ultralytics/skills

ultralytics/skills has 38 GitHub stars and 0 forks. GitHot measured 0 stars in 24 hours. Last checked 2026-11-12 22:24 UTC. Last push 2026-11-08.

Comment ces chiffres sont mesurés

Agent skills for Ultralytics YOLO, covering models, datasets, training, tuning, inference, tracking, and export.

38 étoiles 0 forks

Progression mesurée des étoiles

in 24h: 0
in 7 days: Historique en cours
in 15 days: Historique en cours
in 1 month: Historique en cours

Variations nettes sur des périodes qui se chevauchent, pas des totaux quotidiens. Les mesures absentes restent vides. Les étoiles indiquent l’intérêt, pas la qualité.

README.md

Original

English | 简体中文

🧠 Ultralytics Agent Skills

Agent skills for Ultralytics Platform, the ultralytics Python package, and the yolo CLI. They teach AI coding agents (Claude Code, Codex, Cursor, and any agent that reads the Agent Skills format) the full computer-vision lifecycle: data/annotation → training → tuning → inference/tracking → export/deployment.

🧩 Skills

Skill	Use it for
`yolo`	Platform and local lifecycle overview, core CLI/Python grammar, routing to the other skills
`yolo-models`	Choosing family/size/task variant: YOLO26/11/v8, YOLO-World, YOLOE, SAM, RT-DETR; exact weight names
`yolo-datasets`	Platform upload/annotation, data.yaml, formats, converters, splitting, validation
`yolo-training`	Platform cloud/remote training, local train/val, arguments, results, troubleshooting
`yolo-tuning`	Platform experiment comparison, improvement playbook, `model.tune()`, Ray Tune
`yolo-inference`	Platform Predict/endpoints, local predict, Results API, tracking, Solutions
`yolo-export`	Platform/local ONNX/TensorRT/CoreML/Core AI/OpenVINO/LiteRT/NPU export, quantization, benchmarking
`platform-cli`	Scripting Platform with `ul cloud`: resources, dataset inspection, cloud training, model analysis, exports, deployments, billing

Each skill is a SKILL.md of procedures, decision tables, and gotchas, with Codex and ChatGPT presentation metadata in agents/openai.yaml. Version-volatile catalogs (weight names, argument tables, export formats) live in companion files beside it. Package facts are grounded against ultralytics v8.4.208; Platform flows are grounded against the current Platform documentation.

📦 Install

Install the ultralytics package, including all requirements, in a Python>=3.8 environment with PyTorch>=1.8.

pip install ultralytics

On Python 3.11+ this also installs the ul CLI used by the platform-cli skill; it ships in the ultralytics-platform dependency, which can also be installed on its own.

For alternative installation methods, including Conda, Docker, and building from source via Git, please consult the Quickstart Guide.

Claude Code

claude plugin marketplace add ultralytics/skills
claude plugin install yolo@ultralytics

Develop from a local clone

git clone https://github.com/ultralytics/skills
claude plugin marketplace add ./skills
claude plugin install yolo@ultralytics

Try it without installing anything

Launch a session with the plugin loaded from the repo directory, leaving your installed plugins untouched:

git clone https://github.com/ultralytics/skills && cd skills
claude --plugin-dir .

Codex

codex plugin marketplace add ultralytics/skills
codex plugin add yolo@ultralytics

Restart Codex to use the installed skills.

Develop from a local clone

git clone https://github.com/ultralytics/skills && cd skills
codex plugin marketplace add .
codex plugin add yolo@ultralytics

Restart Codex after edits. To pull new versions of the GitHub source, run codex plugin marketplace upgrade ultralytics, then codex plugin add yolo@ultralytics.

Other agents (Cursor, Gemini CLI, ...)

npx skills add ultralytics/skills

Single skills, global install, manual copy

npx skills add ultralytics/skills --skill yolo-training # one skill only
npx skills add ultralytics/skills -g                    # global (~/.claude/skills/) instead of ./.claude/skills/

Or simply copy (or symlink) the folders under skills/ into your agent's skills directory.

🛠️ Design notes

Skills are divided by lifecycle stage and user intent, not by model family.
Frontmatter is portable: name + description only.
Each lifecycle skill covers both the Platform UI and the local Python/CLI path where applicable.
Every skill defers to the installed version at runtime: yolo checks (version), yolo cfg (valid arguments), and error messages beat any table in these files. The Platform CLI skill defers to ul version, ul cloud <resource> <operation> --help, and the live /openapi.json.

🔒 Data and network access

The plugin contains only instructions; nothing runs when it loads. The commands it suggests can download packages, weights, datasets, and the public Platform API schema, and ul cloud commands upload the files you pass them to Ultralytics Platform with your API key. The ultralytics package also sends anonymous usage analytics; turn them off with yolo settings sync=False.

💡 Contribute

Ultralytics thrives on community collaboration, and we deeply value your contributions! Please see our Contributing Guide for details on how to get involved. We also invite you to share your feedback through our Survey. A huge 🙏 thank you to all our contributors!

📄 License

Ultralytics offers two licensing options to accommodate diverse needs:

AGPL-3.0 License: Ideal for students, researchers, and enthusiasts passionate about open collaboration and knowledge sharing. This OSI-approved open-source license promotes transparency and community involvement. See the LICENSE file for details.
Enterprise License: Designed for commercial applications, this license permits the seamless integration of Ultralytics software and AI models into commercial products and services, bypassing the copyleft requirements of AGPL-3.0. For commercial use cases, please inquire about an Ultralytics Enterprise License.

📮 Contact

For bug reports or feature suggestions related to these skills, please use GitHub Issues. For general questions, discussions, and community support, join our Discord server!