Use computer_use for live Windows or macOS GUI work that structured tools cannot complete. Discover an approved app and window, act from fresh observations, and verify every requested result.
The skill exposes the agent to untrusted, user-generated content from public third-party sources, creating a risk of indirect prompt injection. This includes browsing arbitrary URLs, reading social media posts or forum comments, and analyzing content from unknown websites.
The required workflow uses computer use to observe and interact with live GUI applications and documents, which can display arbitrary outsider-authored text.
f6dbc1d
If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.