/skill-doctor: Claude Code Tells You Which Skills It Never Touched
Claude Code 2.1.261 adds a /skill-doctor command that reports which loaded skills went unused, plus configurable limits for how much Bash and task output Claude sees inline.
What it is
A release of Anthropic's Claude Code CLI, version 2.1.261.
What it does
Adds `/skill-doctor` to show which loaded skills went unused in a session, `bashOutputMaxChars` and `taskOutputMaxChars` settings to raise how much command and background-task output is passed inline before falling back to a saved file (up to 128K characters), and `--append-subagent-system-prompt-file` for subagent prompts too large for the command line. It also adds an organization-policy status line explaining why a policy failed to load, for example a proxy that isn't passing the endpoint through.
Why it matters
Bloated skill libraries quietly cost context and money if half of them never fire; `/skill-doctor` turns that guesswork into a list you can act on, and the raised output limits mean fewer silent truncations on verbose build or test commands.
How to use it
Update to 2.1.261 and run `/skill-doctor` after a session to see which loaded skills went unused, then raise `bashOutputMaxChars` in settings if your commands produce large output.