Model Switchboard

# Model Switchboard v3.0 — Safe AI Model Configuration for OpenClaw

> ⛔ **HARD RULE: NEVER edit `openclaw.json` model fields directly.**
> Always use this skill's commands. No exceptions. Ever.

## Why This Exists

Editing `openclaw.json` directly for model changes is the #1 cause of OpenClaw gateway crashes. Wrong model type in wrong slot = instant death. No backup = hours rebuilding. This skill eliminates that entirely.

## How It Works

1. **Validates** model format and role compatibility before any change
2. **Auto-backs up** config before every modification (30 rolling backups)
3. **Uses OpenClaw CLI** (`openclaw models set`) — never raw JSON
4. **Blocks unsafe assignments** (image-gen model as primary LLM = blocked)
5. **Instant rollback** if anything goes wrong
6. **Canvas UI** for visual model management

## Quick Reference

```bash
SWITCHBOARD="$SKILL_DIR/scripts/switchboard.sh"

# View current setup
$SWITCHBOARD status

# Change models
$SWITCHBOARD set-primary "anthropic/claude-opus-4-6"
$SWITCHBOARD set-image "google/gemini-3-pro-preview"
$SWITCHBOARD add-fallback "openai/gpt-5.2"
$SWITCHBOARD remove-fallback "openai/gpt-5.2"
$SWITCHBOARD add-image-fallback "openai/gpt-5.1"

# Preview before applying
$SWITCHBOARD dry-run set-primary "openai/gpt-5.2"

# Discovery & recommendations
$SWITCHBOARD discover          # List all available models
$SWITCHBOARD recommend         # Get optimal suggestions

# Redundancy (3-deep failover)
$SWITCHBOARD redundancy        # Assess current redundancy
$SWITCHBOARD redundancy-deploy # Preview optimal config
$SWITCHBOARD redundancy-apply  # Apply optimal config
$SWITCHBOARD redundancy-apply 4  # Custom depth

# Backup & restore
$SWITCHBOARD backup            # Manual backup
$SWITCHBOARD list-backups      # Show all backups
$SWITCHBOARD restore latest    # Undo last change

# Import/Export (portable model configs)
$SWITCHBOARD export config.json
$SWITCHBOARD import config.json

# Cron model validation
$SWITCHBOARD validate-c...

# 🔀 Model Switchboard

**Never crash your OpenClaw gateway again.**

Safe, redundant AI model configuration with automatic fallback chains, validation, and one-command recovery. Built for every OpenClaw user — from first install to production deployment.

## The Problem

Editing `openclaw.json` directly for model changes is the #1 cause of OpenClaw gateway crashes:

- Put an image model as primary LLM → **gateway dead**
- Typo in model reference → **gateway dead**
- Provider goes down, no fallback → **gateway dead**
- No config backup → **hours rebuilding from scratch**

## The Solution

Model Switchboard gives you:

- **3-deep fallback chains** — if Model A fails, B catches it, then C
- **Provider diversity** — never stack the same provider in a chain
- **Pre-flight validation** — blocks unsafe assignments before they touch config
- **Auto-backup** — snapshots config before every change (30 rolling)
- **One-command recovery** — `restore latest` = 2 seconds, not 5 hours
- **Per-task routing** — right model for the right job (conversation, coding, images, heartbeat)
- **Visual dashboard** — see your entire model tree at a glance
- **Setup wizard** — guided first-time configuration

## Quick Start

```bash
# First time? Run the setup wizard:
./scripts/switchboard.sh setup

# See current model assignments:
./scripts/switchboard.sh status

# Auto-generate 3-deep redundant config:
./scripts/switchboard.sh redundancy-apply

# Change a model (validates + backs up first):
./scripts/switchboard.sh set-primary anthropic/claude-opus-4-6

# Something broke? Instant recovery:
./scripts/switchboard.sh restore latest
```

## All Commands

**Getting Started**
- `setup` / `init` — First-time setup wizard with provider detection

**Model Assignment**
- `status` — Show current model tree + health
- `set-primary <model>` — Set primary LLM
- `set-image <model>` — Set image/vision model
- `add-fallback <model>` — Add LLM fallback
- `remove-fallback <model>` — Remove LLM fallback
- `add-...