Gasoline’s run_accessibility_audit tool surfaces WCAG violations as actionable findings your AI can fix.

The Problem

Accessibility is important but easy to overlook during development. Running audits manually, interpreting violations, and knowing how to fix them adds friction. Most developers defer a11y until it becomes a compliance requirement.

With Gasoline, your AI runs the audit and fixes the issues in one step.

How It Works

  1. Your AI calls run_accessibility_audit
  2. Extension runs axe-core against the live page
  3. Violations returned with severity, elements, and remediation guidance
// AI runs audit on the login form
{
  "scope": "#login-form",
  "tags": ["wcag2a", "wcag2aa"]
}

// Response:
{
  "violations": [{
    "id": "label",
    "impact": "critical",
    "description": "Form elements must have labels",
    "nodes": [{
      "target": ["input#email"],
      "html": "<input type=\"email\" id=\"email\" placeholder=\"Email\">"
    }]
  }]
}

MCP Tool Parameters

Parameter Description
scope CSS selector to scope the audit (default: full page)
tags WCAG tags to test: wcag2a, wcag2aa, wcag2aaa, best-practice

What’s Reported

Each violation includes:

Field Description
Rule ID axe-core rule identifier (e.g., color-contrast)
Impact Critical, serious, moderate, or minor
Description Human-readable explanation
Affected elements CSS selectors and HTML snippets
Help URL Link to detailed remediation guidance

Supported WCAG Levels

Tag Coverage
wcag2a Level A (minimum compliance)
wcag2aa Level AA (standard compliance)
wcag2aaa Level AAA (enhanced compliance)
best-practice Industry best practices beyond WCAG

Common Findings

Category Examples
Images Missing alt text, decorative images not marked
Color Insufficient contrast ratio
Forms Missing labels, unlabeled inputs
ARIA Invalid roles, missing required attributes
Structure Missing landmarks, heading order
Keyboard Not focusable, missing focus indicators

The AI Advantage

Your AI doesn’t just report violations — it fixes them. Because it has access to both the violation details (element selectors, rule descriptions) and your source code, it can:

  • Add missing ARIA labels to form inputs
  • Fix color contrast ratios in CSS
  • Add alt text to images
  • Correct heading hierarchy
  • Add keyboard navigation handlers
  • Re-run the audit to verify the fix

Use Cases

Pre-Commit Checks

“Run an accessibility audit on this page before I commit.”

Catch issues before they reach production.

Component Development

“Check the accessibility of this new modal component.”

Scope the audit to just the component you’re building.

Compliance Remediation

“Fix all critical accessibility issues on this page.”

Your AI audits, identifies violations, and applies fixes in one flow.

Post-Refactor Verification

“Did my refactor break any accessibility?”

Quick audit after UI changes catches regressions before users hit them.