MM

michaelwybraniec/mcp-agentic-sdlc

Developer tools

Эта страница переведена автоматически. Технические детали смотрите в оригинале. Открыть оригинал

10 stars 5 forks Качество 47 Тренд 22NOASSERTION

A comprehensive framework for managing software development lifecycle with AI agents, combining structured development processes with intelligent workflow management.

Обзор

A comprehensive framework for managing software development lifecycle with AI agents, combining structured development processes with intelligent workflow management. MCP Agentic SDLC consists of two main components: 1. - ASDLC.md - AI-Human collaborative development process - Structured development phases with agentic integration - Continuous feedback loops - Balanced responsibility distribution 2. - AWP.md - Human-AI collaboration workflow - Context management - Progress tracking - Standardized procedures Start with workflow.md for a comprehensive explanation of the server architecture, workflow, and how all components work together. After running the init tool, the Kanban board at (Cursor Simple Browser when available, otherwise your system browser). Markdown under backlog- / stays the source of truth; the board is a live read-only compile. - MCP tool — recompile kanban/backlog.

README

Показан фрагмент оригинального README. Открыть оригинал.

MCP Agentic SDLC

A comprehensive framework for managing software development lifecycle with AI agents, combining structured development processes with intelligent workflow management.

Overview

MCP Agentic SDLC consists of two main components:

  1. Agentic Software Development Lifecycle (ASDLC) - ASDLC.md

    • AI-Human collaborative development process
    • Structured development phases with agentic integration
    • Continuous feedback loops
    • Balanced responsibility distribution
  2. Agentic Workflow Protocol (AWP) - AWP.md

    • Human-AI collaboration workflow
    • Context management
    • Progress tracking
    • Standardized procedures

New to MCP Agentic SDLC? Start with workflow.md for a comprehensive explanation of the server architecture, workflow, and how all components work together.

Live Kanban quickstart

After running the init tool, the Kanban board starts automatically at http://localhost:4173 (Cursor Simple Browser when available, otherwise your system browser). Markdown under backlog-/ stays the source of truth; the board is a live read-only compile.

To start manually:

cd agentic-sdlc/kanban && npm run watch
  • backlog_sync MCP tool — recompile kanban/backlog.json from task markdown (pass startTaskId / completeTaskId when starting or finishing tasks)
  • backlog:///snapshot — agent-readable compiled state
  • Kanban toolbar — copy-paste AWP commands (awp refine, awp start, awp test, awp fix, awp next, awp auto, …)

See workflow.md — Section 11 for column rules, agent commits panel, time tracking, and backlog-complete flow.

Architecture

graph TB
    subgraph MCP["MCP Agentic SDLC"]
        direction TB
        
        subgraph ASDLC["Agentic Software Development Lifecycle"]
            direction LR
            ProblemDef["Problem Definition"]
            Design["Design Phase"]
            Development["Development Phase"]
            Testing["Testing Phase"]
            Deployment["Deployment Phase"]
            Maintenance["Maintenance Phase"]
            
            ProblemDef --> Design
            Design --> Development
            Development --> Testing
            Testing --> Deployment
            Deployment --> Maintenance
            Maintenance -.-> ProblemDef
            
            subgraph Core["Collaboration Core"]
                AIAgent["AI Agent"]
                Human["Human"]
                AIAgent -.-> Human
                Human -.-> AIAgent
            end
            
            Core -.-> ProblemDef
            Core -.-> Design
            Core -.-> Development
            Core -.-> Testing
            Core -.-> Deployment
            Core -.-> Maintenance
        end
        
        subgraph AWP["Agentic Workflow Protocol"]
            direction LR
            Context["Context Management"]
Открыть полный README на GitHub

Установка

This server does not publish a one-line install command.

Open the repository installation guide Новичок в MCP? Начните с руководства на mcpskill.org →