Structured skill packs for AI agents (Claude Code, Codex, Cursor) to operate the Horizon AI toolchain — HBDK compilation, HMCT/Plugin quantization, UCP on-board inference, and LLM compression.
Skill Claude CodeCodex
A tool for comparing two PyTorch FX Graph files, which describe the operations in a machine-learning model, and linking differences back to model source code.
Skill Claude CodeCodex
A code-generation workflow for quantizing and compiling neural-network models for Horizon hardware. Quantization reduces numerical precision so models can be prepared for deployment.
Skill Claude CodeCodex
Um gerador de código que transforma um modelo já quantizado num ficheiro HBM compilado para implantação em hardware Horizon. HBM é o formato de modelo usado pelo processo de compilação para execução no dispositivo.
Skill Claude CodeCodex
Um gerador de código para preparar modelos de aprendizagem automática para quantização no hardware Horizon J6. Quantização reduz a precisão numérica do modelo para permitir a execução no dispositivo; QAT significa treino consciente da quantização.
Skill Claude CodeCodex
A guide for reading Horizon model-check results after preparing a model for quantization, a process that makes models use lower-precision numbers. It examines model structure and quantization settings using model_check_result.txt.
Skill Claude CodeCodex
A guide for fixing accuracy problems in PyTorch models on Horizon J6 hardware before deployment. It focuses on calibration and quantization-aware training, where model numbers are adapted for faster, lower-precision processing.
Skill Claude CodeCodex
A routing entry point for Horizon J6 development tasks. It sends requests to focused tools for model performance tests, HBM inference deployment, inference code generation, and Perfetto trace analysis.
Skill Claude CodeCodex
A tool for measuring a J6 development board’s processor load, DDR memory bandwidth, and memory use over SSH. It can collect these readings while computer-vision or language models run, or while the board is idle.
Skill Claude CodeCodex
Generate X86-side Python client code for hbminfer, the Python SDK that connects to BPU boards via gRPC to deploy and run HBM model inference. Use this skill when the user wants to write Python code using hbminfer client APIs — including HbmRpcSession (standard/flexible mode), HTensor for transmission optimization…
Skill Claude CodeCodex
A guide for generating C++ code for UCP-based neural-network inference, the process of running a trained model to produce predictions. It covers loading a .hbm model, preparing tensors, processing input, running inference, reading outputs, and releasing resources.
Skill Claude CodeCodex needs its repo
A repository instruction file for generating software assets such as logos, icons, favicons, illustrations, and hero images. It describes available generation routes and how to handle requests without an image API key.
Skill Claude CodeCodex
Use this skill whenever the user wants to quickly locate, diagnose, or analyze likely performance bottlenecks in a UCP model inference Perfetto trace. Trigger on requests about UCP inference traces, .pftrace files, inference latency, slow inference, pipeline stalls, CPU/BPU gaps, OpInfer delays, dispatch or response…
Skill Claude CodeCodex
Capture UCP (Horizon Robotics inference SDK) Perfetto traces from a J6 development board. Use this skill whenever the user wants to grab UCP traces, capture Perfetto traces on a Horizon dev board, debug UCP scheduling issues with trace data, or pull .pftrace files from a remote board. Also trigger when the user…
At most 3 mods per repository are shown here, and a mod shipped inside a plugin is left to that plugin's page — the rest are on their repository pages: