AI-Research-SKILLs

A library of 98 Agent Skills covering AI research tasks end to end.

Free Development & Engineering ★ 13701

View on GitHub →

#48 of 58 skill repos

13.7k GitHub stars

+218 stars this week

Orchestra-Research author

// about

AI Research SKILLs lets an AI agent carry out AI research work such as model architecture selection, fine-tuning, post-training, and inference. Each skill follows the SKILL.md format and includes quick-reference docs, deeper reference material, helper scripts, and examples drawn from real GitHub issues.

// how to install

Run npx @orchestra-research/ai-research-skills, which auto-detects supported coding agents and installs selected skills.

// works with Claude CodeOpenCodeCursorCodexGemini CLI

// what’s inside

This repository bundles the following skills:

LitGPT: Skill for training and fine-tuning models with the LitGPT framework.
Mamba: Skill for working with the Mamba state-space model architecture.
Axolotl: Skill for fine-tuning language models using Axolotl.
PEFT: Skill for parameter-efficient fine-tuning methods.
Unsloth: Skill for fast, memory-efficient fine-tuning.
TRL: Skill for reinforcement-learning-based post-training.
vLLM: Skill for serving models efficiently.
llama.cpp: Skill for running inference with llama.cpp.
LangChain: Skill for building agent workflows with LangChain.

AI-Research-SKILLs

// about

// what’s inside

// similar skills in Development & Engineering