Skip to content

ParkerRex/ai-sdlc

BranchesTags

Repository files navigation

AI-SDLC CLI

Markdown-driven software-lifecycle powered by AI agents

PyPI version License Python version AI Powered

uv pip install ai-sdlc

AI-SDLC GitHub

📋 Table of Contents

🚀 Getting Started

📖 Usage & Workflow

🏗️ Technical Details

🛠️ Development

📚 Resources

🤖 What is AI-SDLC?

AI-SDLC transforms software development into a structured, AI-assisted workflow that takes you from initial idea to production-ready code through 8 carefully designed steps.

✨ Key Features

  • 🎯 Structured Workflow: 8-step process from idea → PRD → architecture → tasks → tests
  • 🤖 AI-Powered: Leverages AI agents for automated processing and iteration
  • 📝 Markdown-Driven: Everything lives in version-controlled markdown files
  • 🔄 Iterative: Built-in support for refining ideas and requirements with AI chat
  • 🚀 Production-Ready: Generates comprehensive task lists and test plans
  • 📦 Zero Config: Works out of the box with sensible defaults

💡 Why AI-SDLC?

Traditional development often jumps straight to coding, missing crucial planning steps. AI-SDLC ensures you:

  • ✅ Never skip important planning phases
  • ✅ Document decisions and rationale
  • ✅ Generate comprehensive implementation plans
  • ✅ Create thorough test strategies
  • ✅ Maintain project history in version control

🚀 Quick Start

Get up and running with AI-SDLC in under 2 minutes:

1️⃣ Install Dependencies

# Install uv (fast Python package manager)
curl -LsSf https://astral.sh/uv/install.sh | sh

# Install AI-SDLC
uv pip install ai-sdlc

2️⃣ Initialize Your Project

# Create a new directory or use existing project
mkdir my-awesome-project && cd my-awesome-project

# Initialize AI-SDLC
aisdlc init

3️⃣ Start Your First Feature

# Create a new feature
aisdlc new "Add user authentication system"

# Check your progress
aisdlc status

4️⃣ Follow the Workflow

  1. Fill out the generated markdown in doing/add-user-authentication-system/0-idea-*.md
  2. Run aisdlc next to generate an AI prompt for the next step
  3. Use the prompt with your preferred AI tool (Claude, ChatGPT, Cursor, etc.) and save the response
  4. Repeat until all 8 steps are complete
  5. Archive with aisdlc done

💡 Tip: AI-SDLC works with any AI tool - use your favorite AI assistant!

🎯 Flexible Usage

Option 1: Full CLI Workflow - Use the complete workflow with aisdlc commands that generate prompts for your AI tool

Option 2: Prompts Only - Just use the prompt templates manually with your preferred AI tool:

  • Copy prompts from prompts/ directory
  • Use with any AI chat interface (Cursor, Claude, ChatGPT, etc.)
  • Perfect for one-off projects or custom workflows

📦 Installation

Prerequisites

Tool Install (macOS example)
Python 3.13+ brew install [email protected] or pyenv install 3.13.0
uv curl -LsSf https://astral.sh/uv/install | sh
AI Editor (optional) Cursor, VS Code with AI extensions, or any AI chat interface

Install AI-SDLC

# Option 1: Using uv (recommended)
uv pip install ai-sdlc

# Option 2: Using pip
pip install ai-sdlc

# Verify installation
aisdlc --help

📖 Basic Workflow

The AI-SDLC workflow follows an 8-step process from idea to tests:

flowchart TD
    I[01-idea]-->P1[02-prd]-->P2[03-prd-plus]-->A[04-architecture]
    A-->SP[05-system-patterns]-->T[06-tasks]-->TP[07-tasks-plus]-->TESTS[08-tests]

    %% Iteration loop for steps 1-5
    CHAT[💬 Iterate with AI Chat]
    I -.-> CHAT
    P1 -.-> CHAT
    P2 -.-> CHAT
    A -.-> CHAT
    SP -.-> CHAT
    CHAT -.-> I
    CHAT -.-> P1
    CHAT -.-> P2
    CHAT -.-> A
    CHAT -.-> SP

    %% Agent mode for steps 7-8
    AGENT[🤖 Use AI Agent Mode]
    TP --- AGENT
    TESTS --- AGENT
Loading

🎯 Workflow Modes

Mode Steps Description Best For
💬 Chat Mode 1-5 Interactive iteration with AI chat Refining ideas, requirements, architecture
📝 Manual Mode 6 Fill out markdown manually Creating detailed task lists
🤖 Agent Mode 7-8 Automated processing with AI agents Task review, test generation

📋 Step-by-Step Process

  1. Initialize project: aisdlc init
  2. Start new feature: aisdlc new "Your feature idea"
  3. Progress through steps: aisdlc next (repeat for each step)
  4. Check status: aisdlc status
  5. Complete feature: aisdlc done

📋 Command Reference

Command Description Example
aisdlc init Initialize AI-SDLC in current directory aisdlc init
aisdlc new <idea> Start new feature with idea description aisdlc new "Add user authentication"
aisdlc next Progress to next step in workflow aisdlc next
aisdlc status Show current project status aisdlc status
aisdlc done Archive completed feature to done/ aisdlc done
aisdlc --help Show help information aisdlc --help

Working with steps:

  • Each step creates a markdown file in doing/<feature-slug>/
  • Fill out the generated markdown before running aisdlc next
  • AI agents process your input and generate the next step
  • Alternative: Use prompt templates directly with any AI chat interface

⚙️ How It Works

flowchart TD
    I[01-idea]-->P1[02-prd]-->P2[03-prd-plus]-->A[04-architecture]
    A-->SP[05-system-patterns]-->T[06-tasks]-->TP[07-tasks-plus]-->TESTS[08-tests]

    %% Iteration loop for steps 1-5
    CHAT[💬 Iterate with AI Chat]
    I -.-> CHAT
    P1 -.-> CHAT
    P2 -.-> CHAT
    A -.-> CHAT
    SP -.-> CHAT
    CHAT -.-> I
    CHAT -.-> P1
    CHAT -.-> P2
    CHAT -.-> A
    CHAT -.-> SP

    %% Agent mode for steps 7-8
    AGENT[🤖 Use AI Agent Mode]
    TP --- AGENT
    TESTS --- AGENT
Loading

Workflow modes explained:

  • Steps 1-5 (💬 Chat Mode): You manually fill out markdown files and iterate with AI chat to refine your ideas, requirements, and architecture
  • Step 6 (Tasks): Manual step to create implementation tasks
  • Steps 7-8 (🤖 Agent Mode): Automated processing using AI agents for task review and test generation

Running aisdlc next:

  1. Reads the previous markdown file
  2. Merges it into the prompt for the next step
  3. For steps 7-8: Calls AI agent (requires compatible AI editor or API)
  4. Writes the new markdown and bumps .aisdlc.lock

Using prompts manually:

  1. Copy the appropriate prompt from prompts/ directory
  2. Paste your previous step's content into the <prev_step> placeholder
  3. Use with any AI chat interface (Cursor, Claude, ChatGPT, etc.)
  4. Save the output as the next step's markdown file

🏗️ Project Structure

.
├── ai_sdlc/                # main Python package
│   ├── cli.py              # entry point for `aisdlc`
│   ├── commands/           # sub-commands: init | new | next | status | done
│   ├── scaffold_template/  # default templates for new projects
│   └── utils.py            # shared helpers
├── prompts/                # LLM templates for each SDLC step
│   ├── 0.idea.instructions.md   # initial idea analysis
│   ├── 1.prd.instructions.md    # product requirements
│   ├── 2.prd-plus.instructions.md # enhanced requirements
│   ├── 3.system-template.instructions.md # system architecture
│   ├── 4.systems-patterns.instructions.md # design patterns
│   ├── 5.tasks.instructions.md  # implementation tasks
│   ├── 6.tasks-plus.instructions.md # task list review & handoff preparation
│   └── 7.tests.instructions.md         # test generation
├── tests/                  # pytest suite (unit + integration)
│   ├── unit/               # unit tests
│   └── integration/        # integration tests
├── doing/                  # active features (created by init)
├── done/                   # completed features (created by init)
├── .aisdlc                 # TOML config (ordered steps, dirs, diagram)
├── .aisdlc.lock           # current workflow state
├── pyproject.toml          # build + dependency metadata
├── CHANGELOG.md            # version history
└── README.md               # you are here

🏛️ Architecture

AI-SDLC is built around a simple but powerful concept: markdown-driven development with AI assistance.

Core Components

  1. CLI Interface (ai_sdlc/cli.py)

    • Entry point for all commands
    • Handles argument parsing and command routing

  2. Command System (ai_sdlc/commands/)

    • Modular command structure
    • Each command handles a specific workflow step

  3. Prompt Templates (prompts/)

    • LLM prompts for each SDLC step
    • Structured to guide AI through development process

  4. State Management

    • .aisdlc - Project configuration
    • .aisdlc.lock - Current workflow state
    • File-based state tracking

Workflow Engine

The workflow engine processes each step by:

  1. Reading the previous step's output
  2. Merging it with the appropriate prompt template
  3. Calling Cursor agent to generate next step
  4. Writing output and updating state

🔧 Technology Stack

Layer Main libs / tools Why
CLI Python 3.13, click-style argparse (stdlib) modern syntax, zero deps runtime
Package mgmt uv fast, lock-file driven reproducibility
Dev tooling Ruff, Pyright, pytest lint + format, type-check, tests
AI Integration Pluggable AI agents works with any AI editor or API
Packaging setuptools, PEP 621 metadata slim install

pyproject.toml excerpt

[project]
name = "ai-sdlc"
requires-python = ">=3.13.0"

[project.optional-dependencies]
dev = [
  "pytest>=7.0",
  "pytest-mock>=3.0",
  "ruff>=0.0.292",
  "pyright>=1.1.350"
]

🛠️ Development Setup

Prerequisites

Tool Install (macOS example)
Python 3.13+ brew install [email protected] or pyenv install 3.13.0
uv curl -LsSf https://astral.sh/uv/install | sh
AI Editor (optional) Cursor, VS Code with AI extensions, or any AI interface
Node 20 + pnpm* brew install node pnpm (only if you touch TS helpers)

Setup

git clone https://github.com/your-org/ai-sdlc.git
cd ai-sdlc
uv venv && source .venv/bin/activate
uv sync --all-features        # installs runtime + dev deps from uv.lock
pre-commit install            # optional hooks

🧪 Testing & Quality

# install dev extras
uv pip install -e .[dev]

# lint + format
uv run ruff check ai_sdlc tests
uv run ruff format ai_sdlc tests

# type-check
uv run pyright

# run all tests
uv run pytest

# run specific test types
uv run pytest tests/unit/      # unit tests only
uv run pytest tests/integration/  # integration tests only

Integration tests spin up a temp project dir and exercise the CLI flow.

🔧 Troubleshooting

Common Issues

"AI agent command not found"

  • AI-SDLC generates prompts that work with any AI tool
  • No specific AI tool installation required
  • Use the generated prompts with your preferred AI assistant

"Permission denied" errors

  • Check file permissions in your project directory
  • Ensure you have write access to the current directory

"Invalid .aisdlc configuration"

  • Verify your .aisdlc file has valid TOML syntax
  • Run aisdlc init to regenerate default configuration

"Lock file corruption"

  • Delete .aisdlc.lock and run aisdlc status to regenerate
  • The tool handles corrupted lock files gracefully

Getting Help

  1. Check the command help: aisdlc --help or aisdlc <command> --help
  2. Review the CHANGELOG.md for recent changes
  3. Open an issue on GitHub with:
    • Your OS and Python version
    • Full error message
    • Steps to reproduce

🗺️ Roadmap

Planned Features

  • Pluggable AI providers – flag --model to swap GPT-4o, Claude, Gemini, etc.
  • 09-release-plan step (CI/CD & deployment playbook)
  • Context-window management (summaries / embeddings for large projects)
  • Repomix integration for giant monorepos
  • Template customization - custom prompt templates per project
  • Parallel workflows - multiple features in development simultaneously
  • Enhanced AI provider integrations (OpenAI API, Anthropic API, etc.)

Future Considerations

  • Web UI for workflow visualization
  • Team collaboration features
  • Metrics and analytics for development velocity
  • Integration with project management tools

🤝 Contributing

We welcome contributions! Here's how to get started:

Development Workflow

  1. Fork and clone the repository
  2. Create a feature branch: git checkout -b feat/your-feature
  3. Make your changes with tests
  4. Run quality checks: ruff check, pyright, pytest must pass
  5. Open a PR with a clear description

Code Standards

  • Follow existing code style (enforced by Ruff)
  • Add tests for new functionality
  • Update documentation for user-facing changes
  • Keep commits atomic and well-described

Areas for Contribution

  • Bug fixes - Check GitHub issues
  • Documentation - Improve clarity and examples
  • Testing - Expand test coverage
  • Features - See roadmap above
  • Prompt engineering - Improve LLM prompt templates

📄 License

MIT © 2025 Parker Rex See LICENSE for details.

About

No description, website, or topics provided.

Resources

Readme

License

View license
Activity

Stars

174 stars

Watchers

10 watching

Forks

48 forks
Report repository

Releases 3

v0.6.3 Latest
Jun 4, 2025
+ 2 releases

Packages

No packages published

Contributors 2

Languages