harmonyos-device-automation

Pass

Audited by Gen Agent Trust Hub on Sep 16, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill uses npx -y @midscene/harmony@1 to download and execute the Midscene automation package from the NPM registry. This is a standard and expected behavior for a skill designed to facilitate mobile automation using these tools.
  • [COMMAND_EXECUTION]: The skill utilizes the Bash tool to run HDC (HarmonyOS Device Connector) commands and the Midscene CLI. Specifically, it enables the execution of shell commands on the connected device via the runhdcshell functionality, which is essential for the skill's primary purpose of device control.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted visual data from device screenshots via the take_screenshot and act commands. This constitutes an attack surface where instructions embedded in a device's UI could attempt to influence the agent. However, this is an inherent aspect of vision-based automation, and no malicious implementation details were found.
  • [SAFE]: No obfuscation, data exfiltration, or persistence mechanisms were detected. API keys and configuration settings are handled through standard environment variables and .env files, following best security practices for secret management.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 16, 2026, 02:48 AM