expo-skill-eval

End-to-end testing for Expo skills evaluates trigger accuracy, code quality, and runtime rendering by running automated scripts. It uses Python for orchestration to measure performance and verify output against expected quality standards on iOS and Android devices.

To benchmark Expo skill performance, you need a macOS environment with Xcode, Android SDK, Python, and bun installed. These dependencies allow the evaluation scripts to measure accuracy and runtime rendering across different mobile platforms.

Yes, you can compare different Expo skills by benchmarking their performance and accuracy against various inputs and configurations. This process measures trigger accuracy and runtime rendering to ensure they meet required specifications.

Output verification confirms that Expo skill outputs match expected results and meet quality standards. The evaluation platform checks runtime rendering and code quality to ensure the skill works as intended across different device configurations.

Python and bun are required dependencies because Python scripts handle the orchestration and automation of the end-to-end tests, while bun manages the JavaScript runtime environment needed to accurately evaluate Expo skill rendering.

Automated benchmarking should be used when testing new Expo skills or comparing multiple skills to ensure they meet required specifications. It provides comprehensive output verification and measures performance across various inputs faster than manual testing.