Specification-Driven Development · v1.0.14

From Idea
to Production
Automatically

GSI-Protocol is an AI-powered SDD workflow that transforms any requirement into production-ready code through four structured phases — powered by Gherkin, Structure, and Implement.

4
Phases
3
AI Platforms
1
Command
claude-code ~ project
# Run complete SDD workflow in one command
$ /sdd-auto Add user authentication with JWT
→ Phase 1/4: Generating Gherkin specification...
✓ features/user-auth.feature created
→ Phase 2/4: Designing architecture...
✓ docs/features/user-auth/architecture.md created
→ Phase 3/4: Implementing code...
✓ src/auth/jwt.service.ts created
✓ src/auth/auth.middleware.ts created
→ Phase 4/4: Verifying implementation...
✓ All scenarios passed
✓ Feature complete. 4 files generated.
G
Gherkin
BDD specification in plain language
S
Structure
Language-agnostic architecture
I
Implement
Production-ready code generation

// Features

Built for how
developers actually work

Every aspect of GSI-Protocol is designed to eliminate the friction between idea and production code.

Multi-Platform Support

Works seamlessly with your existing AI assistant. No lock-in, no migration required.

Claude Code Codex (OpenAI) GitHub Copilot

Language Agnostic

Architecture design is independent of programming language — TypeScript, Python, Go, Rust — it works everywhere.

Automated Workflow

Execute the complete 4-phase development cycle with a single /sdd-auto command.

BDD Integration

Built-in Gherkin support turns business requirements into living documentation and executable specs.

Project-Aware Context

GSI-Protocol automatically detects your project's tech stack, conventions, and existing patterns. Architecture decisions adapt to what you already have — no manual configuration needed.

Role-Based Phases

Mirrors real team workflow: PM → Architect → Engineer → QA, each phase with a clear responsibility.

// 4-Phase Workflow

Structured from
requirement to release

Each phase has a clearly defined role, input, and output. No ambiguity, no guesswork.

1
PM Role

Specification

Converts your requirement into a structured Gherkin feature file with Given/When/Then scenarios.

/sdd-spec <requirement>
→ features/*.feature
2
Architect Role

Architecture

Designs a language-agnostic architecture document from your Gherkin spec, defining modules and interfaces.

/sdd-arch <feature>
→ docs/*/architecture.md
3
Engineer Role

Implementation

Generates production-ready source code based on the architecture document, respecting your project's conventions.

/sdd-impl <feature>
→ Source code files
4
QA Role

Verification

Verifies the implementation against every Gherkin scenario and generates a conclusion report.

/sdd-verify <feature>
→ docs/*/conclusion.md
Or run all 4 phases at once: /sdd-auto <your requirement>

// Quick Start

Up and running
in under a minute

The installer guides you through platform selection and installs everything automatically.

Recommended — uvx Recommended
$ uvx --from gsi-protocol-installer gsi-install
Alternative — pipx
$ pipx run gsi-protocol-installer

The installer will guide you through:

  1. 1 Select AI platform(s) — Claude Code, Codex, or GitHub Copilot
  2. 2 Choose installation scope — global or project-specific
  3. 3 All workflow commands installed and ready to use

// Commands

Complete reference

Command Description
/sdd-auto Execute complete 4-phase workflow automatically
/sdd-spec Generate Gherkin specification from requirements
/sdd-arch Design architecture from specification
/sdd-impl Implement code based on architecture
/sdd-verify Verify implementation against specification
/sdd-test Filter business scenarios and generate unit test shells
Claude Code / Codex
/sdd-auto <requirement>
GitHub Copilot
@workspace /sdd-auto <req>
Requirements
Python 3.10+ · Git

// Get started today

Stop writing boilerplate.
Start shipping features.

Install GSI-Protocol in under a minute and transform how you build software with AI.