sdlc-agents 0.1.0

A LangGraph-based multi-agent SDLC orchestration library

sdlc-agents

AI-powered multi-agent CLI for code-change workflows, built with LangGraph.

Agent Workflow

START → Planner → Research → Coder → Code Reviewer → Documentation → Final Report → END

Agent	Role
PlannerAgent	Gathers context and creates an execution plan
ResearchAgent	Produces actionable implementation tasks
CoderAgent	Plans safe code changes with snippets
CodeReviewerAgent	Reviews for correctness, security, quality
DocumentationAgent	Generates docs and release notes

Installation

pip install -e ".[dev]"

CLI Usage

sdlc-agent version
sdlc-agent dry-run "Add logging to auth module"
sdlc-agent run "Add logging to auth module"

Environment Variables

Variable	Default	Description
LLM_PROVIDER	openai	Provider: mock, openai, github_models
OPENAI_API_KEY		OpenAI API key
MODEL_NAME	gpt-4o-mini	Model name
TEMPERATURE	0.2	LLM temperature
GITHUB_TOKEN		GitHub token for GitHub Models
GITHUB_MODELS_BASE_URL	https://models.github.ai/inference	GitHub Models endpoint

Copy .env.example to .env and fill in your values.

Dry-Run Example (No LLM)

sdlc-agent dry-run "Add error handling to payment service"

Real Run with GitHub Models

$env:LLM_PROVIDER="github_models"
$env:GITHUB_TOKEN="your_token"
sdlc-agent run "Add error handling to payment service"

Testing

python -m pytest tests/ -v