React Doctor

We'd reach for React Doctor the moment AI-generated code starts landing in your React PRs. It's not just another linter; it's the guardrail that tells you when your agent wrote something clever but wrong. The health score gives you a quick number to chase, and the GitHub Action blocks merges when things regress, which is huge for keeping quality high without manual policing. When you should pick it: you're using Codex, Copilot, or any AI agent that writes React, or you're tired of useEffect traps and prop drilling sneaking into your codebase. The $30/mo team plan is affordable for what it does, and the free personal tier makes trying it on your own projects a no-brainer. When to pass: if your team is all-in on Vue or Svelte, or if your code is so stable that you don't need an extra CI gate, this likely won't earn its keep. Also, it's a linter, not a test suite—if you're expecting it to catch logic errors, you'll be disappointed. Compared to ESLint, which is single-file and rule-based, React Doctor's cross-file analysis is what sets it apart. ESLint catches style and obvious bugs, but it can't see the render loop caused by a useEffect dependency that spans two files. That's where React Doctor shines. In practice, we found the inline PR comments genuinely useful—they explain not just what's wrong, but why it matters, and suggest fixes. The autofix for some rules saves time, though not everything is auto-fixable yet. Watch out for the potential for false positives; with any static analysis, you'll occasionally get a warning that's a false alarm, so you'll need to tune rules or add suppressions as you go. But the ability to respect your existing ESLint/oxlint config means it plays well with your current setup. One more thing: the interactive TUI is a nice touch for