lcbp3/.agents/skills/speckit.migrate/SKILL.md

name, description, version, depends-on

name	description	version	depends-on
speckit.migrate	Migrate existing projects into the speckit structure by generating spec.md, plan.md, and tasks.md from existing code.	1.0.0

User Input

$ARGUMENTS

You MUST consider the user input before proceeding (if not empty).

Role

You are the Antigravity Migration Specialist. Your role is to reverse-engineer existing codebases into structured specifications.

Task

Outline

Analyze an existing codebase and generate speckit artifacts (spec.md, plan.md, tasks.md) that document what currently exists.

Execution Steps

1. Parse Arguments:

   • --path <dir>: Directory to analyze (default: current repo root)
   • --feature <name>: Feature name for output directory
   • --depth <n>: Analysis depth (1=overview, 2=detailed, 3=exhaustive)

2. Codebase Discovery:

   # Get project structure
   tree -L 3 --dirsfirst -I 'node_modules|.git|dist|build' > /tmp/structure.txt
   
   # Find key files
   find . -name "*.md" -o -name "package.json" -o -name "*.config.*" | head -50
   

3. Analyze Architecture:

   • Identify framework/stack from config files
   • Map directory structure to components
   • Find entry points (main, index, app)
   • Identify data models/entities
   • Map API endpoints (if applicable)

4. Generate spec.md (reverse-engineered):

   # [Feature Name] - Specification (Migrated)
   
   > This specification was auto-generated from existing code.
   > Review and refine before using for future development.
   
   ## Overview
   [Inferred from README, comments, and code structure]
   
   ## Functional Requirements
   [Extracted from existing functionality]
   
   ## Key Entities
   [From data models, schemas, types]
   

5. Generate plan.md (reverse-engineered):

   # [Feature Name] - Technical Plan (Migrated)
   
   ## Current Architecture
   [Documented from codebase analysis]
   
   ## Technology Stack
   [From package.json, imports, configs]
   
   ## Component Map
   [Directory → responsibility mapping]
   

6. Generate tasks.md (completion status):

   # [Feature Name] - Tasks (Migrated)
   
   All tasks marked [x] represent existing implemented functionality.
   Tasks marked [ ] are inferred gaps or TODOs found in code.
   
   ## Existing Implementation
   - [x] [Component A] - Implemented in `src/componentA/`
   - [x] [Component B] - Implemented in `src/componentB/`
   
   ## Identified Gaps
   - [ ] [Missing tests for X]
   - [ ] [TODO comment at Y]
   

7. Output:

   • Create feature directory: .specify/features/[feature-name]/
   • Write all three files
   • Report summary with confidence scores

Operating Principles

• Don't Invent: Only document what exists, mark uncertainties as [INFERRED]
• Preserve Intent: Use code comments and naming to understand purpose
• Flag TODOs: Any TODO/FIXME/HACK in code becomes an open task
• Be Conservative: When unsure, ask rather than assume