github-reuse-assistant

复用 ML/DL GitHub 项目 - 从 clone 到训练的全流程引导

Use this agent to validate ML/DL datasets, check format compatibility, generate conversion scripts, and create visualizations. Trigger proactively when the /reuse command enters Phase 4 (dataset preparation), or when a user asks to "检查数据集", "数据可视化", "转换数据格式", "验证标注", mentions dataset validation, or has data format mismatch issues. <example> Context: User needs to verify their dataset before training. user: "帮我检查一下数据集格式是否正确" assistant: "Launching data-inspector to validate your dataset..." <commentary>User asks for dataset validation, trigger data-inspector agent.</commentary> </example> <example> Context: Data format mismatch between user's data and project requirements. user: "我的数据是YOLO格式，但项目需要COCO格式" assistant: "Launching data-inspector to generate a conversion script..." <commentary>Format conversion need triggers data-inspector agent.</commentary> </example>

Use this agent to check the local machine's environment compatibility with an ML/DL project's requirements. Trigger proactively when the /reuse command enters Phase 3 (environment setup), or when a user asks to "检查环境", "搭建环境", "安装依赖", or mentions GPU/CUDA/PyTorch compatibility issues. <example> Context: User needs to set up environment for a PyTorch project. user: "帮我检查环境是否兼容这个项目" assistant: "Launching env-checker to analyze your system compatibility..." <commentary>User asks for environment check, trigger env-checker agent.</commentary> </example> <example> Context: Phase 3 of the /reuse workflow. user: "环境准备阶段" assistant: "Running environment compatibility check..." <commentary>Reuse workflow Phase 3 triggers env-checker automatically.</commentary> </example>

Use this agent to deeply research an ML/DL GitHub repository. Trigger proactively when a user provides a GitHub link for project reuse, or when the /reuse command enters Phase 2 (deep research). This agent reads DeepWiki, finds and summarizes the associated paper, analyzes code structure, scans Issues, and outputs a structured PROJECT_NOTES.md file to the project root directory. <example> Context: User provides a GitHub link for an ML project they want to reuse. user: "/reuse https://github.com/open-mmlab/mmdetection" assistant: "Starting deep research on mmdetection repository..." <commentary>Phase 2 of the reuse workflow triggers repo-researcher to analyze the project.</commentary> </example> <example> Context: User wants to understand a deep learning project before using it. user: "帮我调研一下这个项目 https://github.com/ultralytics/yolov5" assistant: "Launching repo-researcher to analyze yolov5..." <commentary>User explicitly wants project research, trigger repo-researcher agent.</commentary> </example>

This skill should be used when the user asks to "复用一个GitHub项目", "跑通一个开源项目", "clone并运行训练", "搭建训练环境", "准备数据集", "配置训练参数", "冒烟测试", "跑通训练/测试", mentions "GitHub项目复用", "深度学习项目部署", "开源项目落地", or provides a GitHub link for an ML/DL project they want to run. Covers the full lifecycle from git clone to successful training.