Requirements Analyzer Skill

Purpose

Extract requirements by analyzing ACTUAL test assertions and code implementations.

Key Difference

This does NOT guess. It reads:

• Test assertions to understand what's verified
• Code implementations to understand what exists
• Dependencies to understand platform requirements

Output

Updates PROJECT.md with evidence-based requirements.

Usage

python .claude/skills/requirements-analyzer/extract_requirements.py

Evidence Sources

1. Test Files → Functional Requirements

   • Parse test assertions
   • Extract acceptance criteria
   • Map to implementations

2. Code Documentation → Capabilities

   • Parse docstrings
   • Extract function signatures
   • Identify public APIs

3. Implementation Patterns → Non-Functional Requirements

   • Performance optimizations (caching, vectorization)
   • Memory management (clear_cache, gc.collect)
   • Platform dependencies (MLX, Python version)

Output Format

JSON file with:

{
  "functional": [
    {
      "id": "FR-1",
      "description": "...",
      "acceptance_criteria": "...",
      "test_file": "tests/...",
      "test_function": "test_...",
      "evidence": ["assertion1", "assertion2"],
      "status": "✅ Implemented"
    }
  ],
  "non_functional": [...],
  "traceability": [...]
}