SpecWeave

The Enterprise Layer for AI Coding.

Permanent memory, GitHub/JIRA sync, quality gates, autonomous execution. Ship features while you sleep.

npm install -g specweave   # Requires Node.js 20.12.0+

---

Quick Demo

/sw:increment "User authentication"
/sw:auto                              # Ship while you sleep

What happens:

• AI creates spec + plan + tasks
• Executes autonomously for hours
• Tests, fixes failures, syncs to GitHub/JIRA
• You review finished work

[08:23:41] [Planning]      Analyzing T-003: Implement refresh token rotation
[08:24:12] [Implementing]  Writing src/auth/token-manager.ts
[08:25:33] [Testing]       Running tests... FAILED
[08:25:47] [Fixing]        Adjusting implementation...
[08:26:15] [Testing]       Re-running... PASSED
[08:26:22] [Done]          T-003 complete. Moving to T-004...

---

Why SpecWeave?

Every AI coding tool loses context when the chat ends. SpecWeave creates permanent documentation:

.specweave/increments/0001-oauth/
├── spec.md    <- WHAT: User stories, acceptance criteria
├── plan.md    <- HOW: Architecture decisions, tech choices
└── tasks.md   <- DO: Implementation tasks with tests

After 6 months: Search "OAuth" → find exact decisions, who approved, why it was built that way.

---

Key Differentiators

Lazy Plugin Loading (99% Token Savings)

SpecWeave loads plugins on-demand based on your prompt keywords:

Say "React frontend" → frontend plugin loads. Say "Kubernetes deploy" → k8s plugin loads. No manual configuration.

Self-Improving Skills

SpecWeave learns from corrections. When you fix something, it captures the learning:

## Skill Memories
<!-- Auto-captured by SpecWeave reflect -->
- Always use `vi.hoisted()` for ESM mocking in Vitest 4.x+
- Prefer native `fs` over fs-extra in new code

Next time, it won't make the same mistake.

Structured Documentation (No Root Bloat)

Everything stays organized in .specweave/:

.specweave/
├── increments/####-name/     # Feature specs + tasks
├── docs/internal/            # Living documentation
│   ├── architecture/adr/     # Architecture Decision Records
│   └── specs/                # Feature specifications
└── config.json               # Project configuration

Your project root stays clean. No scattered markdown files.

15+ AI Agents Working Together

Agents auto-activate based on context. Mention "security" → security expertise loads.

---

Install

Prerequisites

Node.js 20.12.0+ required (recommend Node.js 22 LTS).

node --version   # Check version

Getting SyntaxError: Unexpected token 'with'? Your Node.js is too old. Upgrade instructions →

New Project

npm install -g specweave
mkdir my-app && cd my-app
specweave init .

Then describe what you want in Claude Code:

"Build a calculator app with React"

Existing Project

npm install -g specweave
cd your-project
specweave init .

Then:

/sw:increment "Add dark mode"   # Create spec + plan + tasks
/sw:auto                        # Ship while you sleep

---

Core Commands

53 total commands →

---

External Integrations

...