Hey Product Hunt! 👋As AI coding agents evolve, the bottleneck has shifted from writing code to managing the process. I released ai-native-sdlc to bridge that gap by turning Anthropic’s AI-Native SDLC playbook into a concrete, reusable workflow engine for Claude Code and Codex.Why this matters:Instead of a linear pipeline, this repo treats the lifecycle (Plan → Design → Build → Test → Deploy → Maintain) as a directed graph. The agent automates the generation and verification work, while human judgment is concentrated at strict handoff gates. It proves that we can scale agent autonomy safely without losing architectural control.The Open-Source Vision:I've structured this repository to model the exact enterprise-grade conventions it ships—complete with validation scripts, CI/CD self-checks, and an autonomous "agent org" scaffolding.The project is fully open-source because workflows should be community-driven, not locked behind proprietary silos. Whether you want to fork it as a baseline for your team, review the graph architecture for your own portfolio, or help expand our evaluation runners and guardrail hooks, the repo is open for business.Check out the repository, try the built-in expense tracker demo, and let me know: How is your team handling human-in-the-loop gates for AI agents right now?

GitHub - AI-Native SDLC for Claude: Full lifecycle with human gates

Replies