sandbox_run

THE RISK

High Risk

Why sandbox_run needs a policy

This tool executes arbitrary shell commands inside an ephemeral sandbox. Although the sandbox is destroyed afterwards, the shell command itself can have wide-ranging effects (network calls, file exfiltration, etc.) during its execution. The core action is executing code/commands, making Execute the correct category.

From the tool's definition run a shell command, return the output, and destroy the sandbox

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

POLICY

How to control sandbox_run

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

policy.json

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

sandbox_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 Microsandbox — 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 Microsandbox tools

Destructive sandbox_fs_remove Remove a file or directory inside the sandbox. Destructive sandbox_remove Remove a sandbox. Must be stopped unless force is true. Destructive volume_remove Remove a named volume. Execute sandbox_exec Execute a command inside a running sandbox and return stdout/stderr/exit code. Execute sandbox_shell Execute a shell command or script string inside a running sandbox. Execute sandbox_stop Stop a running sandbox. Use force to kill with SIGKILL instead of graceful SIGTERM. Execute sandbox_create Create and boot a persistent named sandbox with full configuration. Write sandbox_fs_mkdir Create a directory with parent directories inside the sandbox.

All 19 Microsandbox 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 Microsandbox → 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 sandbox_run

What does the sandbox_run tool do? +

Create an ephemeral sandbox, run a shell command, return the output, and destroy the sandbox. It is categorised as a Execute tool in the Microsandbox MCP Server, which means it can trigger actions or run processes. Use rate limits and argument validation.

How do I enforce a policy on sandbox_run? +

Register the Microsandbox MCP server in PolicyLayer and add a rule for sandbox_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 Microsandbox. Nothing to install.

What risk level is sandbox_run? +

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

Can I rate-limit sandbox_run? +

Yes. Add a rate_limit block to the sandbox_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 sandbox_run completely? +

Set action: deny in the PolicyLayer policy for sandbox_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 sandbox_run? +

sandbox_run is provided by the Microsandbox MCP server (superradcompany/microsandbox-mcp). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.

Enforce policy on every Microsandbox tool call.

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

CHECK YOUR STACK →