Another
ReviewAudited by ClawScan on May 10, 2026.
Overview
This skill appears to do what it claims, but it gives an AI broad Android device control, including screenshots, text input, WiFi debugging, and arbitrary ADB shell commands.
Use this skill mainly on test devices or test profiles. Before using it on a personal device, verify the Another desktop app source, avoid sensitive apps or accounts, require confirmation before shell/WiFi/destructive actions, and disconnect or disable WiFi debugging when finished.
Findings (4)
Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.
If used on a personal or production device, an agent could change settings, affect app data, or expand device access beyond the immediate USB session.
The skill exposes raw ADB shell execution and WiFi debugging controls, but does not define safe command boundaries, require confirmation for destructive commands, or explain how WiFi debugging should be disabled after use.
`another_shell` | Run an arbitrary adb shell command and get the output ... `another_wifi_enable` | Switch a USB device to WiFi debugging mode
Use this only with explicit user approval for shell, WiFi, app-changing, or account-affecting actions; prefer allowlisted commands and disable/disconnect WiFi debugging when finished.
The agent may be able to interact with signed-in apps or private device state if those are visible or accessible during the task.
These tools let the agent operate the Android device UI and apps using whatever access is already present on the device. This is aligned with the skill purpose, but it is privileged delegated control.
`another_send_text` | Type text into the focused input field ... `another_launch_app` | Launch an app by package name ... `another_open_url` | Open a URL in the device's default browser
Prefer a test device or separate test profile, avoid sensitive accounts, and supervise actions in apps that can spend money, send messages, or change account data.
Users must independently trust and verify the local Another desktop app/MCP server before granting it device control.
The skill has no reviewed code or install package, and the metadata does not provide a homepage/source for validating the external desktop/MCP component it requires.
Source: unknown; Homepage: none; Install specifications: No install spec — this is an instruction-only skill.
Install the desktop app only from a verified source and confirm the local MCP server is the intended one before connecting devices.
Private screen contents or shell output could enter the agent conversation/context while the device is being controlled.
Screenshots and device outputs flow through a local MCP endpoint to the agent. This is expected for the skill, but the artifact does not describe data handling or privacy boundaries.
`another_take_screenshot` | Capture the screen as a PNG image ... url: http://localhost:7070/mcp
Keep sensitive information off-screen, avoid personal devices when possible, and use a controlled test environment for automation.