computer-use-automation
Pass
Audited by Gen Agent Trust Hub on Jul 18, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill is a set of educational patterns and two diagnostic Python scripts (
action_safety_linter.pyandtool_choice_advisor.py). Analysis of the scripts shows they only use the Python standard library, perform local data processing (JSON/text parsing), and emit reports to stdout. There are no network operations, exfiltration patterns, or attempts to execute untrusted code. The skill's primary focus is actually improving agent safety by recommending programmatic APIs over brittle computer-use GUIs and flagging destructive actions that lack confirmation gates.
Audit Metadata