Computer Use Automation
Automates local browser and desktop UI task fixtures while generating deterministic action summaries and replay artifacts.
The Computer Use Automation skill equips Claude Code with deterministic UI execution capabilities for local browser and desktop task fixtures. Designed for sandbox testing and task evaluation, it coordinates fixture execution via dedicated helper scripts, producing reproducible action logs, summaries, and replay artifacts within designated workspace directories for streamlined review and auditability.
Key Features
01Generates structured action summaries for post-run evaluation
02Executes deterministic local browser and desktop task fixtures
03Supports controlled UI automation benchmarking and sandbox validation
04Provides CLI-driven execution workflows via standardized helper scripts
05Creates inspectable replay artifacts in configured workspace directories
060 GitHub stars
Use Cases
01Replaying automated browser fixtures to evaluate agent behavior
02Running controlled UI task benchmarks in isolated sandbox environments
03Generating reproducible UI interaction artifacts for testing and debugging