analyze-mobile-app
Warn
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: MEDIUMCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION] (MEDIUM): The skill relies on executing adb shell commands to interact with a mobile device. This includes simulating user input (tap, text) and reading system properties. While necessary for the skill's function, these commands allow the agent to perform sensitive actions on the connected device.
- [PROMPT_INJECTION] (LOW): Surface for Indirect Prompt Injection. 1. Ingestion points: Extracts UI hierarchy and text from external apps via adb shell uiautomator dump. 2. Boundary markers: Absent; the agent does not distinguish between app data and instructions. 3. Capability inventory: Can simulate device input and write files to the local disk. 4. Sanitization: Absent; UI text is processed without validation, potentially allowing a malicious app to influence the agent's logic.
Audit Metadata