benchmark

Detects web performance regressions by comparing Core Web Vitals and bundle sizes against baselines.

Updated Mar 23, 2026
One-click install
npx skills add https://github.com/dhnpmp-tech/project-agent --skill benchmark-dhnpmp-tech
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: benchmark
Source: https://github.com/dhnpmp-tech/project-agent/tree/main/.claude/skills/gstack/benchmark
Command: npx skills add https://github.com/dhnpmp-tech/project-agent --skill benchmark-dhnpmp-tech

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Web applications slow down gradually as each pull request adds small delays and larger bundles, and nobody notices until the app feels sluggish. This Skill measures real page performance, captures baselines, and flags regressions before they reach production. ## Core Features & Use Cases - Baseline Capture & Comparison: Record TTFB, FCP, LCP, DOM timings, transfer sizes, and request counts per page, then compare any later run against the saved baseline with regression thresholds. - Resource & Bundle Analysis: Identify the slowest and largest resources, track JavaScript and CSS bundle growth, and check results against industry performance budgets. - Trend Tracking: Review historical benchmark data to spot gradual performance degradation over days or weeks. - Use Case: Before merging a PR, run a benchmark against your staging URL to discover that LCP doubled and the JS bundle grew 60%, then pinpoint the new dependency responsible. ## Quick Start Ask the AI to run /benchmark against your app's URL with the --baseline flag first, then re-run it after making changes to see a regression report.

Frequently Asked Questions about benchmark

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I detect performance regressions before merging a PR?

Capture a baseline with /benchmark <url> --baseline on the main branch, then run /benchmark <url> on your feature branch. The skill compares TTFB, FCP, LCP, transfer size, and bundle size, flagging anything exceeding regression thresholds.

What metrics does a web performance benchmark measure?

It measures TTFB, First Contentful Paint, Largest Contentful Paint, DOM Interactive, DOM Complete, full load time, total requests, transfer size, and JS/CSS bundle sizes using the browser's Performance API via the browse daemon.

Can I benchmark only the pages changed in my branch?

Yes, use /benchmark --diff to benchmark only pages affected by the current branch. It diffs against the PR base branch to determine which pages were touched, or you can specify pages explicitly with --pages.

What counts as a performance regression in the report?

Timing metrics regress when they increase over 50% or by more than 500ms; bundle sizes regress above 25% growth. Smaller increases (20% timing, 10% bundle, 30% request count) are flagged as warnings rather than regressions.

Why does the benchmark need a baseline to be useful?

Without a baseline the skill can only report absolute numbers, which lack context since acceptable load times vary by page type. A baseline enables relative comparison against your own app's historical performance, which is how regressions are detected.