# act

Execute a natural language action. The AI will plan and perform multi-step operations in a single invocation, useful for transient UI interactions (e.g., Spotlight, dropdown menus) that disappear between separate commands.

Agent View of the PolicyLayer registry record for `act`. HTML page: https://policylayer.com/tools/android/act

## Facts

- Tool: `act`
- Server: Android (`@midscene/android-mcp`) — https://policylayer.com/tools/android.md
- Install: `npx -y @midscene/android-mcp`
- Homepage: https://www.npmjs.com/package/@midscene/android-mcp
- Risk category: Execute (High risk)
- Registry record: grade F, identity unverified
- Server rate-limited: no
- Parameters: 9 (1 required)
- Recommended policy verdict: Rate-limited

## Parameters

| Parameter | Type | Required | Description |
| --- | --- | --- | --- |
| `prompt` | string | yes | Natural language description of the action to perform, e.g. "press Command+Space, type Safari, press Enter" |
| `deepThink` | boolean | no | Plan this action with deep thinking (richer context and sub-goal decomposition). Helps with complex multi-step instructions at the cost of speed. Defaults to th |
| `deepLocate` | boolean | no | Use deep locate for every element this action targets. Improves precision for small or ambiguous targets at the cost of speed. Defaults to the server --deep-loc |
| `android.deviceId` | string | no | Android device ID (from adb devices) |
| `android.useScrcpy` | boolean | no | Enable scrcpy accelerated screenshots |
| `android.aiActContext` | string | no | Background knowledge passed to aiAct. Default: no extra context. |
| `android.waitAfterAction` | number | no | Wait time in milliseconds after each action execution. Default: 300ms. |
| `android.replanningCycleLimit` | integer | no | Maximum number of replanning cycles for aiAct. Default: model adapter default. |
| `android.screenshotShrinkFactor` | number | no | Screenshot shrink factor before sending images to AI. Default: 1; high values may reduce recognition quality, especially on mobile. |

Parameters from the server's own tool schema.

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

```json
{
  "jsonrpc": "2.0",
  "id": 1,
  "method": "tools/call",
  "params": {
    "name": "act",
    "arguments": {
      "prompt": "<prompt>"
    }
  }
}
```

## Why act is rated High

This tool triggers execution of multi-step automated operations on an Android device based on natural language prompts. While not destructive by itself, it can perform any sequence of device interactions (opening apps, modifying settings, sending messages, etc.) whose actual effects depend on how the natural language is interpreted.

From the tool's own definition: "Tool description states it 'Execute[s] a natural language action' and 'perform[s] multi-step operations', enabling the AI to control Android device interactions autonomously."

## Use case

AI agents invoke act to trigger actions in Android. 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 Android:

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

## Other tools on Android (22)

- `ClearInput` — Destructive — https://policylayer.com/tools/android/clearinput.md
- `android_connect` — Execute — https://policylayer.com/tools/android/android-connect.md
- `AndroidBackButton` — Execute — https://policylayer.com/tools/android/androidbackbutton.md
- `AndroidHomeButton` — Execute — https://policylayer.com/tools/android/androidhomebutton.md
- `AndroidRecentAppsButton` — Execute — https://policylayer.com/tools/android/androidrecentappsbutton.md
- `CursorMove` — Execute — https://policylayer.com/tools/android/cursormove.md
- `DoubleClick` — Execute — https://policylayer.com/tools/android/doubleclick.md
- `DragAndDrop` — Execute — https://policylayer.com/tools/android/draganddrop.md
- `KeyboardPress` — Execute — https://policylayer.com/tools/android/keyboardpress.md
- `Launch` — Execute — https://policylayer.com/tools/android/launch.md
- `LongPress` — Execute — https://policylayer.com/tools/android/longpress.md
- `Pinch` — Execute — https://policylayer.com/tools/android/pinch.md
- `PullGesture` — Execute — https://policylayer.com/tools/android/pullgesture.md
- `RunAdbShell` — Execute — https://policylayer.com/tools/android/runadbshell.md
- `Scroll` — Execute — https://policylayer.com/tools/android/scroll.md
- `Swipe` — Execute — https://policylayer.com/tools/android/swipe.md
- `Tap` — Execute — https://policylayer.com/tools/android/tap.md
- `Terminate` — Execute — https://policylayer.com/tools/android/terminate.md
- `take_screenshot` — Read — https://policylayer.com/tools/android/take-screenshot.md
- `android_disconnect` — Write — https://policylayer.com/tools/android/android-disconnect.md
- `assert` — Write — https://policylayer.com/tools/android/assert.md
- `Input` — Write — https://policylayer.com/tools/android/input.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=android · API: https://policylayer.com/registry/api · Policy library: https://policylayer.com/policies/android