ADD (Agent Driven Development) solves this by bringing structured, specification-driven SDLC practices to AI-native development—just as TDD revolutionized testing and DDD revolutionized domain modeling, ADD provides a proven methodology for human-AI collaboration at scale.
The human is the architect and decision maker; the agent is the builder. Gather requirements via one-question-at-a-time interviews. Never batch questions. When stepping away, declare autonomy ceilings explicitly — the agent may climb local → dev → staging but never merges to main or ships to production without human approval.
Every feature flows through the document hierarchy: PRD → Feature Spec → Implementation Plan → User Test Cases → Automated Tests → Implementation. No link may be skipped. Specs live in specs/, plans in docs/plans/. Code changes without a corresponding spec are rejected on review.
Project maturity: beta.
- local: auto-promotion not declared.
Production deploys and merges to the default branch always require human approval.
None in flight. Last shipped: specs/codex-agent-prefixing.md (v0.11.0), specs/init-defaults.md, specs/doctor.md, specs/codex-install-manifest.md (v0.10.2). Current focus: GA promotion (docs/milestones/v1.0-ga.md) — marketplace approval pending.
.add/config.json— project configurationdocs/prd.md— product requirementsspecs/— feature specificationscore/rules/— behavioral rules (ADD-managed projects)