writ

Automate AI-assisted software development with contract-first specs and multi-agent SDLC.

8|1|Updated Feb 22, 2026
One-click install
npx skills add https://github.com/sellke/writ --skill writ
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: writ
Source: https://github.com/sellke/writ/tree/main
Command: npx skills add https://github.com/sellke/writ --skill writ

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) components.

What problem does it solve?

Writ provides an AI-powered development workflow framework that enforces contract-first specifications, a multi-agent SDLC, and automated quality gates to improve reliability, speed, and auditability of AI-assisted software projects.

Core Features & Use Cases

  • Contract-first specifications: lock the contract before building any artifact, enabling precise alignment between ideas and implementation.
  • Multi-agent SDLC: dedicated agents handle architecture checks, coding (TDD), reviews, testing, and documentation with feedback loops.
  • Automated quality gates: architecture checks, lint/type checks, security reviews, and coverage thresholds to prevent drift.
  • Cross-story continuity: What Was Built records propagate downstream for traceability and dependency integration.
  • Platform adapters: native support for Cursor, Claude Code, and OpenClaw.

Quick Start

Install Writ in your project and begin planning your first spec with /plan-product and /create-spec.

Frequently Asked Questions about writ

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I enforce contract-first specifications in an AI-assisted software development workflow?

You can enforce contract-first specifications by locking the contract before building any artifact. This AI-powered workflow uses dedicated commands to plan specs and create precise alignment between ideas and implementation.

How does a multi-agent SDLC improve the reliability of automated software pipelines?

A multi-agent SDLC assigns dedicated agents to handle architecture checks, TDD coding, reviews, testing, and documentation. This creates repeatable pipelines with deterministic gates, feedback loops, and drift-aware context.

Can I use automated quality gates to prevent specification drift in AI-generated code?

Yes, automated quality gates prevent drift by enforcing architecture checks, lint and type checks, security reviews, and coverage thresholds. These gates ensure AI-generated code meets deterministic standards before shipping.

What is the best way to start a contract-first software project using a multi-agent workflow?

The best way to start is to install the workflow framework in your project and use the planning commands to define your product scope and create your first specification before generating any code artifacts.

Does this contract-first development workflow support platform adapters for Cursor and Claude Code?

Yes, the workflow includes native platform adapters for Cursor, Claude Code, and OpenClaw, allowing you to integrate the multi-agent SDLC and automated gates directly into your existing development environments.

How do cross-story continuity records maintain traceability across parallel development pipelines?

Cross-story continuity maintains traceability by propagating What Was Built records downstream. This ensures dependency integration and context awareness across parallel execution within the multi-agent workflow.