# vllm-omni-skills

A collection of AI assistant skills that teach a coding agent how to install, configure, serve and troubleshoot vLLM-Omni, a framework for omni-modality model inference across text, image, video and audio.

The package provides one skill per operational area of vLLM-Omni: installation and environment setup, hardware backend configuration for CUDA, ROCm, NPU or XPU, launching and scaling API servers, adding image, video or audio generation models such as FLUX, Wan2.2 or Qwen3-TTS, distributed and disaggregated inference, performance tuning and quantization, and contributing new models, CI/CD pipelines and release notes. Skills activate automatically based on what a developer asks the agent to do, rather than needing to be invoked by name.

It installs into Cursor and Codex by copying or symlinking the skills directory into the tool's own skills folder, or into Claude Code by adding the repository as a plugin marketplace and installing individual skill plugins from it, such as the PR-review skill on its own.

It is written in Python and licensed under Apache-2.0.

Website: https://github.com/hsliuustc0106/vllm-omni-skills