Multi-Agent Workflow

LDF supports using multiple AI agents to review your specs and code from different perspectives. This catches issues that a single agent might miss.

Overview

Agent	Strength	Use For
Your Primary Tool	Context retention, implementation	Development, code generation
ChatGPT	Pattern recognition, alternatives	Spec review, code audit, security
Gemini	Edge cases, thoroughness	Gap analysis, architecture review

Quick Start

1. Generate an Audit Request

ldf audit --type spec-review --spec user-auth

This creates a structured audit request with your spec content.

2. Send to External Agent

1. Open ChatGPT or Gemini
2. Paste the prompt from multi-agent/prompts/chatgpt/spec-review.md
3. Paste the generated audit request
4. Submit for review

3. Import Feedback

Save the response and import it:

ldf audit --import audit-response.md --spec user-auth

Available Audit Types

Type	Command	Best Agent
Spec Review	ldf audit --type spec-review	ChatGPT
Code Audit	ldf audit --type code-audit	ChatGPT
Security	ldf audit --type security	ChatGPT
Pre-Launch	ldf audit --type pre-launch	ChatGPT
Gap Analysis	ldf audit --type gap-analysis	Gemini
Edge Cases	ldf audit --type edge-cases	Gemini
Architecture	ldf audit --type architecture	Gemini
Full (All Types)	ldf audit --type full	Both

When to Audit

Phase	Recommended Audit
After drafting requirements	spec-review, gap-analysis
After design complete	architecture
Before implementation	edge-cases
Before merge	code-audit
Security-sensitive features	security
Before production release	pre-launch, full

Detailed Documentation

For comprehensive documentation including:

• Detailed workflow diagrams
• Audit templates
• Best practices
• Automated workflow (API integration)
• FAQ

See the Multi-Agent README.

Prompts Reference

ChatGPT Prompts

• spec-review.md - Review requirements and design specs
• code-audit.md - Review implementation code
• security.md - Security-focused review
• pre-launch.md - Production readiness review

Gemini Prompts

• gap-analysis.md - Find missing requirements
• edge-cases.md - Identify corner cases
• architecture.md - Validate architecture decisions

Security Best Practices

API Key Management

Always use environment variables for API keys:

Service	Environment Variable
OpenAI/ChatGPT	OPENAI_API_KEY
Google Gemini	GOOGLE_AI_API_KEY

Important: Never store API keys in .ldf/config.yaml - this file may be committed to version control and could expose your credentials.