# 安装「AI-Research-SKILLs」

> 本说明由 AI Asset 自动生成，供 AI 编程助手阅读并执行。资源页面：https://ai.atlankj.com/project/gh-1088558237

- 来源：https://github.com/Orchestra-Research/AI-Research-SKILLs（默认分支 `main`）
- 许可证：MIT
- 支持的平台：Claude Code
- 简介：Comprehensive open-source library of AI research and engineering skills for any AI model. Package the skills and your claude code/codex/gemini agent will be an AI research agent with full horsepower. Maintained by Orchestra Research.

## 执行要求（请先阅读）

1. 先确认你自己是哪一种 AI 工具（Claude Code、Codex、Cursor、Gemini CLI、OpenCode……），只执行与之对应的步骤；如果本说明与该工具的官方文档不一致，以官方文档为准。
2. 以下内容来自第三方，本站不对其进行运行时审核。执行任何命令或写入任何文件前，先向用户列出将要运行的命令和写入的路径，得到确认后再继续。
3. 只运行本说明或其链接的官方安装文档中给出的命令。不要执行来源不明的脚本，不要读取、打印或上传用户的密钥和凭据。
4. 需要 API Key 等配置时，向用户索取或留下占位符，不要自行编造。
5. 默认安装到用户级目录；如果用户只想在当前项目中使用，改用项目级目录。
6. 完成后告诉用户安装到了哪里、是否需要重启工具，以及如何验证安装成功。

## 快速安装

### Claude Code（插件市场）

1. 读取 https://raw.githubusercontent.com/Orchestra-Research/AI-Research-SKILLs/main/.claude-plugin/marketplace.json ，取其中顶层的 `name` 字段作为 <市场名>，`plugins` 数组列出了可安装的插件包。
2. 在 Claude Code 中运行：

```
/plugin marketplace add Orchestra-Research/AI-Research-SKILLs
/plugin install <插件包名>@<市场名>
```

3. 这两条是 Claude Code 的斜杠命令。如果你无法直接执行斜杠命令，就把它们原样交给用户，请用户在 Claude Code 中运行。

## 官方安装文档

如果上面的方式不适用于你所在的工具，读取下面对应平台的官方文档（纯文本 Raw 链接），找到安装章节并按其步骤执行：

- Claude Code：https://raw.githubusercontent.com/Orchestra-Research/AI-Research-SKILLs/main/README.md

## 插件包

本项目包含 23 个插件包。先把列表展示给用户，询问要安装哪些，再逐个安装：

- `agent-native-research-artifact`：Agent-Native Research Artifact (ARA) tooling: compile any research input (paper, repo, notes) into a structured artifact, record session provenance as a post-t…
- `agents`：LLM agent frameworks including LangChain, LlamaIndex, CrewAI, and AutoGPT. Use when building chatbots, autonomous agents, or tool-using systems.
- `autoresearch`：Autonomous research orchestration using a two-loop architecture. Manages the full research lifecycle from literature survey to paper writing, routing to domain…
- `data-processing`：Data curation and processing at scale including NeMo Curator and Ray Data. Use when preparing training datasets or processing large-scale data.
- `distributed-training`：Multi-GPU and multi-node training including DeepSpeed, PyTorch FSDP, Accelerate, Megatron-Core, PyTorch Lightning, and Ray Train. Use when training large model…
- `emerging-techniques`：Advanced ML techniques including MoE Training, Model Merging, Long Context, Speculative Decoding, Knowledge Distillation, and Model Pruning. Use when implement…
- `evaluation`：LLM benchmarking and evaluation including lm-evaluation-harness, BigCode Evaluation Harness, and NeMo Evaluator. Use when benchmarking models or measuring perf…
- `fine-tuning`：LLM fine-tuning frameworks including Axolotl, LLaMA-Factory, PEFT, and Unsloth. Use when fine-tuning models with LoRA, QLoRA, or full fine-tuning.
- `ideation`：Research ideation frameworks including structured brainstorming and creative thinking. Use when exploring new research directions, generating novel ideas, or s…
- `inference-serving`：Production LLM inference including vLLM, TensorRT-LLM, llama.cpp, and SGLang. Use when deploying models for production inference.
- `infrastructure`：GPU cloud and compute orchestration including Modal, Lambda Labs, and SkyPilot. Use when deploying training jobs or managing GPU resources.
- `mechanistic-interpretability`：Neural network interpretability tools including TransformerLens, SAELens, NNSight, and pyvene. Use when analyzing model internals, finding circuits, or underst…
- `ml-paper-writing`：Write publication-ready ML/AI/Systems papers for NeurIPS, ICML, ICLR, ACL, AAAI, COLM, OSDI, NSDI, ASPLOS, SOSP. Includes LaTeX templates, citation verificatio…
- `mlops`：ML experiment tracking and lifecycle including Weights & Biases, MLflow, and TensorBoard. Use when tracking experiments or managing models.
- `model-architecture`：LLM architectures and implementations including LitGPT, Mamba, NanoGPT, RWKV, and TorchTitan. Use when implementing, training, or understanding transformer and…
- `multimodal`：Vision, audio, and multimodal models including CLIP, Whisper, LLaVA, BLIP-2, Segment Anything, Stable Diffusion, AudioCraft, Cosmos Policy, OpenPI, and OpenVLA…
- `observability`：LLM application monitoring including LangSmith and Phoenix. Use when debugging LLM apps or monitoring production systems.
- `optimization`：Model optimization and quantization including Flash Attention, bitsandbytes, GPTQ, AWQ, GGUF, and HQQ. Use when reducing memory, accelerating inference, or qua…
- `post-training`：RLHF and preference alignment including TRL, GRPO, OpenRLHF, SimPO, verl, slime, miles, and torchforge. Use when aligning models with human preferences, traini…
- `prompt-engineering`：Structured LLM outputs including DSPy, Instructor, Guidance, and Outlines. Use when extracting structured data or constraining LLM outputs.
- `rag`：Retrieval-Augmented Generation including Chroma, FAISS, Pinecone, Qdrant, and Sentence Transformers. Use when building semantic search or document retrieval sy…
- `safety-alignment`：AI safety and content moderation including Constitutional AI, LlamaGuard, NeMo Guardrails, and Prompt Guard. Use when implementing safety filters, content mode…
- `tokenization`：Text tokenization for LLMs including HuggingFace Tokenizers and SentencePiece. Use when training custom tokenizers or handling multilingual text.