Claude-Code-Workflow/docs/commands/codex/prep.md

Prep Prompts

One-Liner

Prep prompts are standardized templates for project context preparation — generating structured project core content packages through agent-driven analysis.

Core Content Package Structure

{
  "task_context": "Task context description",
  "keywords": ["keyword1", "keyword2"],
  "project_summary": {
    "architecture": "Architecture description",
    "tech_stack": ["tech1", "tech2"],
    "key_patterns": ["pattern1", "pattern2"]
  },
  "relevant_files": [
    {
      "path": "file path",
      "relevance": "relevance description",
      "priority": "high|medium|low"
    }
  ],
  "integration_points": [
    "integration point 1",
    "integration point 2"
  ],
  "constraints": [
    "constraint 1",
    "constraint 2"
  ]
}

memory:prepare

Function: Delegate to universal-executor agent, analyzing project via Gemini/Qwen CLI and returning JSON core content package for task context.

Syntax:

/memory:prepare [--tool gemini|qwen] "task context description"

Options:

• --tool=tool: Specify CLI tool (default: gemini)
  • gemini: Large context window, suitable for complex project analysis
  • qwen: Gemini alternative with similar capabilities

Execution Flow:

graph TD
    A[Start] --> B[Analyze Project Structure]
    B --> C[Load Documentation]
    C --> D[Extract Keywords]
    D --> E[Discover Files]
    E --> F[CLI Deep Analysis]
    F --> G[Generate Content Package]
    G --> H[Load to Main Thread Memory]

Agent Call Prompt:

## Mission: Prepare Project Memory Context

**Task**: Prepare project memory context for: "{task_description}"
**Mode**: analysis
**Tool Preference**: {tool}

### Step 1: Foundation Analysis
1. Project Structure: get_modules_by_depth.sh
2. Core Documentation: CLAUDE.md, README.md

### Step 2: Keyword Extraction & File Discovery
1. Extract core keywords from task description
2. Discover relevant files using ripgrep and find

### Step 3: Deep Analysis via CLI
Execute Gemini/Qwen CLI for deep analysis

### Step 4: Generate Core Content Package
Return structured JSON with required fields

### Step 5: Return Content Package
Load JSON into main thread memory

Examples:

# Basic usage
/memory:prepare "develop user authentication on current frontend"

# Specify tool
/memory:prepare --tool qwen "refactor payment module API"

# Bug fix context
/memory:prepare "fix login validation error"

Returned Content Package:

{
  "task_context": "develop user authentication on current frontend",
  "keywords": ["frontend", "user", "authentication", "auth", "login"],
  "project_summary": {
    "architecture": "TypeScript + React frontend, Vite build system",
    "tech_stack": ["React", "TypeScript", "Vite", "TailwindCSS"],
    "key_patterns": [
      "State management via Context API",
      "Functional components with Hooks pattern",
      "API calls wrapped in custom hooks"
    ]
  },
  "relevant_files": [
    {
      "path": "src/components/Auth/LoginForm.tsx",
      "relevance": "Existing login form component",
      "priority": "high"
    },
    {
      "path": "src/contexts/AuthContext.tsx",
      "relevance": "Authentication state management context",
      "priority": "high"
    },
    {
      "path": "CLAUDE.md",
      "relevance": "Project development standards",
      "priority": "high"
    }
  ],
  "integration_points": [
    "Must integrate with existing AuthContext",
    "Follow component organization pattern: src/components/[Feature]/",
    "API calls should use src/hooks/useApi.ts wrapper"
  ],
  "constraints": [
    "Maintain backward compatibility",
    "Follow TypeScript strict mode",
    "Use existing UI component library"
  ]
}

Quality Checklist

Before generating content package, verify:

• Valid JSON format
• All required fields complete
• relevant_files contains minimum 3-10 files
• project_summary accurately reflects architecture
• integration_points clearly specify integration paths
• keywords accurately extracted (3-8 keywords)
• Content is concise, avoid redundancy (< 5KB total)

Memory Persistence

• Session Scope: Content package valid for current session
• Subsequent References: All subsequent agents/commands can access
• Reload Required: New sessions need to re-execute /memory:prepare

Related Documentation

• Memory Commands
• Review Prompts
• CLI Invocation System