Model Router Help

Display all available commands and current configuration.

Trigger

User runs /model-router:help

Instructions

Display a comprehensive help screen showing all commands and current config:

Model Router - Intelligent AI Model Routing
============================================

COMMANDS:
  /model-router:setup          First-time setup wizard
  /model-router:help           Show this help (you are here)
  /model-router:status         Show current config and available models
  /model-router:recommend      Get model recommendation for a task
  /model-router:configure      Change any setting
  /model-router:test <model>   Test a model's performance
  /model-router:learn          Analyze usage and suggest improvements
  /model-router:update-models  Pull latest community model recommendations

USAGE EXAMPLES:
  "What model should I use for code review?"
      → Model Router will recommend based on your setup

  "/model-router:recommend reasoning"
      → Get recommendation for reasoning tasks

  "/model-router:test gemini-3-pro"
      → Run test suite on a model, optionally update recommendations

  "Route this to Gemini"
      → I'll use Gemini CLI to handle the task

QUICK REFERENCE - Task Types:
  critical_code   → Production code, security-critical
  code_review     → Code review, refactoring
  reasoning       → Complex reasoning, math, logic
  simple_edits    → Typo fixes, formatting
  bulk_operations → Batch processing, data extraction
  writing         → Documentation, emails
  analysis        → Data analysis, summarization

HOW IT WORKS:
  Model Router is a recommendation system. When you ask for help
  with a task, it suggests the best model based on:

  1. Task type (what you're trying to do)
  2. Your subscriptions (what's available)
  3. Your preferences (quality vs cost)
  4. Model capabilities (benchmarks, strengths)

  You then follow the recommendation by telling Claude to use
  a specific model, or by running the CLI command directly.

CONFIGURATION:
  Config file: ~/.model-router/config.yaml
  Models file: ~/.model-router/models.yaml
  Tracking:    ~/.model-router/tracking.jsonl

DOCUMENTATION:
  Full docs: https://github.com/mikemorgenstern/model-router

Then read ~/.model-router/config.yaml and show current config:

YOUR CURRENT CONFIGURATION:
  Profile: [profile]
  Routing preference: [preference]
  Tracking: [enabled/disabled]

  Subscriptions:
    Claude Pro:    [✓/✗]
    ChatGPT Plus:  [✓/✗]
    Google AI Pro: [✓/✗]

  CLI Tools:
    Codex: [✓ installed / ✗ not found]
    Gemini: [✓ installed / ✗ not found]

  Local Models (Ollama):
    [list models or "Not configured"]

If config doesn't exist, prompt to run setup:

No configuration found. Run /model-router:setup to get started.