Skip to content

Spec-Driven Development Framework for Human-AI Design & Development using V-Model

License

Notifications You must be signed in to change notification settings

kpruntov/SpecLoom

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

23 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

SpecLoom 🧵

The Compliance & Traceability Layer for AI Agents.

npm version License: MIT

SpecLoom is an MCP Server and CLI designed to enforce the V-Model in agile and iterative developemnt environments for AI-generated code. It acts as a Guardian that prevents hallucinations, ensures traceability, and mandates that every line of code serves a documented requirement.

Stop "Vibe Coding". Start Engineering.


🧠 Why SpecLoom?

  • For AI Agents: Provides structured "Context Bundles" (Requirements + Design + Code) so you don't have to guess.
  • For Humans: Enforces "Four-Eyes" review, prevents scope creep, and generates audit-ready documentation automatically.
  • For Teams: Bridges the gap between "Fast Prototyping" and "Enterprise Compliance".

⚡ Quick Start

1. Installation

npm install -g specloom

2. Get Started

Follow the Quickstart Guide to set up your project in 5 minutes.


🔌 AI Integration (MCP)

SpecLoom implements the Model Context Protocol (MCP), acting as the "Brain" for agents like Gemini CLI, Claude Desktop, Cursor, Windsurf, or Cline.

Configuration

Add SpecLoom to your agent's settings:

{
  "mcpServers": {
    "specloom": {
      "command": "npx",
      "args": ["-y", "specloom", "serve"]
    }
  }
}

Available Tools

  • loom_next: Asks "What should I work on?" (Project Manager)
  • loom_context: Asks "Give me the specs and code for this task." (Librarian)
  • loom_validate: Asks "Did I break anything?" (QA)
  • loom_verify: Asks "Does the code meet the requirements?" (Tester)

🛡️ Key Features

  • Strict V-Model Enforcement: No Code without Specs. No Specs without Context.
  • Graph-Based Traceability: Every artifact (User Story, API, Code, Test) is a node in a queryable graph.
  • The "Four-Eyes" Principle: Prevents self-approval of code (Identity separation).
  • Git-Native: All artifacts are JSON files committed alongside your code.

📄 Documentation

License

MIT

About

Spec-Driven Development Framework for Human-AI Design & Development using V-Model

Topics

Resources

License

Contributing

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors