Add ADD (Agent Driven Development) skill
loading diff…
AI-assisted development today is ad-hoc — developers prompt Claude without structure, skip tests, lose context between sessions, and have no way to step away while work continues safely. ADD (Agent Driven Development) brings a complete SDLC methodology to Claude Code: specs before code, strict TDD enforcement, quality gates that scale with project maturity, and a learning system that accumulates knowledge across sessions and projects.
# Bootstrap a new project with structured interview
/add:init
# Create a feature spec via interview
/add:spec
# Generate implementation plan and run full TDD cycle
/add:plan specs/user-auth.md
/add:tdd-cycle specs/user-auth.md
# → RED: writes failing tests from spec
# → GREEN: implements minimum code to pass
# → REFACTOR: improves quality
# → VERIFY: independent sub-agent checks against spec
# Step away for 2 hours — Claude works autonomously
/add:away 2h
# Come back and get a structured briefing
/add:back
# Run retrospective to capture and promote learnings
/add:retro
Created by @MountainUnicorn. Repository: github.com/MountainUnicorn/add
add-agent-driven-development/SKILL.md with full skill documentation following the CONTRIBUTING.md template