onnxruntime

Audit and fix out-of-range output writes in ONNX Runtime operator shape-inference functions. Use when reviewing or fixing a contrib (or standard) op TypeAndShapeInference where a getNumOutputs() guard precedes a write to a higher output index - optional trailing outputs make a smaller output count schema-valid, so getOutputType(index) can run one past the declared outputs at Graph::Resolve.

Build and run ONNX Runtime WebGPU provider tests on Linux WITHOUT a real GPU, using a software Vulkan adapter (Mesa lavapipe). Use when you need to exercise WebGPU EP kernels off-Mac — the Linux webgpu CI leg is build-only, so software Vulkan is how you actually run WebGPU correctness tests locally. SCOPE - lavapipe only validates host-side enforce/shape bugs and MatMul-free kernels; any graph containing MatMul (including the expanded-Attention node tests) crashes lavapipe and runs ONLY on macOS-arm64 Metal, which is the source of truth for those. Covers install (dnf on Azure Linux), the --use_webgpu build flag, the onnxruntime_provider_test target, VK_ICD_FILENAMES, and the lavapipe MatMul crash gotcha.

Patterns and pitfalls for the ONNX domain Attention operator (opset 23/24) CUDA implementation. Use when modifying the dispatch cascade in core/providers/cuda/llm/attention.cc, writing mask/bias CUDA kernels, debugging attention test routing, or adding features to the ONNX Attention op. NOT for contrib domain MultiHeadAttention/GroupQueryAttention.

Use when rebuilding ONNX Runtime CUDA after editing CUTLASS fused-MHA headers (onnxruntime/contrib_ops/cuda/bert/cutlass_fmha/*.h such as kernel_forward.h or fmha_launch_template.h), or when a header edit "passed" an incremental build but test behavior did not change. Explains the nvcc depfile gotcha that produces stale Memory-Efficient-Attention (MEA) kernels and binaries, and how to force a correct recompile. Also covers disk-space frugality on shared GPU dev boxes.

Build ONNX Runtime from source. Use this skill when asked to build, compile, or generate CMake files for ONNX Runtime.

Run ONNX Runtime tests. Use this skill when asked to run tests, debug test failures, or find and execute specific test cases in ONNX Runtime.

Step-by-step checklist for bumping the pinned ONNX dependency / opset in ONNX Runtime (e.g. ONNX 1.21 / opset 26 → 1.22 / opset 27). Use when integrating a new ONNX release or release-candidate, updating the cmake/deps.txt onnx pin or the cmake/external/onnx submodule, regenerating cmake/patches/onnx/onnx.patch, raising kMaxSupportedOpset, or adding a new opset's CPU kernels. Covers the file taxonomy, archive-hash procedures, patch rebase/mirror rules, the RC→formal strategy, and the optimizer/EP gotchas that the automated audit script misses.

When reviewing or fixing Python code that uses setattr() with user-controlled kwargs to configure C++ extension objects (SessionOptions, RunOptions, etc.) in ONNX Runtime. Use this to apply the allowlist pattern that prevents arbitrary file writes and other attacks via reflected property access.

Lint and format ONNX Runtime code. Use this skill when asked to lint, format, or check code style for C++ or Python files in ONNX Runtime.