AI Performance · Lighthouse

web-performance-skill-cascade

A cascaded, editable performance skill system for AI coding agents: Lighthouse analysis, Core Web Vitals targets, focused modules, report parsing, performance budgets, and before/after demos.

Repository Source bundle Launch article

What it includes

A real Codex `web-performance` skill with progressive disclosure.
Twelve editable modules for Lighthouse, Core Web Vitals, LCP, INP/TBT, CLS, media, fonts, JavaScript, caching, third-party scripts, accessibility/SEO, and verification.
CLI scripts for Lighthouse/PageSpeed collection, report analysis, scenario composition, validation, and green-target budget generation.
Scenario fixtures, a Mirogate Lighthouse dogfood fixture, and before/after performance patch demos.

Why Mirogate built it

Performance work gets messy when every task becomes "make Lighthouse green." This project splits performance guidance into focused modules so an AI coding agent can load the right checks, patch narrowly, and verify the result.

Diagram showing a page, pull request, or Lighthouse report flowing into the web-performance root skill, then conditional Lighthouse, Core Web Vitals, LCP, images, fonts, INP, JavaScript, third-party, CLS, caching, and delivery modules, ending in patch, verify, and residual risk output. — The same cascade shown in the GitHub README, rendered as a PNG for the website.

npm test
node scripts/analyze.mjs --report examples/lighthouse/mirogate-uae-mobile.json
node scripts/compose.mjs --scenario landing-page-lcp
node scripts/budget.mjs --init --target green --stdout

This is not a guarantee of perfect Lighthouse scores, SEO ranking improvement, Google affiliation, or real-user field success. It is a practical performance skill architecture for AI-assisted engineering.