My main focus and hardcore outputs are concentrated in the open-source organization @breakevery, where I am building the underlying infrastructure. 我的主要精力与硬核产出都在开源组织 @breakevery，这里存放着我正在构建的底层基础设施。

"Freedom within the boundary, auditing outside the AI, root privileges to humans." “边界之内自由，AI 之外审计，root 权限归人。”

• What is it / 是什么: A desktop application that grants AI virtual kernel-level privileges within a QEMU RISC-V bare-metal sandbox, while humans retain ultimate root control. 一个让 AI 在 QEMU RISC-V 裸机沙盒中拥有虚拟内核级权限、同时人类保留最终 root 控制权的桌面应用。
• Core Mechanism / 核心机制: AI can write/compile RISC-V assembly, read serial ports, and iterate; all actions are appended to an append-only hash chain, making them auditable and rollback-capable. AI 可编写/编译 RISC-V 汇编、读取串口并迭代；所有动作进入只追加的哈希链，可审计、可回滚。
• Tech Stack / 技术栈: Rust · Tauri · QEMU · RISC-V · Apache-2.0
• Status / 状态: v1.0 Core Completed (2026-10) / v1.0 主体完成

👉 Enter the RiscDom repository to view architecture and design docs 👉 进入 RiscDom 仓库，查看架构与设计文档

Other related repositories / 其他相关仓库：

• riscdom-server — Standalone control plane (HTTP API + minimal Web status page) / 单机控制平面（HTTP API + 最小 Web 状态页）
• riscdom-adminapp — Desktop (Tauri) and Mobile (Android) admin app / 桌面（Tauri）与移动（Android）管理程序

I am looking for partners interested in low-level development (Rust / RISC-V) or AI runtime infrastructure. If you have ideas about "enabling AI to collaborate safely in an auditable sandbox", feel free to open an Issue or PR at breakevery. 我正在寻找对 底层开发（Rust / RISC-V） 或 AI 运行时基础设施 感兴趣的伙伴。如果你对“让 AI 在可审计沙盒中安全协作”这件事有想法，欢迎来 breakevery 提 Issue 或 PR。

• 📧 Email: [email protected]
• 🐦 X (Twitter): @zhanchenyi10

Stay curious, keep creating. / 保持好奇，持续创造。