Frequently Asked Questions

Greenfield projects start from a blank slate — MySpec generates the full four-file bundle (constitution.md, requirements.md, solution.md, tasks.md) from your idea. Brownfield projects start from an existing codebase — MySpec analyzes what's already built, then walks through a change-focused interview and a proposal step before generating a requirements delta scoped to just that change.

Non-negotiable principles and constraints that guide all subsequent phases. This file acts as the project's 'supreme law', defining tech stack choices (e.g., 'Must use TypeScript'), architectural patterns, and coding standards that the AI must adhere to.

A detailed functional specification. We use the EARS+ (Easy Approach to Requirements Syntax) format to ensure requirements are unambiguous, testable, and complete. It defines 'what' the system does without prescribing 'how'.

A comprehensive solution design document. It translates the requirements into technical implementation details, including database schemas, API endpoints, component hierarchies, and system diagrams (Mermaid.js).

An implementation plan. It breaks down the approved design into granular, sequential tasks. Each task is scoped small enough to be easily completed by an AI coding agent or a human developer.

A change-request document generated for Brownfield projects. Instead of writing specs from scratch, proposal.md captures what's being added, changed, or removed in your existing codebase — the delta that the AI Architect then turns into a scoped requirements.md.

MCP (Model Context Protocol) is an open standard that lets AI apps like Claude Desktop, Claude Code, and Cursor connect directly to external tools and data. @myspec/mcp-server is MySpec's MCP server: install it with npx and your AI editor can list, read, create, and update your MySpec projects and spec files directly — no copy-pasting from the browser.