openvinotoolkit

Review OpenVINO changes for correctness, compatibility, performance, security, testing, and maintainability. Use when reviewing a pull request or proposed diff.

Develop and troubleshoot the OpenVINO GGUF frontend (src/frontends/gguf) and the llama.cpp ggml-openvino backend that uses it. Use when checking whether a .gguf or mmproj model is supported; when a .gguf file is rejected as an unsupported architecture; when…

Author, edit, compile, validate, and debug any OpenVINO gh-aw agentic workflow (e.g. ci-doctor, ci-doctor-mq, and any future ones) and their shared safe-output jobs under .github/workflows/shared/agentic-workflows. Use when a user wants to create a new…

Author, edit, debug, and review the OpenVINO GitHub Actions CI infrastructure — regular (non-agentic) workflows under .github/workflows, reusable job_*.yml workflows, custom composite actions under .github/actions, CI helper scripts under .github/scripts,…

Write or review OpenVINO ov::Model graph transformations (passes: MatcherPass, ModelPass, GraphRewrite, BackwardGraphRewrite). Use when the user asks to write, add, or refactor a transformation/fusion/decomposition pass, choose between MatcherPass and…

Debug OpenVINO CPU/GPU plugin inference, accuracy, performance, memory, and device compilation issues, or inspect graph transformations, using debug capabilities such as tensor dumps, execution graphs, profiling, and transformation tracing. Do NOT use for…

Upgrade the PyTorch version used by OpenVINO tests (torch / torchvision / torchaudio) and resolve fallout — missing operator translators, new functionalized `*_copy` aten ops, decomposition changes, FX-only tests failing in TorchScript mode, and accuracy…

Investigate and fix model conversion issues in OpenVINO Frontends (ONNX, PyTorch) — triage, debugging, accuracy comparison, and pre-submission verification.

Adds or modifies a first-party robot integration in the Physical AI Runtime source tree. Use when implementing the Robot protocol under src/physicalai/robot, registering built-in drivers or optional extras, or changing verify.py and robot tests. DO NOT USE…

Runs an exported Physical AI policy on a robot with RobotRuntime and physicalai run. Use when wiring a robot, cameras, PolicySource or TeleopSource, choosing SyncExecution/AsyncExecution/RTCExecution, configuring a runtime YAML recipe, or diagnosing action…

Adds or modifies a camera backend in the Physical AI Runtime source tree. Use when implementing Camera, extending create_camera or discovery under src/physicalai/capture, adding optional vendor SDK extras, or writing capture tests with fake devices. DO NOT…

Adds a built-in preprocessor, postprocessor, or runner to Physical AI Runtime. Use when editing src/physicalai/inference component classes, ComponentRegistry, or their unit tests so a manifest can instantiate a new type. DO NOT USE FOR configuring an existing…

Maintains the Physical AI Runtime configuration package shared with Studio. Use when changing src/physicalai/config, export_config, Config serialization and instantiation, jsonargparse integration, or tests/unit/config. DO NOT USE FOR writing a trusted local…

Configures an exported policy's inference pipeline with Runtime manifest preprocessors, runners, and postprocessors. Use when writing preprocessor/postprocessor lists, choosing a registered type or class_path/init_args, resolving export artifacts, or fixing…

Configures Physical AI Runtime objects with physicalai.config and YAML recipes. Use when writing class_path/init_args for robots, cameras, runtime or inference components, exporting a local object with @export_config, or choosing jsonargparse for typed…

Creates or modifies an external Physical AI robot plugin for Studio. Use when implementing a Runtime Robot driver in your own package, defining studio_catalog.py with RobotCatalogDefinition, RobotProbe or RobotAsset, and registering the…

Enables VLM models for GenAI VLM pipelines.

Validate a newly supported optimum-intel model with OpenVINO GenAI. Use when: checking new model support, verifying model export to OpenVINO IR, running GenAI inference test with llm_bench, benchmarking model accuracy with who-what-benchmark.

Run, write, and fix OpenVINO GenAI Python and WWB tests, and set up their Hugging Face / OpenVINO model and dataset caches. Use when: running tests/python_tests or tools/who_what_benchmark/tests locally, adding a test that downloads a model/dataset/adapter…

Generate OpenVINO GenAI release notes from commits since the latest release. Use when: preparing release notes, summarizing a release branch, or comparing the latest release tag with the current release branch.

Update OpenVINO GenAI site documentation for API or feature changes. Use when: new pipelines, models, or use-cases are introduced; site docs need to reflect new capabilities.

Analyze results of WWB execution with a specific model with all possible backends. Use when: checking wwb logs, troubleshooting wwb fails.

Analyze failed llm_bench execution results for a model. Use when: checking llm_bench log, troubleshooting llm_bench fails.

Open a pull request to openvinotoolkit/openvino.genai. Use when: submitting changes, creating PR, opening pull request, pushing work for review, finalizing a task with a PR.

Validate a HuggingFace tokenizer with OpenVINO Tokenizers and OpenVINO GenAI. Use when: checking if a tokenizer converts and works correctly, verifying tokenizer/detokenizer accuracy, testing normalization steps, checking GenAI Tokenizer compatibility.

Diagnose tokenizer conversion issues for OpenVINO Tokenizers. Use when: tokenizer-checker reports failures, need to pinpoint root cause location (Python conversion vs C++ operation), identify which pipeline stage diverges, determine whether to use…

Fix Python conversion issues in OpenVINO Tokenizers. Use when: tokenizer-diagnostics reports root_cause_location=python, unsupported types need new handlers in hf_parser.py, pipeline step mapping or merging is incorrect, tokenizer_pipeline.py step classes…

Standardize GitHub Actions workflows with minimal functional changes.

Generate release notes for the new NNCF release.

Guidance for implementing Python MCP SDK stdio servers with FastMCP, robust validation, and deterministic tool output.