vibe-coding-ros2
skill
Warn
Health Warn
- No license — Repository has no license file
- Description — Repository has a description
- Active repo — Last push 0 days ago
- Low visibility — Only 5 GitHub stars
Code Pass
- Code scan — Scanned 3 files during light audit, no dangerous patterns found
Permissions Pass
- Permissions — No dangerous permissions requested
Purpose
This project is a development guide and methodology repository for applying AI-assisted pair programming ("VibeCoding") to ROS2 robot deployments. It provides workflows, best practices, and architectural principles rather than executable software.
Security Assessment
The repository consists entirely of documentation and guideline files (written primarily in Chinese). A code scan of 3 files found no dangerous patterns, no hardcoded secrets, and no dangerous permission requests. Since this is a collection of Markdown guides rather than an executable tool, it does not execute shell commands or access sensitive data. Overall risk is rated as Low.
Quality Assessment
The project is actively maintained, with its last push occurring today. However, there are significant quality concerns regarding maturity and licensing. It has very low visibility with only 5 GitHub stars, indicating minimal community review or trust. More importantly, the repository lacks a license file. Without a defined license, the included code, configurations, and text remain under exclusive copyright by default, meaning other developers technically do not have legal permission to use, copy, or modify the content. The project is also tagged as an early "alpha" release.
Verdict
Safe to use for reading and personal reference, but use with caution if you intend to implement or copy the provided configurations into commercial projects due to the missing open-source license.
This project is a development guide and methodology repository for applying AI-assisted pair programming ("VibeCoding") to ROS2 robot deployments. It provides workflows, best practices, and architectural principles rather than executable software.
Security Assessment
The repository consists entirely of documentation and guideline files (written primarily in Chinese). A code scan of 3 files found no dangerous patterns, no hardcoded secrets, and no dangerous permission requests. Since this is a collection of Markdown guides rather than an executable tool, it does not execute shell commands or access sensitive data. Overall risk is rated as Low.
Quality Assessment
The project is actively maintained, with its last push occurring today. However, there are significant quality concerns regarding maturity and licensing. It has very low visibility with only 5 GitHub stars, indicating minimal community review or trust. More importantly, the repository lacks a license file. Without a defined license, the included code, configurations, and text remain under exclusive copyright by default, meaning other developers technically do not have legal permission to use, copy, or modify the content. The project is also tagged as an early "alpha" release.
Verdict
Safe to use for reading and personal reference, but use with caution if you intend to implement or copy the provided configurations into commercial projects due to the missing open-source license.
Vibe-coding development guide for ROS Humble deployment on multiple robot platforms.
README.md
ROS2 机器人开发 VibeCoding 指南
本指南将 VibeCoding 方法论应用于 ROS2 机器人项目开发,涵盖从环境配置、节点开发到部署调试的全流程。
📊 仓库活跃度
🖼️ 概览
ROS2 VibeCoding 是将 AI 结对编程与机器人开发相结合的终极工作流程。ROS2 项目具有独特的复杂性:
- 多节点通信 — Topic/Service/Action 多种通信机制
- 硬件驱动集成 — 相机、激光雷达、IMU、电机驱动
- 实时性要求 — 控制系统对延迟敏感
- 跨平台部署 — x86 开发机 → ARM 部署机(OrinNX/AGX/JetBot)
- 生态依赖 — Navigation2、MoveIt、ROS2 工业应用
核心理念: 规划驱动 + 上下文固定 + AI 结对执行,让机器人代码从「难以维护」变成「可审计、可迭代」。
一句话: VibeCoding = 规划驱动 + 上下文固定 + AI 结对执行
🔑 元方法论 (Meta-Methodology)
递归优化循环
VibeCoding 的核心是一个能自我优化的 AI 系统:
创生 (Bootstrap) → 自省与进化 (Self-Correction) → 创造 (Generation) → 循环与飞跃 (Recursive Loop)
α-Ω 提示词系统
- α-提示词 (生成器): 负责生成其他提示词或技能
- Ω-提示词 (优化器): 负责优化其他提示词或技能
🧭 道
- 凡是 AI 能做的,就不要人工做 — 重复性代码生成、配置模板、文档编写
- 一切问题问 AI — 先问 AI,AI 解决不了再人工
- 上下文是 VibeCoding 的第一性要素 — 垃圾进,垃圾出
- 系统性思考:节点、通信、硬件三维度 — 机器人系统的核心是节点间通信和硬件交互
- 数据与函数即是编程的一切 — ROS2 消息即数据,节点即函数
- 输入→处理→输出刻画整个过程 — 节点的本质
- 先结构,后代码 — 规划好框架,不然后面技术债还不完
- 奥卡姆剃刀定理 — 如无必要,勿增代码
- 帕累托法则 — 关注重要的 20%
- 重复,多试几次 — 实在不行重新开个窗口
- 专注 — 极致的专注可以击穿代码,一次只做一件事
🧩 法
- 一句话目标 + 非目标 — 明确本次开发要做什么、不做什么
- 正交性 — 功能不要太重复,模块职责清晰
- 能抄不写 — 不重复造轮子,先问 AI 有没有合适的仓库
- 官方文档优先 — 先把官方文档喂给 AI
- 按职责拆模块 — 感知、规划、控制、执行分离
- 接口先行,实现后补 — 先定义 msg/srv/action,再实现节点
- 一次只改一个模块 — 降低耦合风险
- 文档即上下文 — 不是事后补,而是开发时同步写
🛠️ 术
- 明确写清: 能改什么、不能改什么
- Debug 法则: 预期 vs 实际 + 最小复现 + 关键日志
- 测试分层: 单元测试、集成测试、硬件测试
- 代码一多就切会话 — 保持上下文清晰
- 逆向思考 — 从需求逆向构建代码
嵌入式平台 Debug(容器/仿真 → 实机部署)
| 阶段 | 环境 | Debug 手段 |
|---|---|---|
| 容器内 Debug | x86 Docker / ARM Docker | ros2 topic echo、rqt、ros2 bag 录制分析 |
| 仿真器验证 | Gazebo / Ignition | 仿真器内节点通信监控、ros2 doctor |
| 实机部署 | Jetson OrinNX / RDK-X5 / 旭日X3 | SSH 远程连接、网络抓包(tcpdump)、远程 GDB / GDBServer |
| 跨平台对比 | 容器 vs 实机 | 确认时钟源(/use_sim_time)、话题带宽、延迟差异 |
工作流: 容器内开发调试 → 仿真器验证功能 → 交叉编译部署到实机 → SSH 远程 Debug
关键点: 实机与容器环境的差异主要在时钟源(仿真用/use_sim_time)、硬件驱动依赖、网络配置
📋 工具链
集成开发环境 (IDE) & 终端
| 工具 | 用途 | 备注 |
|---|---|---|
| VS Code | 主 IDE | + ROS 扩展 |
| Cursor | AI 增强开发 | VS Code 分支 |
| Docker | 容器化开发 | ARM 交叉编译必备 |
| Terminator | 多终端管理 | 分屏操作 |
| tmux | 会话保持 | 服务器开发必备 |
| Warp | AI 终端 | 现代化终端 |
ROS2 开发工具
# 核心工具
sudo apt install -y ros-humble-ros2launch ros-humble-ros2run ros-humble-ros2pkg
sudo apt install -y ros-humble-rqt* ros-humble-rviz2
sudo apt install -y python3-colcon-common-extensions
# 调试工具
sudo apt install -y ros-humble-ros2bag ros-humble-ros2cli
sudo apt install -y ros-humble-topic-monitoring
sudo apt install -y ros-humble-diagnostics
# 仿真工具
sudo apt install -y ros-humble-gazebo-ros-pkgs ros-humble-turtlebot3-*
AI 模型 & 服务 (2026年4月 SOTA)
| 梯队 | 模型 | 适用场景 |
|---|---|---|
| 第一梯队 | Claude 4.7 Opus / GPT-5.4 Pro / Gemini 3.1 Ultra / Kimi K3 / GLM-5 | 机器人架构设计、运动控制算法、导航规划、感知融合、多模态交互、仿真调试 |
| 第二梯队 | Codex 5.5-max / Grok-3 / Qwen 3.5 / Doubao-Pro / Seed-Code 2.0 / DeepSeek V3.2 / Llama 4 | 节点代码生成、传感器驱动、控制器实现、SLAM算法、ROS2集成、边缘部署 |
| 第三梯队 | Mistral Large 3 / Gemini 3.1 Flash / Minimax-M2.7 / Hunyuan-T1 / Ernie-4.5 / Qwen 3 MoE / DeepSeek-Coder-V3 / SWE-1-Max / Tongyi-Qwen-VL / Yi-VL | 视觉里程计、目标检测、图像分割、模型量化加速、代码补全、Bug修复 |
2026年4月要点: GPT-5 / Claude 4.7 Opus / Gemini 3 Ultra 引领多模态融合,Kimi K3 / GLM-5 / Qwen 3 在机器人开发场景达到国际第一梯队水平
⚙️ 核心技能 (Skills)
本项目定义了以下专用技能:
| 技能 | 描述 | 触发词 |
|---|---|---|
| ros2-package-generator | ROS2 功能包生成 | "创建 ROS2 包" / "generate ros2 package" |
| arm64-cross-compile | ARM64 交叉编译 | "交叉编译" / "ARM 编译" |
| ros2-debugging | ROS2 调试技能 | "调试 ROS2" / "ros2 debug" |
| ros2-docker-dev | Docker 开发环境 | "ROS2 Docker" |
| ros2-navigation | 导航系统集成 | "navigation" / "nav2" |
| ros2-perception | 感知节点开发 | "perception" / "detect" |
详见 skills/ 目录。
📚 文档结构
vibe-coding-ros2/
├── README.md # 本指南入口
├── publish.sh # 发布脚本
├── agents/ # Agent 主目录
│ ├── documents/ # 方法论与原则文档
│ │ ├── Methodology_and_Principles/
│ │ ├── Templates_and_Resources/
│ │ └── Tutorials_and_Guides/
│ ├── prompts/ # AI 提示词库
│ │ ├── coding_prompts/
│ │ ├── system_prompts/
│ │ └── user_prompts/
│ ├── robots/ # 机器人类型指南
│ │ ├── common/
│ │ ├── wheeled_vehicle/
│ │ ├── quadruped/
│ │ ├── manipulator/
│ │ ├── humanoid/
│ │ └── multi_rotor_uav/
│ ├── skills/ # 技能定义
│ │ ├── ros2-package-generator/SKILL.md
│ │ ├── arm64-cross-compile/SKILL.md
│ │ └── ros2-debugging/SKILL.md
│ └── memory-bank/ # Memory Bank 模板
│ ├── project-context.md
│ ├── implementation-plan.md
│ └── progress.md
└── i18n/ # 多语言文档
├── README.md
├── en/
│ ├── README.md
│ └── CHANGELOG.md
└── zh-CN/
├── README.md
└── CHANGELOG.md
🚀 快速入门
Step 0: 一键初始化 AI 规则与技能导入
./init-agent.sh --target all
说明:
--target all: 同时生成 VS Code Copilot 与 Cursor 配置--target copilot: 仅生成 Copilot 配置--target cursor: 仅生成 Cursor 配置
脚本会自动生成:
- Copilot 指令文件:
.github/copilot-instructions.md - Cursor 规则文件:
.cursor/rules/vibe-coding-ros2.mdc - MCP 配置模板:
.vscode/mcp.json与.cursor/mcp.json - 全量技能索引:
agents/generated/skill-index.md - 项目准则索引:
agents/generated/context-index.md
Step 1: 选择 AI 工具
推荐使用 Claude Opus 4.5 (VS Code Copilot) 或 Codex CLI
Step 2: 创建 Memory Bank
mkdir -p memory-bank
# 复制模板或让 AI 生成
Step 3: 定义项目上下文
# 项目上下文 - [项目名称]
## 1. 项目概述
- 机器人平台: [ Jetson OrinNX / AGX / x86 ]
- ROS2 发行版: Humble
- 核心功能: [列表]
## 2. 技术栈
- 语言: C++ / Python
- 关键依赖: OpenCV, TensorRT, Navigation2...
## 3. 包结构
[包列表及职责]
## 4. 消息流
[传感器] → [感知] → [规划] → [控制] → [执行器]
Step 4: AI 结对开发
"请阅读 memory-bank 所有文档,帮我实现实施计划第 1 步"
📖 详细文档
| 分类 | 文档 | 内容 |
|---|---|---|
| 入门 | 开发经验 | ROS2 开发坑点与经验 |
| 方法论 | 架构原则 | 节点设计、通信模式 |
| 模板 | 项目模板 | 标准项目结构 |
| 模板 | Memory Bank | 上下文文档模板 |
| 指南 | 调试指南 | rqt, bag, launch 调试 |
| 指南 | Docker 配置 | 容器开发环境 |
| 指南 | 交叉编译 | ARM64 编译 |
| 指南 | Agent 导入指南 | AI 工具加载技能与提示 |
🗺️ 路线图
gantt
title ROS2 VibeCoding 发展路线图
dateFormat YYYY-MM
section 近期 (2026)
完善调试技能: 2026-04, 30d
导航集成提示词: 2026-04, 20d
section 中期 (2026 Q2)
感知节点模板库: 2026-05, 30d
多机协同指导: 2026-05, 20d
section 远期 (2026 Q3)
仿真集成指南: 2026-06, 30d
工业应用模板: 2026-07, 30d
🤝 参与贡献
欢迎提交 Issue 和 Pull Request!
本项目参考 vibe-coding-cn 构建
Reviews (0)
Sign in to leave a review.
Leave a reviewNo results found