open-edge-platform
- 149 skills
- 0 followers
- 6.4k repo stars
- last month last updated
- ▌ Testing 2 · open-edge-platform bundleCreate Scenescape pytest cases (unit, functional, integration, UI, BAT) with SCENESCAPE_SPEC, Zephyr IDs, and positive/negative coverage. Use when adding or modifying tests under tests/, writing pytest, or choosing unit vs functional vs UI.
- ▌ Dlsps User 2 · open-edge-platform bundleDeploy and operate DL Streamer Pipeline Server — a microservice that wraps DL Streamer pipelines behind a REST API for containerized, no-code operation. Use this skill whenever a user wants to: deploy the pipeline server via Docker Compose or Helm; start, stop, or monitor pipeline instances through the REST API; configure pipeline definitions in config.json; publish inference metadata over MQTT, OPC UA, InfluxDB, S3, or ROS2; set up GPU/NPU device access for the container; troubleshoot service-level issues (container startup, REST errors, port conflicts). This skill is NOT for writing new DL Streamer applications or custom GStreamer code — use the dlstreamer-coding-agent skill for that. Trigger on phrases like "pipeline server", "DLSPS", "start pipeline via REST", "deploy video analytics microservice", "config.json pipeline definition".
- ▌ Vss Deploy 2 · open-edge-platform bundleDeploys and manages VSS through setup.sh and its Docker Compose overlays. Use this skill for local lifecycle tasks such as configuration, startup, mode changes, inspection, shutdown, data cleanup, and health checks. It supports summary, search, dual, and unified modes with GPU and vLLM variants.
- ▌ Vss Deploy Helm 2 · open-edge-platform bundleUse this skill whenever a developer needs to deploy VSS to Kubernetes, helm install VSS, configure values.yaml for VSS, or run VSS on k8s with GPU/vLLM for the video-search-and-summarization sample app. This skill is especially useful when translating Docker Compose/setup.sh modes (--summary, --search, --summary-and-search/--unified, dual UI, ENABLE_VLLM, OVMS GPU/NPU) into the actual Helm chart override files and values keys. Prefer this skill for VSS Helm install/upgrade/troubleshooting even if the user only says “put VSS on k8s” or “make values.yaml for VSS”.
- ▌ Scenescape Setup 2 · open-edge-platform bundleDeploy a working Intel® SceneScape installation from scratch (outside the repo). Gathers user-provided streams, camera IDs, scene name, and mapping choice, then runs bootstrap through tracking verification via scripts/deploy_scenescape.sh. Also handles re-running or resuming a single phase of an existing deployment on request (e.g. "recalibrate", "redo scene reconstruction", "resume bootstrap only") via the orchestrator's --phase flag.
- ▌ Vss Search Index 2 · open-edge-platform bundleSearch a video library with natural language via the VSS Pipeline Manager - upload a video (POST /videos), generate its embeddings (POST /videos/search-embeddings/{id}), then run a query (POST /search/query) with optional tag and time filters and read the ranked clip results. Use when the user says "search my videos", "find something in the videos", "when did X happen", or wants to ingest/index a video for search. Requires a search-capable deployment (--search, --dual, or --unified).
- ▌ Anomalib Training 2 · open-edge-platform bundleTrains an anomalib model on a dataset via the Python API or CLI, including training on a custom folder-structured dataset with the Folder datamodule. Use when writing or debugging an anomalib training script/command, choosing Engine/Trainer arguments, or wiring a model + datamodule together. Do not use for adding a new model or datamodule from scratch (see anomalib-adding-a-model / anomalib-adding-a-datamodule), or for the tiled-ensemble or benchmarking pipelines (see anomalib-tiled-ensemble / anomalib-benchmarking).
- ▌ Chatqna Helm Deploy 2 · open-edge-platform bundleDeploy Chat Question-and-Answer Core to Kubernetes using Helm (OpenVINO CPU, OpenVINO GPU, or Ollama), including values.yaml configuration, helm install/upgrade, deployment verification, uninstall, and translation from Docker Compose setup_env.sh variables into Helm override values. Use this skill when the user says "deploy chatqna core to kubernetes", "helm install chatqna-core", "configure values.yaml", "convert compose config to helm", or "translate setup_env.sh to chart values".
- ▌ Model Download User 2 · open-edge-platform bundleDownload and convert AI models using the Model Download microservice. Use this skill whenever a user wants to: download a model from HuggingFace, Ollama, Ultralytics, Geti, or Pipeline Zoo; convert a model to OpenVINO IR format for OVMS; download healthcare AI models (3D Pose, rPPG, AI-ECG) via the HLS plugin; set up the model download service; submit a download or conversion job via the REST API; or ask "how do I get model X working with OVMS?". Also trigger on phrases like "download model", "download weights", "convert to int4", "OVMS-ready model", "prepare model for inference".
- ▌ Vss Summarize Video 2 · open-edge-platform bundleSummarize a video through the VSS Pipeline Manager - start a summary pipeline with POST /summary (full required body), poll GET /summary/{stateId} until complete, then return the summary via GET /summary/{stateId}/raw. Use when the user says "summarize this video", "create a summary", "what happens in this video" (on an ingested video), or wants to run/inspect the summarization pipeline. Requires a summary-capable deployment (--summary, --dual, or --unified).
- ▌ Anomalib Benchmarking 2 · open-edge-platform bundleRuns the anomalib benchmarking pipeline to train/evaluate a grid of model + dataset (+ category) combinations and collect metrics into a results CSV. Use when comparing multiple models/datasets/categories in one sweep, or authoring/editing a benchmark config YAML. Do not use for training a single model (see anomalib-training) or the tiled-ensemble pipeline (see anomalib-tiled-ensemble). For turning measured results into README/docs benchmark tables, see the benchmark-and-docs-refresh skill.
- ▌ Chatqna Docker Deploy 2 · open-edge-platform bundleDeploy Chat Question-and-Answer Core with Docker Compose (OpenVINO CPU, OpenVINO GPU, or Ollama CPU), including env setup, profile selection, startup verification, health checks, and teardown. Use this skill when the user says "deploy chatqna core", "start chatqna container", "run compose", "openvino gpu deploy", or "ollama deploy". Canonical deploy sources are docker/compose.yaml (services and image names) and scripts/setup_env.sh (runtime profile export); Makefile is not the source of truth.
- ▌ Anomalib Adding A Model 2 · open-edge-platform bundleAdds a new anomaly-detection model to anomalib under src/anomalib/models/. Use when implementing a new model architecture (image or video), wiring it into the AnomalibModule base class, registering it so get_model() and the CLI/config (jsonargparse) can discover it, and adding matching tests. Do not use for changes to existing model internals only (no new model), or for datamodule/dataset work (see anomalib-adding-a-datamodule).
- ▌ Anomalib Tiled Ensemble 2 · open-edge-platform bundleRuns and configures the anomalib tiled-ensemble pipeline, which trains/evaluates one model per image tile and merges results (with optional seam smoothing) for high-resolution anomaly detection. Use when the user wants to train with image tiling, mentions "tiled ensemble", or needs to tune tiling/stride/seam-smoothing config. Do not use for regular single-model training (see anomalib-training) or the multi-model benchmarking pipeline (see anomalib-benchmarking).
- ▌ Dlstreamer Coding Agent 2 · open-edge-platform bundleBuild new DL Streamer video-analytics applications (Python, C, C++ or GStreamer command line). Use when: user describes a vision AI pipeline, wants to create a new sample app, combine elements from existing samples, add detection/classification/VLM/tracking/alerts/recording to a video pipeline, or create custom GStreamer elements in Python or C++. Translates natural-language pipeline descriptions into working DL Streamer code using established design patterns.
- ▌ Geti Using The Pipeline 2 · open-edge-platformUse the Geti application end to end through its REST API — the project → dataset → annotate → train → deploy pipeline served by the FastAPI backend in `application/backend/`. Use when a user (not a contributor) wants to create a project, upload media, add annotations, launch a training or quantization job, track job status, configure a source → model → sink inference pipeline, and enable live inference. Covers the `/api/...` endpoints and the async job model, not backend code changes.
- ▌ Multimodal Dataprep User 2 · open-edge-platform bundleDeploy and consume Intel Multimodal DataPrep from prebuilt images or a repository checkout. Use for configuring VDMS or Milvus vector storage, MinIO or local media storage, checking service dependencies, and ingesting, listing, streaming, or deleting videos and images; submitting batch jobs; adding text-summary embeddings; and inspecting telemetry. This service prepares retrieval data but does not execute semantic search. Use multimodal-dataprep-dev for source changes, tests, or image builds.
- ▌ Getitune Training A Model 2 · open-edge-platformTrain a computer-vision model with the getitune library (the Geti training library) using its Python API or CLI. Use when a user wants to train, fine-tune, or evaluate a model with `create_engine(...)` and `engine.train()/engine.test()`, run `getitune train`/`getitune test`, pick or override a recipe under `getitune.recipe.<task>`, choose a device (cpu/gpu/xpu/cuda), warm-start from a checkpoint, or debug a training run. Covers classification, detection, instance/semantic segmentation, and keypoint detection.
- ▌ Getitune Exporting A Model 2 · open-edge-platformExport a trained getitune model (the Geti training library) to a deployable format. Use when a user wants to run `engine.export(...)` or `getitune export`, choose between OpenVINO IR and ONNX, set FP32 vs FP16 precision with `ExportFormat` / `Precision`, or understand where exported artifacts are written and how they load back for inference. Covers the export/load contract between training and OpenVINO/ONNX inference.
- ▌ Getitune Running Inference 2 · open-edge-platformRun inference and evaluation with a getitune model (the Geti training library). Use when a user wants to call `engine.predict()` / `engine.test()` or `getitune predict` / `getitune test`, run inference with a PyTorch checkpoint versus an exported OpenVINO IR (`.xml`) or ONNX (`.onnx`) model, or understand how `OVEngine` loads deployed models via ModelAPI. Covers PyTorch, OpenVINO, and ONNX inference backends.
- ▌ Time Series Analytics User 2 · open-edge-platform bundleBuild a new time-series analytics use case on top of the deployed Time Series Analytics microservice — bring it up with Docker Compose (from a repo clone, or by fetching the compose files from GitHub when no clone exists) using the prebuilt intel/ia-time-series-analytics-microservice image, then author a UDF (Python) + TICKscript pair for the use case (threshold alerting, rate-of-change/spike detection, rolling-window anomaly detection, pretrained-model inference per point, or batch windowed inference over a time window), package it as a tar, deploy it via the REST API, and feed it data. Use when the user describes a sensor/metric monitoring or anomaly-detection scenario, wants to plug their own analytics logic or a trained scikit-learn model into a streaming or windowed-batch pipeline, or asks to wire up MQTT/OPC UA alerting on top of this service. Not for modifying the microservice's own source code — that is time-series-analytics-dev.
- ▌ Getitune Discovering Models 2 · open-edge-platformDiscover which models, recipes, and tasks the getitune library (the Geti training library) supports before training. Use when a user asks what models are available, how to list recipes, how to filter by task or name pattern, how `list_models(...)` and `getitune find` behave, or how to resolve the "model name matches multiple tasks" error. Covers classification, detection, instance/semantic segmentation, and keypoint detection recipes.
- ▌ Getitune Optimizing A Model 2 · open-edge-platformOptimize an exported getitune model (the Geti training library) with post-training quantization. Use when a user wants to run `OVEngine.optimize()` / `engine.optimize()` to produce an INT8 model via NNCF, understands calibration-set requirements, or needs to re-validate and run inference with a quantized model versus the original FP32/FP16 model. Covers OpenVINO NNCF post-training quantization and the accuracy/size trade-off.
- ▌ Getitune Preparing Datasets 2 · open-edge-platformPrepare and point datasets at the getitune library (the Geti training library) for training, testing, and prediction. Use when a user asks which dataset formats are supported, how the `data=` argument of `create_engine(...)` / `--data_root` works, why format auto-detection fails, how to lay out COCO/YOLO/Pascal VOC/Datumaro-native data, how to use a zip archive, or how to pass an Ultralytics YOLO `data.yaml`. Covers Datumaro-based auto-detection and per-task data expectations.
- ▌ Anomalib Adding A Datamodule 2 · open-edge-platform bundleAdds a new dataset/datamodule to anomalib under src/anomalib/data/. Use when wiring a new data source into the AnomalibDataset/AnomalibDataModule base classes, exporting it so anomalib.data.<Name> and the CLI/config (jsonargparse) can discover it, and adding matching tests. Do not use for model architecture work (see anomalib-adding-a-model) or for training an existing datamodule (see anomalib-training).
- ▌ Physicalai Train Adding A Policy 2 · open-edge-platform bundleAdds or modifies a Physical AI Studio policy under library/src/physicalai/policies. Use when creating a new policy family with the config/model/policy split, registering it in the get_policy factory and package exports, or keeping a policy compatible with Lightning training and export. Covers Pi0.5, Pi0, ACT, GR00T, SmolVLA, and LeRobot-wrapped policies.
- ▌ Multimodal Embedding Serving User 2 · open-edge-platform bundleDeploy and consume the Multimodal Embedding Serving microservice — bring it up with setup.sh + docker compose (from a repo clone, or by fetching those same files from GitHub when no clone exists) using the prebuilt intel/multimodal-embedding-serving image, embed text/images/videos over REST on port 9777, choose among 19 models (CLIP/SigLIP/MobileCLIP/CN-CLIP/Blip2/ QwenText), or integrate in-process via the Python SDK wheel. Use when an app needs embeddings for similarity search or retrieval. Not for modifying the service's source — that is multimodal-embedding-serving-dev.
- ▌ Physicalai Train Training A Policy 2 · open-edge-platformTrains, validates, tests, and runs prediction for Physical AI Studio policies via the library Lightning stack. Use when running physicalai fit/validate/test/predict, calling physicalai.train.Trainer and Policy APIs from Python, writing or editing YAML configs under library/configs, wiring a model + datamodule + trainer, resuming from a checkpoint, or debugging a training run. Covers ACT, Pi0, Pi0.5, GR00T, and SmolVLA.
- ▌ Physicalai Train Benchmarking A Policy 2 · open-edge-platformBenchmarks a trained Physical AI Studio policy in a simulation gym and reports success metrics. Use when running physicalai benchmark, editing configs under library/configs/benchmark, adding or changing a Benchmark class in physicalai.benchmark, tuning rollout/episode/env settings, recording rollout videos, or interpreting results.json / results.csv.
- ▌ Physicalai Train Working With Datasets 2 · open-edge-platformWorks with Physical AI Studio datasets and Lightning datamodules built on the LeRobot format. Use when wiring physicalai.data.lerobot.LeRobotDataModule into a training config, choosing a repo_id, converting between the physicalai and lerobot data layouts, defining observation Features/FeatureType, setting normalization, or debugging batch shapes and dataloading.
- ▌ Physicalai Train Exporting And Validating 2 · open-edge-platform bundleExports and validates Physical AI Studio policies for Runtime deployment. Use when working on policy.export(...), the physicalai export CLI, the ONNX/OpenVINO/Torch/ExecuTorch backends, export metadata, numerical parity checks, or the Studio side of the export/load contract that Runtime consumes with InferenceModel(...).
- ▌ Physicalai Runtime Adding A Camera Backend 2 · open-edge-platformAdds or modifies a camera backend under physicalai.capture. Use when implementing a new Camera type, extending create_camera in src/physicalai/capture/factory.py, discovery helpers, optional pip extras for vendor SDKs, SharedCamera transport, or tests under tests/unit/capture with fake devices.
- ▌ Physicalai Runtime Running Policy On Robot 2 · open-edge-platformRuns exported policies on hardware with PolicyRuntime, execution modes, and physicalai run. Use when wiring PolicyRuntime, SyncExecution or RTC execution, runtime YAML configs, action queues, runtime callbacks, or docs/how-to/runtime run-policy-on-robot and execution modes.
- ▌ Physicalai Runtime Loading Exported Policies 2 · open-edge-platform bundleLoads and validates policies exported from Physical AI Studio for Runtime deployment. Use when working on InferenceModel, InferenceModel.from_pretrained, manifest.json, adapter auto-detection (onnx, openvino), backend/device kwargs, Hugging Face Hub policy packages, or the Runtime side of the export/load contract that Studio produces with physicalai export.
- ▌ Physicalai Runtime Adding A Robot Integration 2 · open-edge-platform bundleAdds or modifies robot hardware integrations under physicalai.robot. Use when implementing the Robot protocol, SO101 or Trossen WidowX drivers, robot connect helpers, verify.py checks, optional extras so101 or trossen, or tests in tests/unit/robot.
- ▌ Physicalai Runtime Configuring Inference Pipeline 2 · open-edge-platformConfigures preprocessors, postprocessors, and runners around InferenceModel via manifest specs and ComponentRegistry. Use when editing physicalai.inference.preprocessors or postprocessors, manifest preprocessor/postprocessor lists, instantiate_component, registered type names, or class_path init_args for inference pipeline components.
- ▌ Physicalai Runtime Adding A Camera Backend · open-edge-platformAdds or modifies a camera backend under physicalai.capture. Use when implementing a new Camera type, extending create_camera in src/physicalai/capture/factory.py, discovery helpers, optional pip extras for vendor SDKs, SharedCamera transport, or tests under tests/unit/capture with fake devices.
- ▌ Physicalai Runtime Running Policy On Robot · open-edge-platformRuns exported policies on hardware with PolicyRuntime, execution modes, and physicalai run. Use when wiring PolicyRuntime, SyncExecution or RTC execution, runtime YAML configs, action queues, runtime callbacks, or docs/how-to/runtime run-policy-on-robot and execution modes.
- ▌ Physicalai Runtime Loading Exported Policies · open-edge-platform bundleLoads and validates policies exported from Physical AI Studio for Runtime deployment. Use when working on InferenceModel, InferenceModel.from_pretrained, manifest.json, adapter auto-detection (onnx, openvino), backend/device kwargs, Hugging Face Hub policy packages, or the Runtime side of the export/load contract that Studio produces with physicalai export.
- ▌ Physicalai Runtime Adding A Robot Integration · open-edge-platform bundleAdds or modifies robot hardware integrations under physicalai.robot. Use when implementing the Robot protocol, SO101 or Trossen WidowX drivers, robot connect helpers, verify.py checks, optional extras so101 or trossen, or tests in tests/unit/robot.
- ▌ Physicalai Runtime Configuring Inference Pipeline · open-edge-platformConfigures preprocessors, postprocessors, and runners around InferenceModel via manifest specs and ComponentRegistry. Use when editing physicalai.inference.preprocessors or postprocessors, manifest preprocessor/postprocessor lists, instantiate_component, registered type names, or class_path init_args for inference pipeline components.
- ▌ Metro AI App Recipe · open-edge-platform bundleStand up a complete, ready-to-run computer-vision analytics stack on Intel hardware with one Docker Compose command — point it at your video sources and an OpenVINO/ONNX model to get live annotated WebRTC video plus real-time detection dashboards and alerts (object detection, classification, counting, or zone/line-crossing) for any vertical, with no glue code. See "When to use this skill" for the full component list, trigger conditions, and boundaries.
- ▌ Metro AI App Builder · open-edge-platform bundleConversational orchestrator that turns a plain business objective into a working Intel Edge AI application by asking only business questions — never which framework, model, or device — then discovering the relevant open-edge-platform/skills, proposing a plan, and building the deliverable by DELEGATING to the right skill(s) after you confirm.
- ▌ Image Composer Build · open-edge-platform bundleUse when building OS disk images using the image-composer-tool and YAML template definitions for any supported OS/imageType.
- ▌ Image Composer Custom · open-edge-platform bundleCustomize image-composer templates with extra packages and repos, saving results to user-templates/ so canonical templates stay pristine. Use when users ask for package/repo injection without editing image-templates/.
- ▌ Image Composer List Os · open-edge-platform bundleList all buildable OS targets and template details from image-templates/. Use when users ask what OSes, distros, arches, or image types are available.
- ▌ Pre Install · open-edge-platformInstall pre-orchestrator infrastructure (OpenEBS, MetalLB, cluster setup)
- ▌
- ▌
- ▌
- ▌
- ▌
- ▌ Physicalai Train Adding A Policy · open-edge-platform bundleAdds or modifies a Physical AI Studio policy under library/src/physicalai/policies. Use when creating a new policy family with the config/model/policy split, registering it in the get_policy factory and package exports, or keeping a policy compatible with Lightning training and export. Covers Pi0.5, Pi0, ACT, GR00T, SmolVLA, and LeRobot-wrapped policies.
- ▌ Physicalai Train Training A Policy · open-edge-platformTrains, validates, tests, and runs prediction for Physical AI Studio policies via the library Lightning stack. Use when running physicalai fit/validate/test/predict, calling physicalai.train.Trainer and Policy APIs from Python, writing or editing YAML configs under library/configs, wiring a model + datamodule + trainer, resuming from a checkpoint, or debugging a training run. Covers ACT, Pi0, Pi0.5, GR00T, and SmolVLA.
- ▌ Studio Adding Robot Form UI Fields · open-edge-platformAdds a new interactive robot form UI field for plugin payload schemas. Use when introducing a new `robot_payload_ui` item kind, wiring renderer support under application/ui/src/features/robots/robot-form/robot-schema/components, updating plugin SDK UI-schema validation, and documenting how plugin authors adopt the field.
- ▌ Physicalai Train Benchmarking A Policy · open-edge-platformBenchmarks a trained Physical AI Studio policy in a simulation gym and reports success metrics. Use when running physicalai benchmark, editing configs under library/configs/benchmark, adding or changing a Benchmark class in physicalai.benchmark, tuning rollout/episode/env settings, recording rollout videos, or interpreting results.json / results.csv.
- ▌ Physicalai Train Working With Datasets · open-edge-platformWorks with Physical AI Studio datasets and Lightning datamodules built on the LeRobot format. Use when wiring physicalai.data.lerobot.LeRobotDataModule into a training config, choosing a repo_id, converting between the physicalai and lerobot data layouts, defining observation Features/FeatureType, setting normalization, or debugging batch shapes and dataloading.
- ▌ Physicalai Train Exporting And Validating · open-edge-platform bundleExports and validates Physical AI Studio policies for Runtime deployment. Use when working on policy.export(...), the physicalai export CLI, the ONNX/OpenVINO/Torch/ExecuTorch backends, export metadata, numerical parity checks, or the Studio side of the export/load contract that Runtime consumes with InferenceModel(...).
- ▌ Tme · open-edge-platformReport Intel Total Memory Encryption (TME/MKTME) support, activation state, and encryption algorithm. Use when asked about memory encryption on an Intel platform.
- ▌ Boot Guard · open-edge-platformReport Intel Boot Guard provisioning state, profile (0, 3, 4, or 5), and Verified/Measured Boot modes. Use when asked about Boot Guard, BtG profiles, or firmware root of trust.
- ▌ Secure Boot · open-edge-platformReport UEFI Secure Boot configuration on an Intel edge platform - enabled state, Setup Mode, and enrolled PK/KEK/db keys. Use when asked about Secure Boot status or boot chain key enrollment.
- ▌ Disk Encryption · open-edge-platformReport disk encryption configuration - LUKS/dm-crypt volumes, cipher and key size, swap encryption, and unencrypted volumes. Use when asked whether disks or swap are encrypted.
- ▌ Crypto Capabilities · open-edge-platformReport hardware crypto acceleration available on the CPU - AES-NI, SHA-NI, RDRAND, RDSEED. Use when asked which crypto instructions the platform supports.
- ▌ Trusted Compute Install · open-edge-platformInstall or uninstall the Intel Trusted Compute stack on bare-metal Ubuntu 24.04, via K3s or Docker. Use when asked to install, deploy, verify, or remove Trusted Compute packages.
- ▌ Create Image · open-edge-platformBuild a host OS image using the Image Composer Tool (ICT) from a source template with customizable user credentials.
- ▌ Set Power Profile · open-edge-platformSet how much power your Intel Core Ultra system may use — either a ready-made profile (LowPower 10 W, BalancedLow 15 W, BalancedHigh 20 W, Performance 25 W, or MaxPerformance = platform max / cTDP Level 2) or a Custom envelope with explicit PkgWatt (PL1) and optional SysWatt (psys) caps, burst ratio, and PL1 time window (tau). Runs locally with tools/power-tuning/set_power_profile.sh.
- ▌ Set Thermal Profile · open-edge-platformSet the thermal escalation policy of an Intel host with tools/power-tuning/set_thermal_profile.sh — generate, validate, apply and verify a thermald thermal-conf.xml that stages a response on the CPU package sensor (Fan active < Processor passive < intel_powerclamp passive). Choose a ready-made profile (cool 55/70/80, warm 60/75/85, hot 70/90/95, thermal-max 95/100/104, in °C for Fan/Processor/powerclamp) or a Custom set of trip points, optionally add the CHRG cooling device, and make thermald the sole thermal authority. Also supports disabling thermald to revert to kernel default thermal control.
- ▌ Monitor Power Thermal · open-edge-platformRun a live power and thermal monitor locally on an Intel host using tools/power-tuning/pt_mon.sh (turbostat), sampling package temperature and the RAPL power domains (PkgWatt, CorWatt, GFXWatt, RAMWatt, SysWatt) at a fixed interval and logging to pt_mon.txt. Useful for observing a power profile from set-power-profile under load from generate-platform-stress.
- ▌ Update Install Packages · open-edge-platformUpdate Ubuntu package configuration files for package add and delete operations.
- ▌ Generate Openvino Stress · open-edge-platformGenerate sustained AI inference load on CPU, GPU, or NPU using OpenVINO benchmark_app in a container (K3s pod or Docker). Produces real neural-network inference stress for power/thermal profiling — unlike stress-ng synthetic load. Auto-detects whether K3s or Docker is available; supports both runtimes. Pod specs are generated inline (no external YAML files). Ideal for validating power profiles under realistic AI workloads, measuring inference throughput per watt, and thermal qualification with real compute patterns.
- ▌ Generate Platform Stress · open-edge-platformCreate controlled CPU and integrated-GPU load on an Intel host to see how the platform behaves when it is busy. Choose how many CPU workers to run, how hard each one pushes (per-CPU load %), how many GPU workers to run, and how long the load lasts — then start it with a single command via tools/power-tuning/stress_gen.sh (stress-ng). Ideal for validating a power profile or power cap under real load, checking thermal and power headroom, and running repeatable burn-in or benchmarking workloads.
- ▌ Validate Platform Config · open-edge-platformValidate whether a provisioned platform is correctly configured when run locally on that host, including k3s pod health, binary paths, cloud-init state, network and proxy setup, and device readiness.
- ▌ Create Usb Installation Files · open-edge-platformPackage bootable USB installation artifacts (HookOS, host image, deployment scripts) from a standard Docker build, an ICT image, or a previously built image.
- ▌ Combined Power Thermal Profiling · open-edge-platformOrchestrate a full platform profiling session on an Intel host — apply a power envelope and thermal policy, start the power/thermal monitor, drive a bounded stress load, then summarize the result as a single enclosure report. Chains set-power-profile → set-thermal-profile → monitor-power-thermal → generate-platform-stress and emits one consolidated report (min/mean/max of PkgTmp / PkgWatt / GFXWatt plus throttle/headroom verdict). Ideal for qualifying whether an enclosure can sustain a chosen profile under load before it ships.
- ▌ Shell · open-edge-platformShell scripting standards for Scenescape — shebang, style, and Bash guidelines.
- ▌ Python · open-edge-platformPython coding standards for Scenescape — imports, indentation, patterns, and conventions.
- ▌ Makefile · open-edge-platformMakefile standards for Scenescape — build targets, conventions, and patterns.
- ▌ Security · open-edge-platformOn-demand security review skill for Scenescape — code and configuration security guidance.
- ▌ Javascript · open-edge-platformJavaScript coding standards for Scenescape — code style, conventions, and frontend patterns.
- ▌ Scenescape Setup · open-edge-platform bundleDeploy a working Intel® SceneScape installation from scratch (outside the repo). Gathers user-provided streams, camera IDs, scene name, and mapping choice, then runs bootstrap through tracking verification via scripts/deploy_scenescape.sh. Also handles re-running or resuming a single phase of an existing deployment on request (e.g. "recalibrate", "redo scene reconstruction", "resume bootstrap only") via the orchestrator's --phase flag.
- ▌ Documentation How · open-edge-platformProcedures for updating Scenescape documentation — where to make changes and what to update for each type of modification.
- ▌ Test Verification Gate · open-edge-platformRuntime test verification gate for Scenescape — image freshness checks, rebuild-before-test requirements, and retry policy.
- ▌ External Source Adapter · open-edge-platformWrite or review Scenescape external-source converter/adapter scripts that map native telemetry (MAVLink, ROS 2, NMEA, CAN, UWB/RTLS, proprietary) into the Scene Controller external_source MQTT contract. Use when the task involves external sources, publishers, adapters, converters, source_id, or scenescape/external topics.
- ▌ Geti Annotating And Managing Labels · open-edge-platformCreate projects, manage labels, and annotate media in the Geti application via its REST API. Use when a user wants to create a project with a task type and label set, add/edit/remove labels, upload images or videos, draw or set annotations (classification labels, bounding boxes, polygons) on media or video frames, review dataset statistics, or prepare a dataset so it is trainable.
- ▌ Geti Runtime Running Live Inference · open-edge-platformRun and operate live inference in the Geti application pipeline. Use when a user wants to start, monitor, stop, or recover source -> model -> sink runtime execution, verify production readiness, or troubleshoot live pipeline behavior such as stalls, dropped outputs, and latency regressions.
- ▌ Geti Runtime Configuring Inference Pipeline · open-edge-platformConfigure and validate the Geti runtime inference pipeline in application mode. Use when a user needs to set up or troubleshoot source → model → sink configuration, tune pipeline parameters, diagnose bad predictions or throughput issues, verify deployment settings, or move a project from trained model to stable runtime inference without changing backend implementation code.
- ▌ Geti UI Dev · open-edge-platformDevelop and validate changes in `application/ui/` for the React and TypeScript frontend. Use when touching `application/ui/src/**`, frontend tests, RSBuild or Vitest config, Playwright setup, package scripts, or generated API typings under `src/api`. Helps with Node and npm requirements, install and build commands, lint, typecheck, test workflows, and coordination with backend OpenAPI changes.
- ▌ Geti Library Dev · open-edge-platformDevelop and validate changes in `library/` for the `getitune` Python package. Use when changing library source or tests as well as packaging, recipes and model manifests. This includes Python APIs and CLI behavior across training and export. Covers environment setup; accelerator extras; multi-backend architecture; model and recipe additions; focused checks.
- ▌ Geti Backend Dev · open-edge-platformDevelop and validate changes in `application/backend/` for the FastAPI `geti` service. Use when changing backend source or tests as well as packaging and configuration. This includes API routers and schemas; services and repositories; database code; UI-facing backend contracts. Covers environment setup; targeted tests; OpenAPI generation; local server workflows.
- ▌ Geti Docs Update · open-edge-platformUpdate documentation (READMEs, application docs, or inline docstrings) to reflect changes, fixes, or new features. Use when a PR modifies behavior that should be documented for users or developers.
- ▌ Geti Openapi Sync · open-edge-platformRegenerate and validate the OpenAPI contract between `application/backend/` and `application/ui/`. Use when backend endpoints, schemas, request or response models, or API surface change, or when `application/ui/src/api/openapi-spec.json` or `openapi-spec.d.ts` is stale. Handles backend spec generation, UI spec placement, TypeScript type regeneration, and the smallest backend and UI checks needed to confirm the contract still matches.
- ▌ Getitune Training A Model · open-edge-platformTrain a computer-vision model with the getitune library (the Geti training library) using its Python API or CLI. Use when a user wants to train, fine-tune, or evaluate a model with `create_engine(...)` and `engine.train()/engine.test()`, run `getitune train`/`getitune test`, pick or override a recipe under `getitune.recipe.<task>`, choose a device (cpu/gpu/xpu/cuda), warm-start from a checkpoint, or debug a training run. Covers classification, detection, instance/semantic segmentation, and keypoint detection.
- ▌ Getitune Exporting A Model · open-edge-platformExport a trained getitune model (the Geti training library) to a deployable format. Use when a user wants to run `engine.export(...)` or `getitune export`, choose between OpenVINO IR and ONNX, set FP32 vs FP16 precision with `ExportFormat` / `Precision`, or understand where exported artifacts are written and how they load back for inference. Covers the export/load contract between training and OpenVINO/ONNX inference.
- ▌ Getitune Running Inference · open-edge-platformRun inference and evaluation with a getitune model (the Geti training library). Use when a user wants to call `engine.predict()` / `engine.test()` or `getitune predict` / `getitune test`, run inference with a PyTorch checkpoint versus an exported OpenVINO IR (`.xml`) or ONNX (`.onnx`) model, or understand how `OVEngine` loads deployed models via ModelAPI. Covers PyTorch, OpenVINO, and ONNX inference backends.
- ▌ Geti Using The Pipeline · open-edge-platformUse the Geti application end to end through its REST API — the project → dataset → annotate → train → deploy pipeline served by the FastAPI backend in `application/backend/`. Use when a user (not a contributor) wants to create a project, upload media, add annotations, launch a training or quantization job, track job status, configure a source → model → sink inference pipeline, and enable live inference. Covers the `/api/...` endpoints and the async job model, not backend code changes.
- ▌ Getitune Discovering Models · open-edge-platformDiscover which models, recipes, and tasks the getitune library (the Geti training library) supports before training. Use when a user asks what models are available, how to list recipes, how to filter by task or name pattern, how `list_models(...)` and `getitune find` behave, or how to resolve the "model name matches multiple tasks" error. Covers classification, detection, instance/semantic segmentation, and keypoint detection recipes.
- ▌ Getitune Optimizing A Model · open-edge-platformOptimize an exported getitune model (the Geti training library) with post-training quantization. Use when a user wants to run `OVEngine.optimize()` / `engine.optimize()` to produce an INT8 model via NNCF, understands calibration-set requirements, or needs to re-validate and run inference with a quantized model versus the original FP32/FP16 model. Covers OpenVINO NNCF post-training quantization and the accuracy/size trade-off.
- ▌ Getitune Preparing Datasets · open-edge-platformPrepare and point datasets at the getitune library (the Geti training library) for training, testing, and prediction. Use when a user asks which dataset formats are supported, how the `data=` argument of `create_engine(...)` / `--data_root` works, why format auto-detection fails, how to lay out COCO/YOLO/Pascal VOC/Datumaro-native data, how to use a zip archive, or how to pass an Ultralytics YOLO `data.yaml`. Covers Datumaro-based auto-detection and per-task data expectations.
- ▌ Geti Import Export Datasets · open-edge-platformImport and export datasets in the Geti application via its REST API. Use when a user wants to bring an existing dataset into Geti (COCO/YOLO/VOC/Datumaro), migrate data from another tool, export a project's dataset or a dataset revision, remap labels or filter subsets during transfer, or track the async jobs that perform these operations.
- ▌ Testing · open-edge-platform bundleReview/generate unit, integration, and regression test expectations