Benchmark

功能

• Establishes baselines for page load times, Core Web Vitals, and resource sizes.
• Compares performance metrics before and after PRs.
• Tracks performance trends over time.
• Collects detailed metrics via JavaScript evaluation and browse daemon commands.

使用场景

• Use when optimizing web page speed and Core Web Vitals.
• Use when identifying performance regressions introduced by code changes.
• Use when monitoring and reporting on performance trends over time.

非目标

• Does not perform code optimization directly; focuses on measurement and reporting.
• Does not replace full performance profiling tools but provides a regression detection layer.
• Does not handle non-web performance metrics.

工作流

1. Setup: Initialize directories and check for updates.
2. Page Discovery: Auto-discover pages or use specified pages; identify pages affected by current branch in diff mode.
3. Performance Data Collection: Use browse daemon to gather TTFB, FCP, LCP, DOM metrics, resource analysis, bundle size, and network summary.
4. Baseline Capture: Save collected metrics to a baseline file if in `--baseline` mode.
5. Comparison: Compare current metrics against baseline, identifying regressions and warnings.
6. Resource Analysis: Report on the top slowest resources.
7. Performance Budget Check: Compare metrics against industry standards.
8. Trend Analysis: Show performance trends from historical data if in `--trend` mode.
9. Save Report: Write the final report to markdown and JSON files.

实践

• Performance measurement
• Regression testing
• Benchmarking
• Web performance analysis

先决条件

• Claude Code
• Git
• Bun v1.0+
• Node.js (Windows only)
• gstack browse daemon