Front-End Checklist: 385 Rules for Web Quality Audits
/Front-End-Checklist
Front-End Checklist turns scattered front-end best-practice knowledge into 385 discrete, checkable rules split across 11 categories. The payoff is depth in the categories most tools skimp on — accessibility runs 95 rules deep and SEO runs 94, more than double what HTML or JavaScript get. It's a checklist to walk through, not a linter that runs itself, so the value only shows up if someone, human or agent, actually works the list.
What is the Front-End Checklist
Front-End Checklist is an open-source rule corpus for front-end quality review, spanning HTML, CSS, JavaScript, performance, accessibility, SEO, security, and more. Each of its 385 rules gets its own page with an explanation, a Critical/High/Medium/Low priority tag, remediation guidance, and verification steps. Browse it on frontendchecklist.io, query it through the MCP server, or read it end to end in the README.
385 rules across 11 categories
- ✓385 rules across 11 categories, from a 5-rule Privacy section to a 95-rule Accessibility section
- ✓Every rule is tagged Critical, High, Medium, or Low so you can triage instead of reading start to finish
- ✓Rule pages pair an explanation with remediation guidance and verification steps, not just a checkbox
- ✓An MCP server at mcp.frontendchecklist.io exposes 11 tools, including review_code, audit_url, search_rules, get_workflow, and get_checklist_rules
- ✓Installable skills (npx skills add frontendchecklist/skills) package single rules, like https, into a focused agent workflow
- ✓A companion project, UX Patterns for Devs, covers choosing a UI pattern before you audit the implementation with this checklist
How to use the Front-End Checklist
- •Work through the category navigator on frontendchecklist.io before a launch or a pull request
- •Point an MCP-compatible agent at mcp.frontendchecklist.io and ask it to review a component or audit a live URL
- •Install a single-rule skill like https when you only need one focused check, not the full 385-rule pass
- •Call search_rules or get_workflow from inside an agent session before writing accessibility, SEO, or security code
- •Read the README straight through when you want the whole checklist in one file, offline from the site
Strengths
- ✓Covers categories generic linters ignore — privacy, internationalization, breadcrumb accessibility, file-upload accessibility
- ✓Priority tags mean 385 rules doesn't mean 385 equally urgent tasks
- ✓The MCP server puts the same rule corpus in front of a human reviewer and an auditing agent
- ✓Each rule explains the fix, not just the problem
What the Front-End Checklist does not cover
- △The license isn't stated in the repo's published facts — confirm it in the repository before using this in a commercial audit product
- △385 rules is a lot for a small project; there's no built-in way to scope the list down beyond browsing category by category
- △It's a checklist, not an automated scanner — nothing here rewrites your code, it only tells you what to fix
- △MCP tools like audit_url still depend on the agent driving them; a weak agent session gets a weak audit even with the right tool
Front-End Checklist alternatives
Frequently asked questions
Front-End Checklist is free to use, and the repository is public on GitHub with a documented contribution workflow (pnpm install, pnpm dev). The exact license isn't stated in the published repo facts, so check the LICENSE file in the repository before using it in a commercial product.
Front-End Checklist covers 11 categories: HTML, CSS, JavaScript, performance, accessibility, SEO, security, images, testing, privacy, and internationalization, totaling 385 rules. Accessibility (95 rules) and SEO (94 rules) are the two deepest categories.
Front-End Checklist ships an MCP server at mcp.frontendchecklist.io with 11 tools, including review_code for pasted code, audit_url for public pages, and search_rules for keyword lookup, so agents can use it directly. Point any MCP-compatible agent at it and name the server in your prompt, since some clients discover MCP tools lazily.
Front-End Checklist tags every rule Critical, High, Medium, or Low. Critical covers site-breaking, compliance-sensitive, or security-sensitive issues to fix first; High covers major impact on UX, accessibility, performance, or discoverability; Medium is standard best practice; Low is situational.
Direct CI/CD pipeline integration isn't clearly documented in the Front-End Checklist repository. The published workflows are browsing the website, querying the MCP server, or reading the README, not a CLI step wired into a pipeline, so check the repo directly if you need that.
The Front-End Checklist website (frontendchecklist.io) is for browsing and filtering rules interactively as a human. The MCP server (mcp.frontendchecklist.io) exposes the same 385-rule corpus as 11 tools, like review_code and audit_url, for agent workflows.
Who should try it — and who should skip
Teams doing a pre-launch review, and solo developers who want a second pass without hiring an auditor, get the most out of this — especially once you connect an MCP-capable agent and let it run search_rules or audit_url for you. If what you actually want is a single automated score with no manual review, this isn't it: Front-End Checklist hands you a rule corpus to work through, not a grade it calculates for you.
Related repositories
Still deciding about Front-End-Checklist?
One click hands the question to an AI along with this page — see what it says about Front-End-Checklist.