app_stop

What app_stop does on Claude Mobile

AI agents invoke app_stop to trigger actions in Claude Mobile. 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 app_stop is rated High

This tool executes a command on a controlled device (Android via ADB or iOS Simulator via simctl) to terminate an application. While not strictly destructive (the app and its data remain) and not directly harmful to the user's data in most cases, it performs an irreversible operational action in real-time with potential side effects (unsaved work loss, interrupted transactions, state corruption).

From the tool's definition Tool name: 'app_stop'. Description: 'Force stop an app'. The verb 'force stop' indicates termination of a running process on a mobile device, which is an external operation with side effects that cannot be precisely predicted (stopping an app mid-operation…

The rule that runs app_stop safely

PolicyLayer is an MCP gateway: it sits between your AI agents and Claude Mobile, and checks every tool call against a rule you set before the call runs. Nothing changes on the server itself. For app_stop, this is the rule to start with:

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

Questions about app_stop