adb

Pass

Audited by Gen Agent Trust Hub on Mar 17, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill is centered around executing bash scripts (located in the scripts/ directory) and raw adb commands. These scripts perform high-privilege operations such as modifying device settings (adb-settings-theme), managing packages (packagename uninstall), and capturing screen data.
  • [PROMPT_INJECTION]: The skill creates a surface for indirect prompt injection by processing output from external, potentially untrusted Android devices.
  • Ingestion points: Commands like scripts/adb-logcat-package, scripts/wearableservice-items, and dumpsys ingest arbitrary text from the device's logs and system services.
  • Boundary markers: There are no instructions or delimiters provided to ensure the agent ignores or sanitizes instructions embedded within device output.
  • Capability inventory: The skill provides significant capabilities, including file system access and the ability to execute system commands, which could be exploited if an attacker-controlled application on the connected device writes malicious instructions to the logcat.
  • Sanitization: The skill does not implement any sanitization or validation of the data retrieved from the Android device before presenting it to the agent.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 17, 2026, 08:31 AM