Skip to content

quick-start.md

Latest commit

 

History

History
215 lines (150 loc) · 5.79 KB

quick-start.md

File metadata and controls

215 lines (150 loc) · 5.79 KB
Raw

Getting Started

DNA by Humaan is one npm package with two modes:

Mode How When
Build intelligence npx @superhumaan/dna-by-humaan doctor → full platform Always — one install for the whole squad folder
Production protection import from "@superhumaan/dna-by-humaan/runtime" When you have a Node.js API server

Portfolio squads: install on the parent folder that holds every related app — not inside each repo. Portfolio install guide →

Prerequisites

  • Node.js 20+
  • Git

Install (recommended)

cd /path/to/your-project          # or portfolio parent for multi-app squads
npx @superhumaan/dna-by-humaan doctor
npm install                       # if doctor added @superhumaan/dna-by-humaan

One dna doctor installs everything:

Layer Included
.DNA/ intelligence config, neuralNetwork, behaviour, CellularMemory, immune system
103 stem packs + /dna-* commands guidelines, expectations, workflows
AI workbench AGENTS.md, Cursor + Claude rules/skills — always on
Detection stack, monorepo apps, AI tools, GitHub, healthcare/legal, drift
Feature factory 9-role loop, quality gates, same template every feature
Runtime + CI auto-wired observer, GitHub Actions, pre-push hook, Docker
965 knowledge packs foundation from scan + marketplace

Doctor opens GitHub browser sign-in when needed and auto-wires the runtime observer for Express, Fastify, and Next.js.

For backend projects, the runtime package is added automatically when missing:

pnpm add @superhumaan/dna-by-humaan   # only if not already added by doctor
import { dnaRuntime } from "@superhumaan/dna-by-humaan/runtime";

Upgrade DNA (including Lab UI)

/labs is served by the package copy your API process resolves. Nested installs and long-lived Node processes are the usual reason Lab looks “stuck” on an old UI.

npx @superhumaan/dna-by-humaan@latest update   # CLI + packs + nested Lab installs
# or: npx dna lab installs --fix

# REQUIRED every time:
# 1. Restart the API / dna lab serve process that mounts Lab
# 2. Hard-refresh /labs (Cmd+Shift+R)
curl -s http://localhost:<api>/api/dna/labs/health | jq '{dnaVersion,labUi}'

Full checklist: Lab upgrade DX.

Portfolio install (multi-product squads)

cd ~/work/your-squad-folder    # parent of product-a, product-b, shared-api…
npx @superhumaan/dna-by-humaan doctor
dna analyze
dna context cursor

Commit .DNA/, DNA/, AGENTS.md, ai/, .cursor/, and .claude/ at the parent so the squad shares one brain.

Full manifest →

Initialise (interactive wizard)

Use init only when you want the four-question wizard before doctor:

dna init -y
dna doctor
dna validate

Interactive onboarding:

dna init
  1. Project name
  2. What are you building?
  3. Platform (web, mobile, desktop, CMS)
  4. Optional platform features

Describe features in plain language in chat — no copy-paste prompts:

dna feature "I want providers to record phone calls and transcribe notes"

Generated structure

your-project/
├── AGENTS.md
├── ai/                      # feature factory
├── .cursor/                 # rules, skills, commands, stems
├── .claude/
├── .github/workflows/       # dna-ci, preview, security
├── .DNA/
│   ├── config.dna.json
│   ├── behaviour/
│   ├── knowledge/
│   ├── stems/               # 103 prompt stem packs
│   ├── CellularMemory/
│   └── runtime/
└── DNA/Impressions/

Commit .DNA/, DNA/, workbench files, and CI workflows.

Daily workflow

dna scan
dna context cursor
dna plan compliance
dna validate

Type / in Cursor for stem packs and /dna-* commands.

Platform features (admin, SSO, deploy)

dna platform list
dna plan feature admin-portal --quote "Admin portal with Google directory sync"
dna context platform

See Platform Catalog.

Brownfield (existing project)

dna analyze --deep
dna document --from-code
dna plan ivf --quote "Integrate DNA without a rewrite"
dna context ivf

See Brownfield / IVF.

Production protection

pnpm add @superhumaan/dna-by-humaan
dna runtime install

See Runtime Observer.

Feature factory (automatic)

Installed during dna doctor. Describe what you want in Cursor or Claude — no prompts to copy. Optional terminal start: dna feature "...".

Each feature gets a local quality report (SonarQube-style SAST + lint/typecheck) in .DNA/reports/quality/. Agents run dna quality report --feature before marking work complete.

Close-out gates (automatic): dna docker build then dna github push. GitHub permissions are granted once during dna doctor via browser login — no manual tokens.

Optional: local reference repos

If you clone DNA's reference production apps (a production app, a production app, ops tooling, a production app), point the CLI at them:

export DNA_REFERENCE_ROOT=~/Projects
dna platform projects

See Platform Catalog.

Next steps

Topic Guide
Portfolio install One squad, many products
Core concepts Concepts
All commands CLI Reference
Knowledge packs Marketplace
Team rollout Team Testing
Naming (Humaan vs DNA) Naming conventions