Choose by the work you need to do

Compare AI skills

Choose up to three resources. Compare their purpose, license and requirements, then read the scope of any recorded checks before installing.

Clear selectionBrowse the library →

This comparison does not rank or certify resources. Client labels are declared targets, not compatibility test results. Recorded outcomes can include passes and failures; they cover only the stated scenario and current package.

Web & HTML

Evidence-led Web Quality Audit

Organize a web review across accessibility, SEO, performance and browser practices using scoped evidence.

Purpose
Web quality audit
Author / publisher
Addy Osmani
License
MIT · License guidance
Version
afa8da942115.bb1
Access
Free download. External tools and services may have separate costs.
Declared clients
Claude Code, Codex, Cursor

Source & package

Inspect the original source ↗

8 files · 1 script/code file. License file detected.

Current package SHA-256dd1896332340b2c621d195dfea8004a885615a7b2515b7074ffbb9273bf590a3

Tools & requirements

Choose representative URLs and states, and use authorized browser/audit tools. The optional static HTML helper needs Bash, jq, grep and find. Set BB_SKILL_DIR to the installed web-quality-audit directory and pass the target project/file separately; run with Bash rather than assuming an executable bit.

Independent MIT project maintained by Addy Osmani; not official Google guidance or certification. The package does not provide Chrome DevTools MCP, Lighthouse, credentials, user telemetry or a model. Tool, model and infrastructure access may have separate costs. Check current W3C and Google documentation; automated checks cover a subset of issues and cannot establish WCAG conformance, application security, search rankings or field Core Web Vitals. Full skill/agent runtime evaluation has not been performed. Packaging links are adapted; required cross-skill reference documents are bundled, while related SKILL.md links open pinned upstream source and do not install other skills.

Evidence & limits

Source review checked 2026-10-03. Full skill/agent runtime evaluation has not been performed.

1 current outcome record. Read the actual environment, result and limitations.