expertise

Expertise Dashboard

Display the current state of the agent expertise system for this project.

Steps

1. Check for .expertise/ directory in the project root. If it doesn't exist, tell the user:

   • The expertise system isn't initialized for this project
   • They can initialize it by running: bash $CLAUDE_PLUGIN_ROOT/scripts/init-expertise.sh
   • Stop here.

2. Read .expertise/config.yaml and display:

   • Configured max_lines limit
   • List of configured agents

3. For each model file in .expertise/models/, display:

   • Filename
   • Line count
   • Whether it's empty (only has the initial comment) or has content
   • Top-level YAML keys (first level only) to show what categories the agent is tracking

4. Check docs/solutions/ (managed by compound-engineering plugin). If it exists, count .md files per subdirectory and display a summary.

5. Format as a readable summary. Example output:

## Agent Expertise System

Config: max_lines=5000, 6 agents configured

### Mental Models

| Agent | Lines | Status | Tracking |
|-------|-------|--------|----------|
| ash-reviewer | 142 | active | patterns_discovered, architecture, open_questions |
| elixir-reviewer | 89 | active | patterns_discovered, codebase_observations |
| liveview-reviewer | 0 | empty | — |
| python-reviewer | 0 | empty | — |
| typescript-reviewer | 0 | empty | — |
| team-lead | 34 | active | project_patterns, workflow_observations |

### Institutional Knowledge (docs/solutions/)

12 solution docs across 5 categories:
  - runtime-errors/: 4 docs
  - best-practices/: 3 docs
  - database-issues/: 2 docs
  - test-failures/: 2 docs
  - build-errors/: 1 doc

Important

This is a read-only command. Never modify any files. Only read and report.