a11y-pilot

image

An AI-powered command-line interface functioning as an ESLint for web accessibility. It scans frontend codebases for WCAG accessibility violations and intelligent auto-fixes them by orchestrating the GitHub Copilot CLI as the refactoring engine.

View on GitHub

Tech Stack

  • Languages: JavaScript, Node.js
  • Tools: GitHub Copilot CLI, AST Parsing, Commander.js

Features

  • Scans frontend files for 8 high-impact accessibility rules.
  • Reports issues with file, line number, severity, and precise WCAG references.
  • Context-rich prompt engineering system to intelligently build fixing instructions.
  • Auto-fixing integration that invokes GitHub Copilot CLI directly from the terminal.
  • CI-friendly JSON output formatting for automated pipelines.

Challenges and Solutions

  • Challenge: Automating Copilot CLI dynamically to execute correct accessibility fixes.

    Solution: Designed an execution shell wrapper that intelligently streams contextual instructions.

Learning Outcomes

  • Developed advanced knowledge of Abstract Syntax Tree (AST) parsing.
  • Learned how to translate complex WCAG accessibility guidelines into actionable code rules.
  • Built seamless integrations with third-party AI CLIs.