adb-magisk

Warn

Audited by Gen Agent Trust Hub on Sep 4, 2026

Risk Level: MEDIUMCOMMAND_EXECUTIONPRIVILEGE_ESCALATIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: Multiple Python scripts utilize the subprocess.run module to execute arbitrary shell commands via the Android Debug Bridge (ADB).
  • scripts/adb-magisk-launch.py executes adb shell am start to launch activities.
  • scripts/adb-magisk-enable-zygisk.py and scripts/adb-magisk-install-module.py execute various adb shell input tap and adb shell getprop commands.
  • [PRIVILEGE_ESCALATION]: The skill explicitly interacts with the device's root subsystem using the su command to perform actions that bypass standard application sandboxing.
  • workflow/magisk-verification.toon executes su -c 'id' and su -c 'cat /data/adb/magisk.db' to verify root functionality.
  • workflow/module-management.toon performs sensitive file system operations such as rm -rf /data/adb/modules/{{ module_name }} which requires root privileges.
  • [INDIRECT_PROMPT_INJECTION]: The skill has a large ingestion surface for untrusted data that could influence agent behavior.
  • It processes external Magisk module ZIP files in scripts/adb-magisk-install-module.py without content validation.
  • It parses command output from the device (e.g., pm list packages, mount, ls /data/adb/modules/) in various workflows to determine its next execution steps.
  • [DATA_EXPOSURE]: The skill accesses protected areas of the device filesystem, such as /data/adb/, to verify installation state and list modules, as seen in workflow/magisk-verification.toon.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Sep 4, 2026, 12:32 PM
Security Audit — agent-trust-hub — adb-magisk