/doctor vs /debug
People often think /doctor and /debug both just find what is wrong. They look in different places. /doctor checks your setup: the install, the settings files, your extensions and CLAUDE.md. It proposes fixes and applies them after you agree. /debug turns on the log of a running session and reads it, to find a problem that shows up while you work. Run /doctor when something will not load, and /debug when something goes wrong mid-task.
/doctor Command
It checks your setup and offers to fix what is wrong.
/doctor is a bundled skill that runs a checkup on your Claude Code setup. It looks at the install, the settings files, your extensions and your CLAUDE.md. It flags things like a broken settings file, an extension you never use, or CLAUDE.md lines Claude could learn from the code. Then it proposes fixes and applies them only after you agree. If Claude Code will not start at all, run claude doctor in your shell instead.
/debug Command
It turns on the debug log and reads it to find a problem.
/debug is a bundled skill that turns on debug logging for the session and then reads that log to find a problem. Logging is off unless you started with claude --debug, so /debug records from the moment you run it. Add a few words about the problem to point Claude at it.
In one line
- /doctor Command
- It checks your setup and offers to fix what is wrong.
- /debug Command
- It turns on the debug log and reads it to find a problem.