adb-screen-detection

Pass

Audited by Gen Agent Trust Hub on Mar 17, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection.
  • Ingestion points: Text is ingested from an external Android device screen through the adb-ocr-extract.py script.
  • Boundary markers: The documentation does not specify the use of delimiters or 'ignore embedded instructions' prompts when processing OCR results.
  • Capability inventory: The skill includes the ability to perform physical interactions on the device via adb-tap-coordinate.py based on detected elements, creating a feedback loop where injected instructions on the screen could trigger specific actions.
  • Sanitization: There is no evidence of sanitization or filtering of the text extracted from the screen before it is processed by the agent.
  • [COMMAND_EXECUTION]: The skill executes local Python scripts using uv run and interacts with the system-level adb (Android Debug Bridge) tool. While these capabilities are essential for the skill's stated purpose of Android automation, they provide the agent with direct control over a connected device's interface and file system (e.g., saving screenshots to /tmp/).
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 17, 2026, 08:31 AM