Skills.

Give your AI a better way to work.

A skill is a set of written instructions that teaches an AI how to do one job the way it should be done: review a pull request, plan a migration, write the release notes.

Install one here and it travels with your account into Claude, Claude Code, Cursor and every other client you sign in with.

Filter by category

55,543 results · page 101 of 1,852

  • evaluating-cosmos-policySkillWeb & browsing

    Lets your agent run NVIDIA Cosmos Policy robot manipulation evaluations in LIBERO and RoboCasa GPU simulations.

    Ready to connect★ 13k

    github.com/orchestra-research/ai-research-skills13k stars

    View details
  • evaluating-llms-harnessSkillAI & models

    Lets your agent test and compare language models on 60+ standard academic benchmarks like MMLU and HumanEval.

    Ready to connect★ 13k

    github.com/orchestra-research/ai-research-skills13k stars

    View details
  • evolving-ai-agentsSkillAI & models

    Gives your agent guidance on evolving and optimizing AI agents using automated, benchmark-driven improvement loops.

    Ready to connect★ 13k

    github.com/orchestra-research/ai-research-skills13k stars

    View details
  • experiment-tracking-swanlabSkillMonitoring & ops

    Lets your agent log and track machine learning experiments with SwanLab runs and dashboards.

    Ready to connect★ 13k

    github.com/orchestra-research/ai-research-skills13k stars

    View details
  • faissSkillSearch

    Lets your agent run fast similarity search over large collections of vector embeddings.

    Ready to connect★ 13k

    github.com/orchestra-research/ai-research-skills13k stars

    View details
  • fine-tuning-openvla-oftSkillDev tools

    Lets your agent fine-tune and evaluate OpenVLA-OFT robot policies for action generation on LIBERO and ALOHA setups.

    Ready to connect★ 13k

    github.com/orchestra-research/ai-research-skills13k stars

    View details
  • fine-tuning-serving-openpiSkillDatabases & data

    Lets your agent fine-tune and run OpenPI robot models like pi0 on custom datasets using JAX or PyTorch.

    Ready to connect★ 13k

    github.com/orchestra-research/ai-research-skills13k stars

    View details
  • fine-tuning-with-trlSkillAI & models

    Lets your agent fine-tune language models using reinforcement learning methods like SFT, DPO, and PPO.

    Ready to connect★ 13k

    github.com/orchestra-research/ai-research-skills13k stars

    View details
  • gguf-quantizationSkillCloud & infra

    Teaches your agent how to compress AI models into GGUF format so they run efficiently on consumer CPUs and Macs.

    Ready to connect★ 13k

    github.com/orchestra-research/ai-research-skills13k stars

    View details
  • gptqSkillDocs & knowledge

    Lets your agent shrink large language models to 4-bit so they fit and run faster on consumer GPUs.

    Ready to connect★ 13k

    github.com/orchestra-research/ai-research-skills13k stars

    View details
  • grpo-rl-trainingSkillProductivity

    Gives your agent expert guidance on fine-tuning language models with GRPO reinforcement learning and TRL.

    Ready to connect★ 13k

    github.com/orchestra-research/ai-research-skills13k stars

    View details
  • guidanceSkillAI & models

    Lets your agent force an AI model to output exact formats like JSON or code using grammar rules.

    Ready to connect★ 13k

    github.com/orchestra-research/ai-research-skills13k stars

    View details
  • hqq-quantizationSkillDatabases & data

    Lets your agent shrink large language models to 4/3/2-bit precision without needing calibration data.

    Ready to connect★ 13k

    github.com/orchestra-research/ai-research-skills13k stars

    View details
  • huggingface-accelerateSkillDev tools

    Lets your agent run PyTorch training across multiple GPUs with a few lines of code.

    Ready to connect★ 13k

    github.com/orchestra-research/ai-research-skills13k stars

    View details
  • huggingface-tokenizersSkillDev tools

    Lets your agent split text into tokens using fast tokenizers and train custom vocabularies.

    Ready to connect★ 13k

    github.com/orchestra-research/ai-research-skills13k stars

    View details
  • implementing-llms-litgptSkillFiles & storage

    Lets your agent write and train large language model code using LitGPT, with pretrained models like Llama and Mistral.

    Ready to connect★ 13k

    github.com/orchestra-research/ai-research-skills13k stars

    View details
  • instructorSkillAI & models

    Lets your agent turn LLM text responses into validated structured data like JSON with typed fields.

    Ready to connect★ 13k

    github.com/orchestra-research/ai-research-skills13k stars

    View details
  • knowledge-distillationSkillCloud & infra

    Teaches your agent how to compress large language models into smaller ones that keep most performance.

    Ready to connect★ 13k

    github.com/orchestra-research/ai-research-skills13k stars

    View details
  • langchainSkillDocs & knowledge

    Lets your agent build LLM-powered apps with agents, chains, memory, and document retrieval across many AI providers.

    Ready to connect★ 13k

    github.com/orchestra-research/ai-research-skills13k stars

    View details
  • langsmith-observabilitySkillDatabases & data

    Lets your agent trace, evaluate, and monitor LLM applications to debug issues and test model outputs.

    Ready to connect★ 13k

    github.com/orchestra-research/ai-research-skills13k stars

    View details
  • llama-cppSkillDocs & knowledge

    Lets your agent run language models locally on Macs, CPUs, or non-NVIDIA GPUs without cloud services.

    Ready to connect★ 13k

    github.com/orchestra-research/ai-research-skills13k stars

    View details
  • llama-factorySkillAI & models

    Gives your agent expert guidance for fine-tuning large language models with LLaMA-Factory, including QLoRA and multimodal setups.

    Ready to connect★ 13k

    github.com/orchestra-research/ai-research-skills13k stars

    View details
  • llamaguardSkillCloud & infra

    Lets your agent check text against safety categories like violence, hate, and self-harm before sending or using it.

    Ready to connect★ 13k

    github.com/orchestra-research/ai-research-skills13k stars

    View details
  • llamaindexSkillDocs & knowledge

    Lets your agent index documents and answer questions over them using retrieval-augmented search.

    Ready to connect★ 13k

    github.com/orchestra-research/ai-research-skills13k stars

    View details
  • llavaSkillMedia

    Lets your agent answer questions about images and hold multi-turn conversations that combine pictures with text.

    Ready to connect★ 13k

    github.com/orchestra-research/ai-research-skills13k stars

    View details
  • long-contextSkillDocs & knowledge

    Teaches your agent how to extend a model's context window to handle long documents up to 128k+ tokens.

    Ready to connect★ 13k

    github.com/orchestra-research/ai-research-skills13k stars

    View details
  • mamba-architectureSkillMedia

    Teaches your agent how Mamba state-space language models work, including their speed and long-sequence trade-offs.

    Ready to connect★ 13k

    github.com/orchestra-research/ai-research-skills13k stars

    View details
  • miles-rl-trainingSkillAI & models

    Guides your agent through setting up large-scale reinforcement learning training for big language models.

    Ready to connect★ 13k

    github.com/orchestra-research/ai-research-skills13k stars

    View details
  • ml-paper-writingSkillDev tools

    Helps your agent draft and polish machine learning research papers for top AI conferences.

    Ready to connect★ 13k

    github.com/orchestra-research/ai-research-skills13k stars

    View details
  • mlflowSkillCloud & infra

    Lets your agent track machine learning experiments, version models, and deploy them to production.

    Ready to connect★ 13k

    github.com/orchestra-research/ai-research-skills13k stars

    View details

What is a skill?

A skill is plain text, usually a SKILL.md file and the scripts it refers to, written for an AI rather than for a person. It carries the steps, the house rules and the examples a good answer needs, so you stop pasting the same briefing into every new chat.

55,111 of the 55,543 skills listed here can be served through ahel today, and they come from public repositories. Each one has its own page with the instructions themselves on it, so you can read what a skill will tell your AI to do before you install it.

Install one and every AI you use gets it

Installing a skill adds it to your gateway and turns it on in the same step. Claude Code surfaces it as a slash command; any client can read the full instructions with the skill_read tool.

Nothing is copied into a project folder. The instructions are served from your account, so the same skill is there in every AI you connect, and turning it off removes it from all of them at once.

See how to connect your AI