GitHub Copilot: Strong Autocomplete but Weak Project Context and a Security Caveat

A nearly year-long evaluation of GitHub Copilot finds it excels at fast inline code completions with broad editor support across VS Code, JetBrains, Neovim, and Visual Studio, but falls short on deep project-wide understanding compared to rivals like Cursor. In a direct test, Copilot's workspace agent missed key components of a codebase's error-handling strategy that Cursor answered comprehensively on the first attempt. Pricing ranges from $10/month for individuals to $39/seat/month for enterprise, making it the most affordable paid AI coding tool for basic autocomplete and chat. A separate security incident this month revealed that Copilot Autofix reviewed a GitHub Actions workflow containing a script-injection flaw and marked it clean, meaning it missed rather than introduced the vulnerability. The takeaway for teams relying on Autofix as a security layer is that a clean result should be read as 'nothing obvious detected,' not as a thorough audit.