The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses a Python script scripts/inspect-adb.py to interact with the system's adb executable. Analysis of this script shows that it uses subprocess.run with shell=False (by passing an argument vector) for all commands, which is a secure practice that prevents command injection.
[SAFE]: The instructions consistently advocate for security best practices, such as never bypassing device authorization, using stable hashes for device serials in logs to protect identifiers, and explicitly warning against using destructive commands like testharness without authorization.
[SAFE]: Data exfiltration risks are mitigated by instructions that recommend bounded diagnostics, redaction of sensitive information in reports, and avoiding the exposure of adb over untrusted networks (Legacy TCP mode).
[SAFE]: The skill correctly identifies that adb kill-server is a disruptive host-wide action and provides targeted troubleshooting alternatives to minimize impact on other processes.

android-debug-bridge