github/accessibility-scanner

What it solves

This project provides a GitHub Action that automatically detects accessibility barriers in digital products. It streamlines the process of identifying accessibility gaps and fixing them by integrating directly into the GitHub ecosystem, converting findings into trackable issues and leveraging AI to propose code fixes.

How it works

The scanner runs as a GitHub Action workflow that scans specified URLs (including dynamic content) using engines like Axe and AccessLint. It can handle authenticated sessions via Playwright and can be extended with plugins (e.g., for detecting low-quality alt text). When a violation is found, the tool creates a GitHub issue. If GitHub Copilot is enabled, the issue can be assigned to the Copilot coding agent to automatically generate proposed fixes and pull requests, which humans then review before merging.

Who it’s for

Developers and QA teams who want to automate accessibility testing as part of their CI/CD pipeline to ensure their websites and repositories remain compliant with accessibility standards (such as WCAG).

Highlights

• AI-Powered Fixes: Integrates with GitHub Copilot to automatically propose and create PRs for accessibility issues.
• Automated Issue Tracking: Automatically creates, reopens, and manages GitHub issues based on scan results.
• Customizable Scanning: Supports multiple scan engines, custom plugins, and specific URL configurations to exclude certain CSS selectors.
• Flexible Authentication: Supports basic username/password login or complex authentication contexts via Playwright.
• Intelligent Caching: Uses a cache key to track results across runs to avoid duplicate issues.