/optimize-claudemd

name: nyann:optimize-claudemd
description: >
  Analyze CLAUDE.md usage patterns and optimize content based on actual
  reference frequency. Removes unused sections, compresses low-density
  content, and adds frequently-used commands that are missing.
arguments:
  - name: force
    description: Skip the 10-session minimum data requirement.
    optional: true
  - name: dry-run
    description: Preview changes without writing.
    optional: true

**Plugin root:** This is a Claude Code plugin, NOT a CLI tool. Do NOT search via `which`, `npm list`, `pip list`, or `brew list`. This file is at `<plugin_root>/commands/`. All scripts: `<plugin_root>/bin/`. Read the matching `skills/*/SKILL.md` for the full flow.

/nyann:optimize-claudemd

Run the `optimize-claudemd` skill (see `skills/optimize-claudemd/SKILL.md`):

1. Analyze usage data from `memory/claudemd-usage.json`. 2. Compute per-section density (references / bytes). 3. Identify unused docs, missing commands, and low-value sections. 4. Preview and confirm optimizations. 5. Apply changes within nyann markers only.

Exit codes:

• `0` — optimization complete (or insufficient data)
• `1` — error (no tracking data, no markers, etc.)

See also:

• `/nyann:doctor` — audit repo health including CLAUDE.md size
• `/nyann:bootstrap` — initial CLAUDE.md generation with tracking opt-in
• `/nyann:explain-state` — view current CLAUDE.md contents