Passed 5 of 20 test cases with the skill, and 0 of 20 without it.
Without skill0%
With skill23%
5 fixed0 passed either way17 still failing0 broke
Pass-rate lift
+23 pp
Tokens
+149%
Agent turns
0%
Response time
-31%
Model gemini-3.6-flashMeasured 7/27/2026v15 of 22 cases passed with skill · 20 comparable
Tokens, turns and time describe resource use, not a dollar price.
What's inside
The instructions your model reads when the skill loads.
SKILL.md483 lines6 filesincludes scripts
Name
Code → PRD
Description
Reverse-engineer any frontend, backend, or fullstack codebase into a complete Product Requirements Document (PRD). Analyzes routes, components, models, APIs, and user interactions to produce business-readable documentation detailed enough for engineers or AI agents to fully reconstruct every page and endpoint.
Code → PRD: Reverse-Engineer Any Codebase into Product Requirements
Mock detection: Automatically distinguishes real API integrations from mock/fixture data
Enum extraction: Exhaustively lists all status codes, type mappings, and constants
Model extraction: Parses Django models, NestJS entities, Pydantic schemas
Automation scripts: codebase_analyzer.py for scanning, prd_scaffolder.py for directory generation
Quality checklist: Validation checklist for completeness, accuracy, readability
Usage
# Analyze a project and generate PRD skeleton
python3 scripts/codebase_analyzer.py /path/to/project -o analysis.json
python3 scripts/prd_scaffolder.py analysis.json -o prd/ -n "My App"
# Or use the slash command
/code-to-prd /path/to/project
Examples
Frontend (React)
/code-to-prd ./src
# → Scans components, routes, API calls, state management
# → Generates prd/ with per-page docs, enum dictionary, API inventory