web-quality-skills

Why The skills predated current Chrome DevTools traces, Lighthouse agent audits, CrUX context, and Agentic Browsing checks. They also needed a clearer distinction between field data, repeatable lab measurements, and observations from one browser session. What changed Use performance traces for diagnosis and live Lighthouse audits for accessibility, SEO, best practices, and agentic browsing when those capabilities are available. Add a shared field, lab, and first-party RUM workflow with low-friction fallbacks. Remove unsourced estimates while preserving source-backed implementation detail, including Speculation Rules, Early Hints, crawler controls, and search-result rendering behavior. Move deeper measurement, security, structured-data, INP, and CLS guidance into focused references, and update package metadata to 2.0.0. Compatibility The six skill names, trigger phrases, installation paths, and framework-agnostic guidance remain intact. Without Chrome DevTools MCP, the skills still use Lighthouse CLI, PageSpeed Insights or CrUX, manual checks, and static analysis. Tool routing is capability-first, with current MCP names kept as examples. Intentional differences v2 starts with live evidence before broad source inspection. It no longer treats one PerformanceObserver session as field data, no longer routes performance through the Lighthouse-for-agents audit when traces are available, and does not promise ranking or field improvements without evidence. It also corrects the DOMPurify Trusted Types policy callback to return a string. These changes are summarized in CHANGELOG.md. Validation Both skill validators pass for all six skills. Manifests, local links, line limits, and the audit script were verified.

Summary Adds two new Phase 2 skill modules focused on AI-native content discovery and structured data markup: skills/answer-engine-optimization — core skill covering content structure, extractable prose, comparison pages, and FAQ patterns for AI citation — deep reference on making content extractable by LLMs — strategies for AI-friendly product/service comparison pages — FAQ writing patterns optimized for AI answer extraction skills/structured-entities — core skill covering Schema.org entity markup (Organization, Product, FAQ, HowTo) for AI knowledge graph indexing — Organization & Brand entity patterns with sameAs disambiguation — Product & Service entity modeling including SoftwareApplication and pricing tiers — FAQ, HowTo & Q&A schema patterns with AI extraction guidelines Motivation Builds on Phase 1 (ai-search-audit + llm-discovery) to teach agents how to reshape actual content and markup for AI-native retrieval and citation. Informed by Norg.ai research and current LLM citation patterns.

This pull request updates the to provide more comprehensive documentation on installing and managing skills using the CLI tool. The new instructions clarify how to install all skills, translate between agents, install specific skills, and list available skills without installing them. Documentation improvements:** Expanded the installation instructions to include examples for using to install all skills, translate between agents, install specific skills, and list available skills without installing them.