ADD — AI-Driven Development

Describe the feature. The agent drives the build. You approve once — exactly where a mistake would actually cost you.

---

Why This Exists

Every AI coding tool can write code fast now. The part that never got solved is trust — how do you know it built the right thing, and how do you know it's correct, without reading every line yourself?

ADD answers both. Freeze the direction before any code is written — spec → scenarios → contract → tests — then give one human approval, at that frozen contract. From there the agent builds and verifies against real evidence: passing tests and checked risks, never a diff that merely looks right.

It's for anyone who builds software with AI in the loop — engineers, architects, testers, designers, product owners, and the people who lead them.

✨ Highlights

• ✅ Approve once, then let it run — one human sign-off at the frozen contract; the agent builds the rest.
• 🔬 Proof, not promises — verified against observed behavior and pre-declared expectations, never just a plausible-looking diff.
• 🔒 Security never gets waved through — any security finding is a hard stop, human in the loop.
• 🌱 Prototype to production — task → milestone → graduate (analytics-gated) → recorded release, one method throughout.
• 🧠 Smarter as you go — lessons fold into a living, compacting foundation carried across milestones.
• 🎨 See it before you build it — a wireframe and a zero-dependency HTML mock, approved before any code.
• 👥 Built for teams — git-native multi-user, N parallel milestones, DAG-scheduled waves.
• 🧩 One slice, many components — monorepo or multi-repo, in one team.
• 🤝 Works with your AI — Claude, Copilot, Cursor, Codex, Gemini; install via npm, pip, or the Claude Code plugin.

Direction before speed. Trust comes from passing tests — not from reading code and finding it plausible.

---

🚀 Get Started

Here's the whole path, from nothing to your first running feature.

Prerequisites: Node ≥ 18 (npm path) or Python ≥ 3.10 (pip path), plus a CLI coding agent — Claude Code, Codex, or similar.

1 · Install into your project

From your project root (an empty folder or an existing repo), pick one ecosystem:

# Node / npm in project folder
npx @pilotspace/add init
npx @pilotspace/add update    # later, to update

or

# Python / pip in project folder
pip install pilotspace-add && pilotspace-add init
pilotspace-add update         # later, to update

or, on Claude Code, install the skill straight from the marketplace — no npm or pip needed:

/plugin marketplace add pilotspace/ADD
/plugin install add@add-method

See a real one: this repo's own .add/ folder.

2 · Spawn your first feature — talk to the agent

In Claude Code, run /add and say what you want to build:

# in claude code -> spawn ADD skill
> /add 'Describe your goal'

Example:

/add 'Let users log in with email + password / SSO, and keep them signed in for 30 days unless they explicitly log out.'

From there the agent runs the on-ramp for you:

1. 🧭 Orients from add.py status (the resume point) — never re-reading your whole repo.
2. 📐 Sizes your request into a milestone (goal · scope · breadth-first tasks · exit criteria) — you confirm the shape.
3. ✍️ Drafts each feature's one-approval front — Spec + Scenarios + Contract + Tests as one bundle — you give one approval, at the frozen contract.
4. ✅ Runs build → verify to green; a security finding always stops back to you.

3 · Resume anytime

/add status | continue

State lives on disk, not in the chat — the agent reports exactly where the project stands. Close your laptop, come back tomorrow, and pick up exactly where you left off. No context rot.

Want more power? ccsk-cli sharpens your agent's skillset for ADD (optional, recommended).

---

⚙️ How ADD Works

Curious how it works end to end? Three pictures, zoomed out one level at a time.

One task · eight steps · one file. Every feature is a single TASK.md that fills in section by section as it moves around the loop — each step produces exactly one durable artifact. The contract freeze is the one human approval; the agent drives the rest. (The artifacts are what you keep — the code is disposable.)

Tasks compound into milestones; milestones grow the project.

---

📚 Learn More

• 📖 Read the book — the full AIDD method, chapter by chapter
• ⚡ 2-minute Getting Started
• 🔍 Full hands-on walkthrough — one real feature, end to end
• 📦 Package source · Changelog
• 🗞️ ADD Across the Org: AI-Driven Development Beyond Code

Releases: @pilotspace/add (npm) · pilotspace-add (PyPI)

---

Star History

---

MIT License · pilotspace/ADD