claude doctor exists as a CLI subcommand but requires interactive input (an Enter press) to complete. This makes it unusable from scripts, hooks, or from within a running Claude Code session (nested sessions are blocked).
It'd be great if it could run fully non-interactively, e.g.:
claude doctor --no-input # run checks, print results, exit 0/1
claude doctor --json # structured output for scripting
My use case: I have hooks/scripts that modify settings.json and want to validate the config afterward without opening a separate interactive session.
claude doctorexists as a CLI subcommand but requires interactive input (an Enter press) to complete. This makes it unusable from scripts, hooks, or from within a running Claude Code session (nested sessions are blocked).It'd be great if it could run fully non-interactively, e.g.:
My use case: I have hooks/scripts that modify
settings.jsonand want to validate the config afterward without opening a separate interactive session.