Front-End Checklist

Measured conclusion: Deep-tested. L2 outcome: tools_ok.

Review frontend code and live pages against 386 quality-gated web development rules.

Add to your agent

Copy this into your MCP client config:

{
  "mcpServers": {
    "Front-End Checklist": {
      "url": "https://mcp.frontendchecklist.io",
      "transport": "streamable-http"
    }
  }
}

Measured tools (11)

• review_code read-only

  Runs a non-exhaustive static heuristic review of pasted HTML, CSS, or JavaScript against multiple frontend best practices and returns prioritized issues, fix guidance, and suggestions for manual verification. Use it when the user requests a review, audit, or debugging of source code. An empty issue list means no provable static issue was found; it does not establish accessibility conformance or runtime correctness.

  Parameters:

  • code (required) — string — The HTML, CSS, or JavaScript source to review (a snippet or a whole file).
  • focus — array — Only check these categories (default: auto-detected from the code).
  • minPriority — string — Lowest priority to report (default: medium).

  Hints: read-only yes · destructive no · idempotent yes · open-world no

• audit_url read-only

  Fetches a public HTTPS page and performs a non-exhaustive static review of its HTML, returning prioritized issues and fix guidance. Use it when the user requests an audit of a deployed public page. Localhost, private networks, and plain HTTP URLs are refused. It does not execute the page, measure live Core Web Vitals, or certify accessibility.

  Parameters:

  • url (required) — string — Public URL to audit. Must use https://. Private IPs and localhost are blocked.
  • focus — array — Only check these categories (default: auto-detected from the fetched HTML).
  • minPriority — string — Lowest priority to report (default: medium).

  Hints: read-only yes · destructive no · idempotent yes · open-world yes

• get_workflow read-only

  Returns the ordered steps of a curated audit workflow, including rule slugs, priorities, and categories. Use it when the user requests a structured audit such as a pre-launch checklist. Available workflows: accessibility-essentials, comprehensive-audit, core-web-vitals, html-foundations, image-optimization, javascript-resilience, launch-checklist, performance-quick-wins, privacy-and-consent, security-audit, seo-audit, testing-checklist.

  Parameters:

  • slug (required) — string — The checklist workflow slug. Available: 'accessibility-essentials', 'comprehensive-audit', 'core-web-vitals', 'html-foundations', 'image-optimization', 'javascript-resilience', 'launch-checklist', 'performance-quick-wins', 'privacy-and-consent', 'security-audit', 'seo-audit', 'testing-checklist'

  Hints: read-only yes · destructive no · idempotent yes · open-world no

• get_checklist_rules read-only

  Returns titles, priorities, verification and remediation prompts for every rule in a curated checklist. Use it when the user requests guidance for an entire checklist. includeContent optionally adds each rule’s long-form body. Available checklists: accessibility-essentials, comprehensive-audit, core-web-vitals, html-foundations, image-optimization, javascript-resilience, launch-checklist, performance-quick-wins, privacy-and-consent, security-audit, seo-audit, testing-checklist.

  Parameters:

  • checklist (required) — string — Checklist slug. Available: 'accessibility-essentials', 'comprehensive-audit', 'core-web-vitals', 'html-foundations', 'image-optimization', 'javascript-resilience', 'launch-checklist', 'performance-quick-wins', 'privacy-and-consent', 'security-audit', 'seo-audit', 'testing-checklist'
  • includeContent — boolean — Include full MDX body content (large). Default false returns title, description, prompts, and metadata only.

  Hints: read-only yes · destructive no · idempotent yes · open-world no

• get_quick_reference read-only

  Returns a compact checklist for one frontend category, filtered by priority and formatted as JSON, Markdown, or a task list. Use it when the user requests a concise checklist for a deployment, review, or team handoff.

  Parameters:

  • category (required) — string — Category name, e.g. 'accessibility', 'performance', or 'seo' (see list_categories).
  • format — string — Output format (default: 'json'). Use 'markdown' for readable docs, 'checklist' for copy-paste task lists.
  • priorityFilter — string — Filter by priority level (default: 'critical+high'). Use 'all' for comprehensive lists, 'critical' for essentials only.

  Hints: read-only yes · destructive no · idempotent yes · open-world no

• get_rule read-only

  Returns complete guidance for one known rule slug: requirements, rationale, verification and remediation prompts, code examples, sources, and related rules. Use it when the user requests detailed guidance for a specific frontend rule. Unknown slugs return similar suggestions.

  Parameters:

  • slug (required) — string — Rule slug: the kebab-case rule ID, e.g. 'alt-text' or 'doctype'. The ID must identify a rule in the published catalogue.
  • includeUrl — boolean — Include the rule's web URL in response (default: false)

  Hints: read-only yes · destructive no · idempotent yes · open-world no

• search_rules read-only

  Finds frontend rules by keywords, category, and priority, returning paginated slugs, titles, and short summaries. Use it when the user requests relevant practices for a technology, concern, or pattern, such as web fonts or accessible forms.

  Parameters:

  • limit — integer — Max results to return (default: 20, max: 100)
  • query — string — Keywords to match, e.g. 'lazy loading images'. Omit to list rules by filters only.
  • cursor — string — Cursor from a previous response, to fetch the next page of results.
  • categories — array — Only return rules in these categories.
  • priorities — array — Only return rules with these priorities.

  Hints: read-only yes · destructive no · idempotent yes · open-world no

• check_rule read-only

  Checks a code snippet against one named frontend rule using static heuristics and returns findings and verification guidance. Without code, returns manual verification guidance. Use it when the user requests a check of a specific rule or a proposed fix. Results do not establish rendered-state or runtime conformance.

  Parameters:

  • slug (required) — string — Rule slug: the kebab-case rule ID, e.g. 'alt-text' or 'doctype'. The ID must identify a rule in the published catalogue.
  • code — string — Code to test against the rule. Omit to get manual verification steps.

  Hints: read-only yes · destructive no · idempotent yes · open-world no

• fix_rule read-only

  Returns step-by-step remediation guidance and priority for one named frontend rule. Use it when the user requests implementation or fix instructions for a known issue. Optional codeSnippet provides context; the tool does not edit files or apply changes.

  Parameters:

  • slug (required) — string — Rule slug: the kebab-case rule ID, e.g. 'alt-text' or 'doctype'. The ID must identify a rule in the published catalogue.
  • codeSnippet — string — Code that violates the rule, so the fix guidance can reference it.

  Hints: read-only yes · destructive no · idempotent yes · open-world no

• explain_rule read-only

  Explains one frontend rule, including its background, impact on users and the business, and related categories. Use it when the user asks why a practice matters or requests educational context. It returns explanation guidance rather than applying a fix.

  Parameters:

  • slug (required) — string — Rule slug: the kebab-case rule ID, e.g. 'alt-text' or 'doctype'. The ID must identify a rule in the published catalogue.

  Hints: read-only yes · destructive no · idempotent yes · open-world no

• list_categories read-only

  Lists available frontend rule categories, their descriptions, and rule counts. Use it when the user asks what the checklist covers or needs valid category names.

  Hints: read-only yes · destructive no · idempotent yes · open-world no

Publisher website Repository

Measured by MCPYP probing. Untested fields are omitted, never estimated. Listing is always free; MCPYP never proxies calls or stores user tokens.