Inference Planz - Help
Display comprehensive help information for the inference-planz plugin.
Overview
Inference Planz is a multi-agent intelligence workflow plugin for Claude Code that helps you:
- •Understand Prompts Deeply: Research agent analyzes your intent
- •Clarify Requirements: Survey agent generates targeted questions
- •Plan Execution: Plan agent creates actionable roadmaps
Commands
| Command | Description |
|---|---|
/inference-planz:run <prompt> | Execute the full planning pipeline |
/inference-planz:planz <prompt> | Alias for run |
/inference-planz:help | Show this help documentation |
/inference-planz:status | Show current session status |
/inference-planz:configure | Configure plugin settings |
Quick Start
code
/inference-planz:run Build a REST API for user authentication with JWT tokens
This will:
- •Research what authentication patterns are best
- •Ask clarifying questions about your requirements
- •Generate a detailed implementation plan
Pipeline Stages
Stage 1: Input Normalization
- •Trims whitespace
- •Detects language and tone
- •Extracts entities, constraints, objectives
Stage 2: Research Agent
- •Deep analysis of user intent
- •Identifies missing details and risks
- •Recommends implementation approaches (A, B, C with tradeoffs)
Stage 3: Survey Agent
- •Generates 5-10 multiple-choice questions
- •Covers goal, scope, constraints, timeline, quality
- •Optimized for unblocking planning decisions
Stage 4: Plan Agent
- •Creates production-grade roadmap
- •Phases with checkpoints and deliverables
- •Includes file structure, interfaces, tests
- •Provides relative sizing estimates
Configuration
Create .claude/planz.config.json in your project:
json
{
"enabled": true,
"debug_mode": false,
"timeouts": {
"research_agent": 60,
"survey_agent": 45,
"plan_agent": 60
}
}
Environment Variables
| Variable | Description | Default |
|---|---|---|
INFERENCE_PLANZ_ENABLED | Enable/disable plugin | true |
INFERENCE_PLANZ_DEBUG | Enable debug output | false |
INFERENCE_PLANZ_TIMEOUT | Total pipeline timeout (seconds) | 180 |
Integration with Other Plugins
Works seamlessly with:
- •inference-confidenz: Confidence scoring for responses
- •inference-continuez: Auto-continuation based on confidence
Troubleshooting
Pipeline timeout: Increase INFERENCE_PLANZ_TIMEOUT or individual agent timeouts
Agent failures: Plugin includes fallback behavior - check logs for details
Empty output: Ensure your prompt is not empty; plugin will show domain selector
Links
- •Repository: https://github.com/Blerbz/blerbz-plugins
- •Issues: https://github.com/Blerbz/blerbz-plugins/issues