What it actually does
A massive, curated library of MCP servers and agent skills focused strictly on production software engineering. Created by Addy Osmani (Chrome Engineering at Google), this repository gives your coding agent the ability to run Test-Driven Development (TDD), perform strict code reviews, enforce accessibility standards, and validate web performance metrics. Instead of your agent just writing raw code, it equips them with the tools to write *production-ready* code.
Who it's for
- 01
Engineering leaders tired of AI agents writing sloppy, untested spaghetti code
- 02
Frontend teams who want their AI agents to strictly adhere to Web Vitals and accessibility guidelines
- 03
Developers building CI/CD pipelines orchestrated by autonomous agents
Where it earns its keep
- Telling your agent to 'Implement the login feature, but you must write passing unit tests first'
- Having an agent review a PR and automatically run Lighthouse to check for performance regressions
- Wiring an agent to your design system so it outputs perfectly accessible HTML components
Use it, or skip it
Reach for it when
- You are using a coding agent (Cursor, Claude Code, Cline) in a professional, enterprise codebase
- You want to enforce strict software engineering principles on AI outputs
Skip it when
- You are just hacking together a quick weekend prototype and don't care about unit tests
- You aren't using a modern MCP-compatible coding agent
10 automations
Ideas, not tutorials. Each one is work a team does by hand today.
- 01Engineering
Automated TDD Enforcer
Agent intercepts feature requests, writes a failing unit test, and refuses to write the implementation until the test logic is approved.
- 02Engineering
Accessibility Auditor
Run the agent over your frontend repo nightly to flag missing ARIA labels and poor color contrast, opening PRs with fixes.
- 03Operations
Web Performance Regression Check
Agent runs Lighthouse on every PR staging branch and blocks merges if the Core Web Vitals score drops below 90.
- 04Engineering
Design System Migrator
Agent systematically replaces raw CSS classes with your company's official design system tokens across a large monorepo.
- 05Engineering
PR Code Reviewer
Agent acts as a senior engineer, leaving inline PR comments focused entirely on edge cases, memory leaks, and Big O complexity.
- 06Engineering
End-to-End Test Generator
Point the agent at a user flow in your app and have it write robust Playwright tests covering happy and sad paths.
- 07Operations
Automated Package Size Monitor
Agent analyzes bundle size changes in Webpack/Vite and alerts the team if a new dependency balloons the app size.
- 08Engineering
Legacy Code Refactorer
Agent breaks down a 2000-line monolithic file into clean, SOLID-compliant modules with perfect unit test coverage.
- 09Engineering
Security Linter
Agent scans new code for common OWASP vulnerabilities (like SQL injection or XSS) before it even reaches code review.
- 10Engineering
Automated Dependency Deprecation
Agent finds usages of deprecated internal libraries and safely rewrites them to the V2 API across the monorepo.
Want one of these running by Friday?
LimeDock builds these as real workflows inside your stack — deployed to your cloud, wired into your Slack and CRM, with the code in your repo. You pay a build fee and your own API keys, nothing else.
Source
Repository stats were read from the GitHub API and reflect the last time we refreshed this entry. The editorial breakdown above is LimeDock’s own analysis — we are not affiliated with addyosmani.
https://github.com/addyosmani/agent-skills