Run doctor checks

omk doctor is a health audit for a single artifact — run it before you trust an eval, and as a CI gate. It does not compare two versions (that's omk eval); it tells you whether one artifact is well-formed enough to be worth measuring.

For the full flag list see the CLI reference. This guide covers the common tasks.

Quick health check

omk doctor                         # audit the current dir or ./skills
omk doctor skills/v1.md            # audit a single skill

You get per-dimension health scores (trigger boundary, doc clarity, instruction precision, dependencies, tool conventions, security, examples), findings, and concrete suggestions, sorted fail → warn → pass. Open the report in omk studio.

Offline / no-LLM mode

On a CI node without claude / codex, or when debugging offline, run only the static rules (readability / metadata / dependencies / samples-contract) — zero LLM calls, zero cost:

omk doctor skills/ --static-only

As a CI gate

omk doctor --gate; echo $?        # quiet; exit 1 on fatal problems, warnings don't block

--gate prints only a stderr summary on failure and signals the result via exit code, so you can wire it into a pipeline. Pair with --static-only for a fast, LLM-free preflight gate.

Machine-readable output

omk doctor skills/ --json > doctor.json

Emits the full DoctorReport as JSON for CI or external tooling to consume.

Auto-fix

omk doctor skills/v1.md --fix

Runs an LLM agent that reads the doctor findings and edits the skill to address them. Review the diff before committing.

Custom dimensions

Append your own LLM-scored dimensions on top of the built-in seven via a YAML file:

omk doctor --dimensions my-dimensions.yaml

(Programmatic registration via registerHealthDimension is also supported — the custom dimension is folded into the same LLM call and report.)

Related

• The three stages — where doctor sits in the doctor → eval → observe loop
• CLI reference: omk doctor — every flag