Repository
Orchestra-Research/AI-Research-SKILLs agent skills
- skills
- 96
- GitHub stars
- 13k
GitHub description: “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.”
- Stars
- 13,405 (947 forks)
- Licence
- MIT
- Last push
- Jun 2026
- Created
- Nov 2025
- Homepage
- orchestra-research.com
- ai
- ai-research
- claude
- claude-code
- claude-skills
- codex
- gemini
- gpt-5
- grpo
- huggingface
- machine-leanring
- megatron
- skills
- vllm
Install all skills
npx skills add Orchestra-Research/AI-Research-SKILLsAdd --skill <name> for a single skill and -a <agent> to choose the agent (see the agent guides).
Skills in Orchestra-Research/AI-Research-SKILLs, ranked
Ranked by score. Sort bymost stars,trending,newest,recently updated
| # | Skill | Repository | Stars | Used in | Tokens | Auto-check | Licence | Updated |
|---|---|---|---|---|---|---|---|---|
| 1 | Generates music from text descriptions with MusicGen and sound effects with AudioGen, using Meta's AudioCraft PyTorch library with melody and style conditioning. | Orchestra-Research/ | 13k | 8 repos | ~3.9k | Automated safety check: Pass | MIT | 3 mo ago |
| 2 | Runs lm-evaluation-harness to benchmark language models on academic suites such as MMLU, GSM8K and HumanEval, compare models and track training checkpoints. | Orchestra-Research/ | 13k | 8 repos | ~3k | Automated safety check: Pass | MIT | 3 mo ago |
| 3 | Guide to using Meta's Segment Anything Model for zero-shot image segmentation with point, box or mask prompts, or automatic mask generation. | Orchestra-Research/ | 13k | 8 repos | ~3.3k | Automated safety check: Pass | MIT | 3 mo ago |
| 4 | Shows how to store documents and embeddings in Chroma, query them by similarity with metadata filters, and persist them to disk for RAG and semantic search projects. | Orchestra-Research/ | 13k | 7 repos | ~2.3k | Automated safety check: Pass | MIT | 3 mo ago |
| 5 | Explains OpenAI's CLIP model for zero-shot image classification, image-text similarity, semantic image search and content moderation, with install steps and code patterns. | Orchestra-Research/ | 13k | 7 repos | ~1.7k | Automated safety check: Pass | MIT | 3 mo ago |
| 6 | Transcribes audio with OpenAI's Whisper: 99 languages, translation to English, language detection, six model sizes and word-level timestamps, from Python or the CLI. | Orchestra-Research/ | 13k | 7 repos | ~1.9k | Automated safety check: Notes | MIT | 3 mo ago |
| 7 | Sets up FAISS for fast nearest-neighbor search over large collections of dense vectors, choosing between Flat, IVF, HNSW and product quantization indexes. | Orchestra-Research/ | 13k | 6 repos | ~1.3k | Automated safety check: Pass | MIT | 3 mo ago |
| 8 | Fine-tune LLMs using reinforcement learning with TRL - SFT for instruction tuning, DPO for preference alignment, PPO/GRPO for reward optimization, and reward model training. | Orchestra-Research/ | 13k | 6 repos | ~2.9k | Automated safety check: Pass | MIT | 3 mo ago |
| 9 | Shows how to load, train and use fast Hugging Face tokenizers, with BPE, WordPiece and Unigram models, padding, truncation and alignment tracking. | Orchestra-Research/ | 13k | 6 repos | ~3.4k | Automated safety check: Pass | MIT | 3 mo ago |
| 10 | Guide to LLaVA for image chat, visual question answering and captioning, with model sizes, CLI and Gradio usage and multi-turn conversation code. | Orchestra-Research/ | 13k | 6 repos | ~2k | Automated safety check: Pass | MIT | 3 mo ago |
| 11 | Parameter-efficient fine-tuning for LLMs using LoRA, QLoRA, and 25+ methods. | Orchestra-Research/ | 13k | 6 repos | ~3.1k | Automated safety check: Pass | MIT | 3 mo ago |
| 12 | Shows how to organize PyTorch training with Lightning's LightningModule and Trainer, covering validation, DDP, callbacks and learning-rate scheduling. | Orchestra-Research/ | 13k | 6 repos | ~2.3k | Automated safety check: Pass | MIT | 3 mo ago |
| 13 | Expert guidance for fast fine-tuning with Unsloth - 2-5x faster training, 50-80% less memory, LoRA/QLoRA optimization | Orchestra-Research/ | 13k | 10 repos | ~577 | Automated safety check: Pass | MIT | 3 mo ago |
| 14 | Teaches an agent to build LM pipelines, RAG systems and agents in DSPy using signatures, modules and optimizers instead of hand-tuned prompts. | Orchestra-Research/ | 13k | 9 repos | ~3.8k | Automated safety check: Pass | MIT | 3 mo ago |
| 15 | Guides an agent through tracking ML experiments with W&B: run logging, config capture, hyperparameter sweeps, artifacts and a model registry. | Orchestra-Research/ | 13k | 9 repos | ~3.1k | Automated safety check: Pass | MIT | 3 mo ago |
| 16 | Uses the Outlines library to constrain model output to a JSON schema, Pydantic model, regex or fixed set of choices when running local models. | Orchestra-Research/ | 13k | 9 repos | ~4k | Automated safety check: Pass | MIT | 3 mo ago |
| 17 | Quick-reference help for fine-tuning language models with Axolotl, covering YAML configs, FSDP, context parallelism, compressed saves and dataset formats. | Orchestra-Research/ | 13k | 8 repos | ~1.2k | Automated safety check: Pass | MIT | 3 mo ago |
| 18 | Guides training and analyzing sparse autoencoders with SAELens to break neural network activations into interpretable features, including superposition and monosemanticity studies. | Orchestra-Research/ | 13k | 5 repos | ~3.2k | Automated safety check: Pass | MIT | 3 mo ago |
| 19 | Serverless GPU cloud platform for running ML workloads. An agent skill from Orchestra-Research/AI-Research-SKILLs. | Orchestra-Research/ | 13k | 5 repos | ~2.1k | Automated safety check: Pass | MIT | 3 mo ago |
| 20 | Shows how to use Pinecone, a managed vector database, for production RAG, semantic search and recommendations: indexes, upserts, queries, filters and namespaces. | Orchestra-Research/ | 13k | 5 repos | ~2k | Automated safety check: Pass | MIT | 3 mo ago |
| 21 | Deploys LLMs with vLLM for high-throughput serving, covering the OpenAI-compatible server, offline batch inference, monitoring and a Docker rollout. | Orchestra-Research/ | 13k | 5 repos | ~2.3k | Automated safety check: Pass | MIT | 3 mo ago |
| 22 | Generates and edits images with Stable Diffusion through Hugging Face Diffusers, covering text-to-image, image-to-image, inpainting, SDXL and custom pipelines. | Orchestra-Research/ | 13k | 5 repos | ~3.2k | Automated safety check: Pass | MIT | 3 mo ago |
| 23 | Benchmarks code generation models with the BigCode Evaluation Harness across HumanEval, MBPP, MultiPL-E and other suites using pass@k metrics. | Orchestra-Research/ | 13k | 4 repos | ~2.9k | Automated safety check: Pass | MIT | 3 mo ago |
| 24 | Runs ML training and batch jobs across clouds with SkyPilot, using spot instances, automatic region selection and managed recovery to cut GPU cost. | Orchestra-Research/ | 13k | 4 repos | ~2.4k | Automated safety check: Pass | MIT | 3 mo ago |
| 25 | Provides PyTorch-native distributed LLM pretraining using torchtitan with 4D parallelism (FSDP2, TP, PP, CP). | Orchestra-Research/ | 13k | 4 repos | ~2.2k | Automated safety check: Pass | MIT | 3 mo ago |
| 26 | 26.Nemo Curator GPU-accelerated data curation for LLM training. An agent skill from Orchestra-Research/AI-Research-SKILLs. | Orchestra-Research/ | 13k | 4 repos | ~2.3k | Automated safety check: Pass | MIT | 3 mo ago |
| 27 | Optimizes transformer attention with Flash Attention for 2-4x speedup and 10-20x memory reduction. | Orchestra-Research/ | 13k | 4 repos | ~2.5k | Automated safety check: Pass | MIT | 3 mo ago |
| 28 | Explains how to run Qdrant, a Rust vector database, for RAG and semantic search, covering collections, points, distance metrics and filtered or batched queries. | Orchestra-Research/ | 13k | 4 repos | ~3.4k | Automated safety check: Pass | MIT | 3 mo ago |
| 29 | Walks through aligning language models with SimPO, a reference-free preference optimization method, using accelerate configs for Mistral 7B, Llama 3 8B and math-focused models. | Orchestra-Research/ | 13k | 4 repos | ~1.5k | Automated safety check: Pass | MIT | 3 mo ago |
| 30 | Guides reinforcement-learning post-training of LLMs with slime, which pairs Megatron-LM training with SGLang rollouts, including GRPO runs on GLM, Qwen3 and Llama 3 models. | Orchestra-Research/ | 13k | 4 repos | ~2.8k | Automated safety check: Pass | MIT | 3 mo ago |
| 31 | Optimizes and serves LLMs on NVIDIA GPUs with TensorRT-LLM, covering quantization, in-flight batching, multi-GPU parallelism and the trtllm-serve command. | Orchestra-Research/ | 13k | 4 repos | ~1.3k | Automated safety check: Pass | MIT | 3 mo ago |
| 32 | Shows how to pull validated, typed data out of LLM responses with Instructor and Pydantic models, including retries on failure and partial streaming. | Orchestra-Research/ | 13k | 6 repos | ~4.2k | Automated safety check: Pass | MIT | 3 mo ago |
| 33 | Constrains language model output with regex, selections and grammars using the Guidance library, so JSON, XML, code or formatted fields come out valid. | Orchestra-Research/ | 13k | 5 repos | ~3.6k | Automated safety check: Pass | MIT | 3 mo ago |
| 34 | Adds distributed and mixed-precision training to a PyTorch script with a few Accelerate lines, then launches it on one GPU, many GPUs or DeepSpeed and FSDP setups. | Orchestra-Research/ | 13k | 5 repos | ~2.1k | Automated safety check: Pass | MIT | 3 mo ago |
| 35 | Guides mechanistic interpretability work with TransformerLens: loading models, caching activations, using HookPoints, activation patching and attention-pattern analysis. | Orchestra-Research/ | 13k | 3 repos | ~3k | Automated safety check: Pass | MIT | 3 mo ago |
| 36 | GGUF format and llama.cpp quantization for efficient CPU/GPU inference. | Orchestra-Research/ | 13k | 3 repos | ~2.6k | Automated safety check: Pass | MIT | 3 mo ago |
| 37 | 37.Llama Cpp Runs LLM inference on CPU, Apple Silicon, and consumer GPUs without NVIDIA hardware. | Orchestra-Research/ | 13k | 3 repos | ~1.5k | Automated safety check: Pass | MIT | 3 mo ago |
| 38 | Uses Ray Data to read, transform and write large datasets across a cluster for ML training and batch inference, with streaming execution and optional GPU steps. | Orchestra-Research/ | 13k | 3 repos | ~1.8k | Automated safety check: Pass | MIT | 3 mo ago |
| 39 | Autonomous AI agent platform for building and deploying continuous agents. | Orchestra-Research/ | 13k | 2 repos | ~2.3k | Automated safety check: Notes | MIT | 3 mo ago |
| 40 | Activation-aware weight quantization for 4-bit LLM compression with 3x speedup and minimal accuracy loss. | Orchestra-Research/ | 13k | 2 repos | ~2.1k | Automated safety check: Pass | MIT | 3 mo ago |
| 41 | Vision-language pre-training framework bridging frozen image encoders and LLMs. | Orchestra-Research/ | 13k | 2 repos | ~4.3k | Automated safety check: Pass | MIT | 3 mo ago |
| 42 | Multi-agent orchestration framework for autonomous AI collaboration. | Orchestra-Research/ | 13k | 2 repos | ~3.4k | Automated safety check: Pass | MIT | 3 mo ago |
| 43 | 43.Gptq Post-training 4-bit quantization for LLMs with minimal accuracy loss. | Orchestra-Research/ | 13k | 2 repos | ~2.9k | Automated safety check: Pass | MIT | 3 mo ago |
| 44 | Half-Quadratic Quantization for LLMs without calibration data. | Orchestra-Research/ | 13k | 2 repos | ~2.9k | Automated safety check: Pass | MIT | 3 mo ago |
| 45 | Implements and trains LLMs using Lightning AI's LitGPT with 20+ pretrained architectures (Llama, Gemma, Phi, Qwen, Mistral). | Orchestra-Research/ | 13k | 2 repos | ~2.8k | Automated safety check: Pass | MIT | 3 mo ago |
| 46 | Compress large language models using knowledge distillation from teacher to student models. | Orchestra-Research/ | 13k | 2 repos | ~3.4k | Automated safety check: Pass | MIT | 3 mo ago |
| 47 | 47.Langchain Framework for building LLM-powered applications with agents, chains, and RAG. | Orchestra-Research/ | 13k | 2 repos | ~3.2k | Automated safety check: Pass | MIT | 3 mo ago |
| 48 | LLM observability platform for tracing, evaluation, and monitoring. | Orchestra-Research/ | 13k | 2 repos | ~2.4k | Automated safety check: Pass | MIT | 3 mo ago |
Questions, answered from the data.
What is the best skill in Orchestra-Research/AI-Research-SKILLs?
AudioCraft Audio Generation from Orchestra-Research/AI-Research-SKILLs ranks first of the 96 skills in Orchestra-Research/AI-Research-SKILLs listed here, with the highest score: its repository has 13k GitHub stars, 8 other GitHub owners carry a copy, its SKILL.md loads about 3.9k tokens and it passes the automated safety check with no findings. Next come LLM Benchmarking with lm-evaluation-harness and Segment Anything Model Guide.
Are the skills in Orchestra-Research/AI-Research-SKILLs official?
None yet. All 96 skills in Orchestra-Research/AI-Research-SKILLs listed here come from community repositories; a skill counts as official when the product's own GitHub organization publishes it.
How do I install all skills from Orchestra-Research/AI-Research-SKILLs?
Run npx skills add Orchestra-Research/AI-Research-SKILLs in your project: the open-source skills CLI installs the repository's skills into your coding agent's skills folder. To install a single skill, open its page here for the exact command.
How are these skills ranked?
By Skill Navigator score, which combines the GitHub stars of the skill's repository (shared across that repo's skills and discounted for large collections), how many other GitHub owners carry a copy of the skill, and automated SKILL.md quality checks, minus penalties for safety-check warnings and for each further skill from the same repository. Skills that fail the safety check are not listed.