skills.rest
sanjanb
sanjanbCOMMUNITY
Explore all skills by this author →
GitHub

benchmark

Detects web performance regressions by benchmarking page load metrics against stored baselines.

1|Updated Jul 23, 2026
Search Tags:#performance#web vitals#bundle size#benchmark#regression detection#page speed#baseline comparison
One-click install
npx skills add https://github.com/sanjanb/my-agent-harness --skill benchmark-sanjanb
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: benchmark
Source: https://github.com/sanjanb/my-agent-harness/tree/main/skills/benchmark
Command: npx skills add https://github.com/sanjanb/my-agent-harness --skill benchmark-sanjanb

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Web performance degrades gradually across pull requests, and teams often notice only after load times have already suffered. This Skill measures page performance on every change, compares results against saved baselines, and surfaces regressions before they ship. ## Core Features & Use Cases - Baseline Capture and Comparison: Record baseline metrics for pages, then compare before/after on every PR to detect regressions in TTFB, FCP, LCP, and full load time. - Resource and Bundle Analysis: Inspect the slowest resources, JavaScript bundle sizes, CSS transfer sizes, and total network weight via the browse daemon's perf command and JavaScript evaluation. - Diff-Aware and Trend Modes: Benchmark only pages affected by the current branch with --diff, or review historical performance trends with --trend. - Use Case: Before merging a feature branch, run a diff-scoped benchmark to confirm the changed pages did not gain load time or bundle weight relative to the stored baseline. ## Quick Start Ask the agent to run /benchmark on your local dev URL with the --baseline flag first, then rerun it after your changes to compare performance against the baseline.

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 pull request?▼

Capture a baseline with /benchmark <url> --baseline before making changes, then rerun the benchmark on your branch. The skill compares TTFB, FCP, LCP, and load times against the baseline and flags regressions.

How to benchmark only the pages changed in my branch?▼

Use /benchmark --diff. It runs git diff against the PR base branch to identify affected files and benchmarks only the pages impacted by the current branch, avoiding a full-site audit.

What performance metrics does this benchmark collect?▼

It collects TTFB, First Contentful Paint, Largest Contentful Paint, DOM Interactive, DOM Complete, and full load time from the Performance API, plus per-resource durations, script and CSS transfer sizes, and total network weight.

Why does the benchmark say NEEDS_SETUP before running?▼

The browse daemon binary must be built once before any browse command works. Run the setup script in the skill directory, which installs bun if missing and compiles the daemon in about ten seconds.

Can I track performance trends over time instead of a single comparison?▼

Yes. Run /benchmark --trend to read historical benchmark reports stored in .gstack/benchmark-reports and visualize how page metrics have changed across previous runs.

Related Skills

Explore More Software Engineering Skills→
ingoatlingoatl

canary

Monitor live production URLs for console errors, performance metrics, and visual baselines.

Community
Advanced
Walrus-ComputingWalrus-Computing
20

benchmark

Detect performance regressions in web pages and app builds against saved baselines.

Official
Advanced
FxHollowFxHollow

benchmark

Measure page performance and compare results against historical baselines.

Community
Advanced
wilfozwilfoz

perf-lighthouse

Run Lighthouse audits via CLI or Node API and parse performance reports.

Community
Intermediate