zechenzhangAGICommunity·9 Agent Skills Included

AI-research-SKILLs

Ninety-eight skills for autonomous AI research, idea to paper

Guides AI agents through the full machine learning research lifecycle: literature survey, ideation, data curation, fine-tuning, distributed training, evaluation, inference serving, and paper writing. Eliminates manual framework setup, undocumented debugging, and citation hallucination with production-ready workflows and verified citation APIs. Includes an autoresearch orchestrator that routes tasks to domain skills and runs experiments continuously until a publishable paper is produced.

11,537|842

↗

npx skills add zechenzhangAGI/AI-research-SKILLs --all -g -y

Available:

Tells the AI agent how the 98 skills are organized across the research lifecycle and enforces the SKILL.md quality and structure standards it must follow when using or extending them.

All Skills in This Repository (9)

Pure Emerald Level Indicators

📦 In Repo

zechenzhangAGI

speculative-decoding

Reduce LLM inference latency using draft models, Medusa heads, and Lookahead Decoding.

long-context

Extend transformer context windows to 32k–128k+ tokens using RoPE, YaRN, and ALiBi.

moe-training

Train Mixture-of-Experts models with DeepSpeed MoE and HuggingFace Transformers.

model-pruning

Prune large language models using one-shot methods like Wanda and SparseGPT.

model-merging

Merge fine-tuned models using linear, SLERP, task_arithmetic, TIES, and DARE methods.

knowledge-distillation

Compress large language models by transferring teacher knowledge to smaller student models.

pinecone

Manage a serverless vector database with hybrid search and metadata filtering.

chroma

Store embeddings and metadata for semantic search and RAG pipelines.

faiss

Perform billion-scale k-NN similarity search with FAISS index types and GPU acceleration.

Community

Intermediate

Frequently Asked Questions

FAQPage Schema

How to install AI-research-SKILLs?▼

Run `npx skills add zechenzhangAGI/AI-research-SKILLs --all -g -y` in your terminal to install all 98 research skills globally for your AI agent.

Can an AI agent run research autonomously with this library?▼

Yes. The autoresearch skill orchestrates the full lifecycle with a two-loop experiment-and-synthesis architecture, routing to domain skills and running continuously via Claude Code /loop or OpenClaw heartbeat.

What ML frameworks do these skills cover?▼

Skills cover vLLM, TensorRT-LLM, SGLang, Axolotl, TRL, DeepSpeed, Megatron-Core, Ray, NeMo, LangChain, DSPy, Chroma, Qdrant, Stable Diffusion, Whisper, OpenVLA, and dozens more across 23 categories.

Does it help write conference papers?▼

Yes. Dedicated skills provide LaTeX templates for NeurIPS, ICML, ICLR, OSDI, and SOSP, plus programmatic citation verification to prevent hallucinated references and academic figure generation.

Do I need research experience to use these skills?▼

No. Each skill includes step-by-step checklists, copy-paste code examples, and troubleshooting guides, so your AI agent handles the technical execution from plain-English goals.