What problem does it solve?
Web performance slowly degrades as small changes accumulate, and teams need an automated way to measure, baseline, compare, and alert on regressions before they reach users. This skill captures real browser metrics to pinpoint regressions in load time, Core Web Vitals, request counts, and bundle sizes so problems are discovered on every PR rather than after release.
Core Features & Use Cases
- Baseline capture: Save deterministic baseline snapshots for pages and branches to enable precise before/after comparisons.
- Comprehensive metrics: Collect TTFB, FCP, LCP, DOM timings, full load, resource sizes, request counts, and top slow resources using a headless browse daemon.
- Comparison & thresholds: Auto-compare current runs against baselines with configurable thresholds and clear regression/warning statuses.
- Trend analysis & reporting: Load historical baselines to show trends, generate human-readable markdown and JSON reports, and store outputs locally.
- Use Case: Run on each pull request to detect JS bundle growth or LCP regressions and receive actionable recommendations before merging.
Quick Start
Run the benchmark against your site and save a baseline with the command "/benchmark https://example.com --baseline" then run "/benchmark https://example.com" to compare and generate a report.