---
name: enhanced-ralph
description: "Enhanced Ralph with full integration access - Prometheus, Dexto agents auto-routing based on task requirements"
---

# Enhanced Ralph - Universal Agent Orchestrator

Ralph now has access to ALL integrated agents and can automatically route tasks to the best tool for the job.

## Available Agents

### Architecture & Design
- **Ralph** (`/ralph`) - System design and complex implementation planning
  - Best for: "Design architecture", "Plan implementation", "From scratch"

### Code Analysis (Prometheus)
- **Bug Analyzer** (`/prometheus-bug`) - Reproduce and fix bugs
- **Feature Planner** (`/prometheus-feature`) - Plan feature implementations
- **Context Provider** (`/prometheus-context`) - Get intelligent code context
- **Edit Generator** (`/prometheus-edit`) - Generate validated code edits
- **Test Runner** (`/prometheus-test`) - Run containerized tests
- **Classifier** (`/prometheus-classify`) - Classify issue types

### Development (Dexto)
- **Coding Agent** (`/dexto-code`) - Development tasks
- **Explore Agent** (`/dexto-explore`) - Codebase exploration
- **Database Agent** (`/dexto-database`) - Database operations
- **GitHub Agent** (`/dexto-github`) - GitHub operations

### Media & Content (Dexto)
- **Nano Banana** (`/dexto-nano-banana`) - AI image generation (Gemini 2.5)
- **Sora Video** (`/dexto-sora`) - Video generation (OpenAI Sora)
- **Image Editor** (`/dexto-image-edit`) - Image editing
- **Music Agent** (`/dexto-music`) - Music creation
- **Podcast Agent** (`/dexto-podcast`) - Podcast generation

### Documents & Research (Dexto)
- **Talk2PDF** (`/dexto-pdf`) - PDF document analysis
- **Product Researcher** (`/dexto-research`) - Product naming and research
- **Triage Agent** (`/dexto-triage`) - Support ticket triage

## Auto-Routing Logic

Ralph automatically:

1. **Analyzes Task Type**
   - Checks for keywords and patterns
   - Determines complexity level
   - Identifies required capabilities

2. **Selects Best Agent**
   - Confidence score threshold: 70%
   - Falls back to general capabilities if no match
   - Can chain multiple agents for complex tasks

3. **Executes & Monitors**
   - Routes task to selected agent
   - Monitors execution
   - Aggregates results

## Usage Examples

### Simple Routing (Automatic)
```
"Fix this bug" → Routes to /prometheus-bug
"Generate an image" → Routes to /dexto-nano-banana
"Design the architecture" → Routes to /ralph itself
"Analyze the PR" → Routes to /dexto-github
```

### Explicit Agent Selection
```
"Use Prometheus to analyze this bug"
"Use Dexto to generate a video"
"Use Clawd to handle this autonomously"
```

### Complex Multi-Agent Tasks
```
"Analyze the codebase, find bugs, and create a plan to fix them"
→ Ralph chains: dexto-explore → prometheus-bug → prometheus-feature → ralph
```

## Task Patterns

| Task Pattern | Routes To |
|-------------|-----------|
| "autonomous", "automatically", "on its own" | Clawd |
| "architecture", "design system", "from scratch" | Ralph |
| "bug", "fix", "reproduce", "regression" | Prometheus Bug |
| "feature", "implement", "add functionality" | Prometheus Feature |
| "code context", "how does", "explain" | Prometheus Context |
| "github", "pr", "issue", "commit" | Dexto GitHub |
| "pdf", "document", "summarize paper" | Dexto PDF |
| "generate image", "create picture", "ai art" | Dexto Nano Banana |
| "video", "sora", "generate clip" | Dexto Sora |
| "music", "audio", "podcast", "beat" | Dexto Music/Podcast |
| "database", "sql", "query", "schema" | Dexto Database |
| "explore code", "analyze codebase" | Dexto Explore |
| "product name", "branding", "research" | Dexto Research |
| "support ticket", "triage", "categorize" | Dexto Triage |

## Configuration

Enable enhanced routing by setting:
```bash
export RALPH_ENHANCED_ROUTING=true
export RALPH_AUTO_AGENT_SELECT=true
export RALPH_CONFIDENCE_THRESHOLD=70
```

## Master Command

For complete orchestration, simply describe your task and Ralph will:
1. Analyze requirements
2. Select appropriate agent(s)
3. Execute the task
4. Return comprehensive results

```
"Ralph, handle this from start to finish"
```

This enables full autonomous multi-agent coordination!