Oh My Multica — Development skill for Claude Code

oh-my-multica

[](https://github.com/xiaohei-info/oh-my-multica/actions/workflows/ci.yml) [](https://github.com/xiaohei-info/oh-my-multica/releases/tag/v1.0.0) [](./LICENSE)

[English](README.md) | [简体中文](README.zh-CN.md)

**A production-grade AI software delivery system built on [Multica](https://github.com/multica-ai/multica).**

[Multica](https://github.com/multica-ai/multica) brings Coding Agents such as Claude Code and Codex into unified workspaces, issues, task queues, and local runtimes. Agents can receive work, report progress and blockers, and operate like team members, while teams centrally manage execution machines and reusable Skills. It provides the foundation for task assignment, lifecycle management, runtime scheduling, and state tracking in multi-Agent collaboration.

Multica focuses primarily on Agent execution and collaboration, but it does not define a complete engineering delivery process for a software project: how a requirement enters design, how executable acceptance criteria are formed, how multiple development tasks run in parallel according to dependencies, what evidence proves an implementation is correct, who performs independent review, when a change may be merged, or how failures recover.

oh-my-multica builds on Multica's excellent mechanism design and adds a more complete software engineering delivery control layer. It advances a requirement into a software change that has passed design, development, verification, review, merge, and final acceptance.

**Multica is a complete Agent Runtime task platform that manages how Agents work, while oh-my-multica manages how software is delivered on top of it.**

Why oh-my-multica

Coding Agents are already good at writing code. The hard problems usually appear outside the code: requirements