# run_in_sandbox

Run a shell command inside the sandbox container. timeout=0 uses default timeout.

Agent View of the PolicyLayer registry record for `run_in_sandbox`. HTML page: https://policylayer.com/tools/agent-jail-mcp/run-in-sandbox

## Facts

- Tool: `run_in_sandbox`
- Server: Agent Jail (`idiscoord13-sketch/agent-jail-mcp`) — https://policylayer.com/tools/agent-jail-mcp.md
- Homepage: https://github.com/idiscoord13-sketch/agent-jail-mcp
- Risk category: Execute (High risk)
- Registry record: grade F, identity unverified
- Server rate-limited: no
- Parameters: 0
- Recommended policy verdict: Rate-limited

## Example call (MCP tools/call, JSON-RPC 2.0)

```json
{
  "jsonrpc": "2.0",
  "id": 1,
  "method": "tools/call",
  "params": {
    "name": "run_in_sandbox",
    "arguments": {}
  }
}
```

## Why run_in_sandbox is rated High

run_in_sandbox triggers real processes with real consequences. An agent gone sideways doesn't fire it once. It starts dozens of builds, sends mass notifications, or burns through compute before anyone looks up.

## Use case

AI agents invoke run_in_sandbox to trigger actions in Agent Jail. 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.

## Recommended policy (PolicyLayer)

Verdict: **Rate-limited**. Enforced by the PolicyLayer MCP gateway (https://policylayer.com/mcp-gateway) before a call reaches Agent Jail:

```json
{
  "version": "1",
  "default": "deny",
  "tools": {
    "run_in_sandbox": {
      "limits": [
        {
          "counter": "run_in_sandbox_rate",
          "window": "minute",
          "max": 10,
          "scope": "grant"
        }
      ]
    }
  }
}
```

## Other tools on Agent Jail (34)

- `cleanup_expired_sandboxes` — Destructive — https://policylayer.com/tools/agent-jail-mcp/cleanup-expired-sandboxes.md
- `destroy_sandbox` — Destructive — https://policylayer.com/tools/agent-jail-mcp/destroy-sandbox.md
- `run_query` — Execute — https://policylayer.com/tools/agent-jail-mcp/run-query.md
- `schedule_task` — Execute — https://policylayer.com/tools/agent-jail-mcp/schedule-task.md
- `add` — Other — https://policylayer.com/tools/agent-jail-mcp/add.md
- `async_tool` — Other — https://policylayer.com/tools/agent-jail-mcp/async-tool.md
- `custom_name` — Other — https://policylayer.com/tools/agent-jail-mcp/custom-name.md
- `greet` — Other — https://policylayer.com/tools/agent-jail-mcp/greet.md
- `log_progress` — Other — https://policylayer.com/tools/agent-jail-mcp/log-progress.md
- `multiply` — Other — https://policylayer.com/tools/agent-jail-mcp/multiply.md
- `my_task` — Other — https://policylayer.com/tools/agent-jail-mcp/my-task.md
- `my_tool` — Other — https://policylayer.com/tools/agent-jail-mcp/my-tool.md
- `tool_with_context` — Other — https://policylayer.com/tools/agent-jail-mcp/tool-with-context.md
- `check_worker_status` — Read — https://policylayer.com/tools/agent-jail-mcp/check-worker-status.md
- `docker_health` — Read — https://policylayer.com/tools/agent-jail-mcp/docker-health.md
- `echo_tool` — Read — https://policylayer.com/tools/agent-jail-mcp/echo-tool.md
- `get_auth_type` — Read — https://policylayer.com/tools/agent-jail-mcp/get-auth-type.md
- `get_client_ip` — Read — https://policylayer.com/tools/agent-jail-mcp/get-client-ip.md
- `get_info` — Read — https://policylayer.com/tools/agent-jail-mcp/get-info.md
- `get_my_emails` — Read — https://policylayer.com/tools/agent-jail-mcp/get-my-emails.md
- `get_user_id` — Read — https://policylayer.com/tools/agent-jail-mcp/get-user-id.md
- `greeting` — Read — https://policylayer.com/tools/agent-jail-mcp/greeting.md
- `introspect` — Read — https://policylayer.com/tools/agent-jail-mcp/introspect.md
- `list_files` — Read — https://policylayer.com/tools/agent-jail-mcp/list-files.md
- `list_sandboxes` — Read — https://policylayer.com/tools/agent-jail-mcp/list-sandboxes.md
- `long_running_task` — Read — https://policylayer.com/tools/agent-jail-mcp/long-running-task.md
- `read_file` — Read — https://policylayer.com/tools/agent-jail-mcp/read-file.md
- `add_expense` — Write — https://policylayer.com/tools/agent-jail-mcp/add-expense.md
- `create_sandbox` — Write — https://policylayer.com/tools/agent-jail-mcp/create-sandbox.md
- `save` — Write — https://policylayer.com/tools/agent-jail-mcp/save.md
- …and 4 more: https://policylayer.com/tools/agent-jail-mcp.md

## For agents

This record is a snapshot. Live verdicts and the full registry:

- Check every server in your MCP config at once: `npx -y policylayer stack`
- Vet a server before you add it: install the mcp-precheck skill — `npx skills add https://policylayer.com` (skill text: https://policylayer.com/skill.md)
- Query the registry over MCP: endpoint `https://api.policylayer.com/mcp` — tools `check_mcp_server`, `check_mcp_stack`, `check_tool`, `search_registry`, `get_change_events`

---

Source: the PolicyLayer MCP registry — one continuously verified record per MCP server. Full record: https://policylayer.com/registry?q=agent-jail-mcp · API: https://policylayer.com/registry/api · Policy library: https://policylayer.com/policies/agent-jail-mcp