所有 Skills

"Two-step image grounding pipeline: extracts referring expressions from (image, caption) pairs and grounds them

tao-analyze-changenet-rca

Performs deep Root Cause Analysis (RCA) on NVIDIA TAO Visual ChangeNet classification experiments with

tao-train-ocrnet

OCRNet for scene text recognition. Recognizes text content from cropped text-region images and supports CTC

tao-train-optical-inspection

Optical Inspection for defect detection using Siamese networks. Compares image pairs to detect manufacturing

tao-run-inference-service

3.1Kbackend-api

Start, query, and stop a network-specific TAO inference microservice ({network_arch}-inference-microservice) by delegating container execution to the appropriate platform skill. Handles container image resolution, job-payload JSON construction, and the service registry. Use when the user wants to run inference on a TAO model checkpoint using a microservice container, deploy a TAO inference endpoint, or stop a running inference container.

tao-analyze-gaps-visual-changenet

3.1Kwriting-content

Performs gap analysis on NVIDIA TAO VCN Classify (Visual Component Net) experiments by invoking the data-services container (`tao_toolkit.data_services` from `versions.yaml`) directly via `docker run … gap_analysis vcn_aoi …` — picks the optimal decision threshold, ranks per-sample weakness, and emits a top-K weakest parquet expanded per-lighting for downstream augmentation. Use when analyzing VCN classification failures, picking SDA augmentation targets, or auditing PASS/NO_PASS boundary cases.

swiftui-expert-skill

3.1Kdesign-ui

适用于编写、审查或重构针对 iOS 和 macOS 的 SwiftUI 代码，涵盖状态管理与 `@Observable` 数据流、视图组合与失效/性能优化、列表与 `ForEach` 标识符管理、Environment 环境上下文用法、本地化、动画、Liquid Glass 样式适配、软弃用 API 迁移，以及使用 Instruments 捕获与分析 `.trace` 文件（定位主线程卡顿、掉帧、CPU 热点或视图过度刷新问题）。

avdlee

google-agents-cli-adk-code

当用户想要“编写智能体代码”、“使用ADK构建智能体”、“添加工具”、“创建回调”、“定义智能体”、“使用状态管理”，或需要ADK（智能体开发工具包）Python API模式和代码示例时，应使用此技能。它是Google ADK技能套件的一部分。它提供了智能体类型、工具定义、编排模式、回调和状态管理的快速参考。不要用于创建新项目（使用google-agents-cli-scaffold）或部署（使用google-agents-cli-deploy）。

google-agents-cli-workflow

此技能应在用户想要“开发智能体”、“使用 ADK 构建智能体”、“本地运行智能体”、“调试智能体代码”、“测试智能体”、“部署智能体”、“发布智能体”、“监控智能体”，或需要 ADK（智能体开发工具包）开发生命周期和编码指南时使用。构建 ADK 智能体的入口点。始终处于活动状态——提供完整的工作流程（脚手架、构建、评估、部署、发布、观察）、代码保留规则、模型选择指南以及 ADK 或任何智能体开发的故障排除步骤。

google-agents-cli-eval

当用户想要“运行评估”、“评估我的ADK智能体”、“编写评估数据集”、“分析评估失败原因”、“比较评估结果”、“优化智能体”，或需要Agent Platform评估方法论和质量飞轮方面的指导时，应使用此技能。涵盖评估指标、数据集模式、LLM作为评判的评分以及常见失败原因。不要用于API代码模式（使用google-agents-cli-adk-code）、部署（使用google-agents-cli-deploy）或项目脚手架（使用google-agents-cli-scaffold）。

google-agents-cli-scaffold

当用户想要“创建 Agent 项目”、“启动新的 ADK 项目”、“帮我构建一个新 Agent”、“为项目添加 CI/CD”、“添加部署配置”、“增强现有项目”或“升级项目”时，应当使用此 Skill。本 Skill 属于 Google ADK（Agent Development Kit）Skill 套件的一部分。涵盖 `agents-cli scaffold create`、`scaffold enhance` 和 `scaffold upgrade` 命令、模板选项、部署目标以及原型优先（prototype-first）工作流。请勿用于编写 Agent 代码（请使用 google-agents-cli-adk-code）或部署运维操作（请使用 google-agents-cli-deploy）。

google-agents-cli-observability

当用户想要“设置追踪”、“监控我的ADK代理”、“配置日志记录”、“添加可观测性”、“调试生产流量”，或需要关于监控已部署的ADK（Agent Development Kit）代理的指导时，应使用此技能。涵盖Cloud Trace、提示-响应日志记录、BigQuery Agent Analytics、第三方集成（AgentOps、Phoenix、MLflow等）以及故障排除。属于Google ADK（Agent Development Kit）技能套件的一部分。不要用于部署设置（使用google-agents-cli-deploy）或API代码模式（使用google-agents-cli-adk-code）。

google-agents-cli-deploy

当用户想要“部署 Agent”、“部署 ADK Agent”、“配置 CI/CD”、“配置 Secret 密钥”、“排查部署故障”，或需要关于 Agent Runtime、Cloud Run 或 GKE 部署目标的指导时，应使用此 Skill。涵盖部署工作流、服务账号、版本回滚和生产基础设施。属于 Google ADK (Agent Development Kit) Skill 套件的一部分。切勿用于 API 代码模式（请使用 google-agents-cli-adk-code）、评估测试（请使用 google-agents-cli-eval）或项目脚手架搭建（请使用 google-agents-cli-scaffold）。

google-agents-cli-publish

当用户想要“发布智能体”、“发布我的ADK智能体”、“向Gemini Enterprise注册智能体”、“发布到Gemini Enterprise”，或需要关于agents-cli publish gemini-enterprise命令的指导时，应使用此技能。当用户想要“在Agent Registry中管理智能体”或“列出/更新/删除已注册的智能体”时，也应使用此技能。涵盖ADK与A2A注册模式、编程式和交互式用法、标志参考、从部署元数据自动检测、Agent Registry集群管理以及故障排除。属于Google ADK（Agent Development Kit）技能套件的一部分。请勿用于部署（请使用google-agents-cli-deploy）。

ultracite

3.1Kfrontend

Ultracite 是一个零配置的 JavaScript/TypeScript 项目代码检查与格式化预设。适用于以下场景：(1) 在项目中设置或初始化 Ultracite（ultracite init），(2) 运行代码检查或格式化命令（check、fix、doctor），(3) 在使用了 Ultracite 的项目中编写或审查 JS/TS 代码——遵循其代码规范，(4) 排查代码检查/格式化问题，(5) 用户在安装了 Ultracite 的项目中提及 'ultracite'、'lint'、'format'、'code quality' 或 'biome/eslint/oxlint'。

haydenbleasel

tao-setup-nvidia-gpu-host

Host setup for TAO GPU backends. Checks and, after user approval, installs NVIDIA driver branch 580, CUDA Toolkit 13.0, and NVIDIA Container Toolkit 1.19.0 for Docker/local-Docker and Kubernetes GPU worker hosts. The `--check-only` path works on any Linux distribution; `--install` automates debian-family (Ubuntu/Debian/Pop!_OS/Mint/Zorin/Raspbian), rhel-family (Fedora/RHEL/Rocky/AlmaLinux), and suse-family (openSUSE/SLES) hosts, and prints actionable manual-install steps for everything else. Use when the user asks to "set up an NVIDIA GPU host", "check TAO Docker GPU runtime", or prepare a Kubernetes GPU worker for TAO.

tao-finetune-huggingface-model

Fine-tune any HuggingFace CV / VLM / LLM model on local NVIDIA GPUs inside an NGC PyTorch container. Use when the user wants to fine-tune a HuggingFace model (full or LoRA), train a vision / VLM / LLM model end-to-end, generate a reproducible HF training pipeline, smoke-test a HuggingFace model locally before scale-up, push a fine-tuned model to the HF Hub with a model card, or emit a self-contained rerun skill for an existing HuggingFace finetune. Supports image classification, object detection, semantic / instance / panoptic segmentation, depth estimation, image-text-to-text VLM (SFT / LoRA), and LLM SFT / DPO / GRPO. Six-step workflow: inspect and qualify, hardware and NGC image, research, generate and smoke, train + eval + infer, push and emit rerun skill.

tao-run-on-local-docker

3.1Kbackend-api

Local or remote Docker execution for TAO SDK job containers using a Docker daemon with NVIDIA GPU runtime. Use

tao-finetune-cosmos-reason

Cosmos3-Nano video QA supervised fine-tuning with FSDP parallelism. Use when training or evaluating video

tao-finetune-cosmos-embed

Cosmos-Embed1 video-text embedding for text-to-video retrieval, video-to-video search, semantic deduplication, and fine-tuning. Use when the user asks to "fine-tune Cosmos-Embed1", "run cosmos-embed inference", "export Cosmos-Embed1", "embed videos", or "search videos with text".

tao-launch-workflow

Shared launch intake for any TAO workflow or action. Use when the user wants to run TAO AutoML, train, evaluate, infer, export, generate TensorRT engines, or launch DEFT/workflow jobs on an execution platform.

hsb-setup

Clone the latest NVIDIA Holoscan Sensor Bridge repo, ask which supported devkit is being used, configure the host per platform, build the correct demo container, run it, and verify HSB connectivity by pinging 192.168.0.2. Use for Holoscan Sensor Bridge setup, build, container launch, and first-connectivity bring-up.

hsb-test

3.1Ktesting-qa

Execute QA test plans on Holoscan Sensor Bridge hardware. Reads a user-provided test document, filters tests by the user's setup, determines which tests can run automatically, executes them with pass/fail evaluation, and produces a structured test results report.

hsb-flash

Flash the FPGA on an HSB board connected to an NVIDIA devkit. Supports HSB Lattice boards (FPGA versions 2407, 2412, 2507, 2510) and Leopard Imaging VB1940 "all-in-one" cameras (FPGA versions 2507, 2510). Uses release-specific YAML manifests and board-type-specific program commands. Lattice and VB1940 commands must never be mixed.