legacy

name	legacy
description	Legacy code reverse engineering. Use this skill to analyze existing code and generate documentation (ADR/SDD/DDD/TDD/VDD) by building a logical understanding tree.

Legacy - Reverse Engineering Documentation

Analyzes existing code and generates documentation automatically.

Core Concept: Recursive Understanding Tree

AI builds a logical understanding tree that grows DEEP, not wide. Each directory is a logical concept, each _node.md is understanding at that level.

Recursive Traversal Algorithm

1. ENTER: Push to stack, create _node.md, form hypothesis.
2. EXPLORE: Read source code, validate understanding.
3. SPAWN: Identify child concepts needing deeper analysis.
4. RECURSE: For each child, repeat process.
5. SYNTH: Combine children insights, update _node.md.
6. EXIT: Generate/Update ADR/SDD/DDD/TDD/VDD docs.

---

Match Flow Protocol

CRITICAL: Before creating any flow, search for existing matching flow using keyword overlap.

• If match found (score >= 2): APPEND new insights as "Legacy Additions".
• If no match: CREATE new flow in DRAFT status.

---

Mandates

• FIRST: Scan all existing flows before any analysis.
• MATCH: Search for existing matching flows before creating new.
• APPEND: Update existing flows with "Legacy Additions" sections only.
• ASK: Stop immediately on conflicts, don't defer.
• PERSIST: State in _traverse.md after EVERY action.