stvlynn/agentic-coding

This repository is a starter template for software projects where most of the coding is done by an AI agent rather than a human writing code by hand. It is not tied to any particular programming language or framework: the idea is that you copy the template and fill in details specific to your own project. The template's main contribution is its documentation structure. There is a file called AGENTS.md (also available as CLAUDE.md) that tells any AI agent opening the project what the ground rules are, how the documentation is organized, and how it should evolve over time. The agent reads this file first before writing any code. The docs folder then branches into sections covering the project overview, frontend conventions (using a pattern called Feature-Sliced Design), backend conventions (using Domain-Driven Design layering), CI/CD and deployment guides, testing expectations, and a log of architecture decisions. For a human getting started, the intended steps are to copy the repository, update the placeholder content in the architecture document, and then begin adding code under the structure the docs describe. The deploy folder contains Docker and Kubernetes configuration that is meant to be adapted for each project. The repository is small and the README is brief. It is a skeleton for a workflow, not a finished tool. The README does not mention a license.