-limited

run_colab_shell 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_colab_shell": {
      "limits": [
        {
          "counter": "run_colab_shell_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 Colab MCP, so you can run agents against it safely.

FAQ

Questions about run_colab_shell

What does the run_colab_shell tool do? +

run_colab_shell. It is categorised as a Execute tool in the Colab 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_colab_shell? +

Register the Colab MCP server in PolicyLayer and add a rule for run_colab_shell: 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 Colab MCP. Nothing to install.

What risk level is run_colab_shell? +

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

Can I rate-limit run_colab_shell? +

Yes. Add a rate_limit block to the run_colab_shell 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_colab_shell completely? +

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

run_colab_shell is provided by the Colab MCP server (kumardev7/colab-mcp). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.