diff --git a/.github/scripts/marketplace/metadata.json b/.github/scripts/marketplace/metadata.json new file mode 100644 index 00000000..3e85d28f --- /dev/null +++ b/.github/scripts/marketplace/metadata.json @@ -0,0 +1,844 @@ +{ + "skills": [ + { + "path": "skills/cuopt-numerical-optimization-api-cli", + "name": "cuopt-numerical-optimization-api-cli", + "description": "LP, MILP, and QP (beta) with cuOpt \u2014 CLI only (MPS files, cuopt_cli). Use when the user is solving LP, MILP, or QP from MPS via command line.", + "metadata": { + "product.primary": "cuOpt", + "classification.category.primary": "accelerated_computing", + "catalog.subdomain": "decision-optimization", + "audience": "developer,application_developer", + "discovery.activity_tags": "configure,optimize,troubleshoot,validate" + } + }, + { + "path": "skills/cuopt-routing-api-python", + "name": "cuopt-routing-api-python", + "description": "Vehicle routing (VRP, TSP, PDP) with cuOpt \u2014 Python API only. Use when the user is building or solving routing in Python.", + "metadata": { + "product.primary": "cuOpt", + "classification.category.primary": "accelerated_computing", + "catalog.subdomain": "decision-optimization", + "audience": "developer,application_developer", + "discovery.activity_tags": "configure,optimize,validate,debug" + } + }, + { + "path": "skills/cuopt-user-rules", + "name": "cuopt-user-rules", + "description": "Base rules for end users calling NVIDIA cuOpt (routing/LP/MILP/QP/install/server). Not for cuOpt internals \u2014 use cuopt-developer for those.", + "metadata": { + "product.primary": "cuOpt", + "classification.category.primary": "accelerated_computing", + "catalog.subdomain": "decision-optimization", + "audience": "developer,application_developer", + "discovery.activity_tags": "configure,validate,troubleshoot,deploy" + } + }, + { + "path": "skills/cuopt-developer", + "name": "cuopt-developer", + "description": "Modify, build, test, debug, and contribute to NVIDIA cuOpt (C++/CUDA, Python, server, CI). Use for solver internals, PRs, DCO, and code conventions.", + "metadata": { + "product.primary": "cuOpt", + "classification.category.primary": "developer_tools", + "catalog.subdomain": "decision-optimization", + "audience": "developer,hpc_developer", + "discovery.activity_tags": "contribute,test,debug,extend" + } + }, + { + "path": "skills/cuopt-install", + "name": "cuopt-install", + "description": "Install cuOpt for Python, C, or server via pip, conda, or Docker; verify the install. For building cuOpt from source, see cuopt-developer.", + "metadata": { + "product.primary": "cuOpt", + "classification.category.primary": "accelerated_computing", + "catalog.subdomain": "decision-optimization", + "audience": "developer,application_developer,devops_engineer", + "discovery.activity_tags": "get_started,configure,validate,deploy" + } + }, + { + "path": "skills/cuopt-numerical-optimization-api-python", + "name": "cuopt-numerical-optimization-api-python", + "description": "Solve LP, MILP, QP (beta) with cuOpt Python API \u2014 linear/quadratic objectives, integer variables, scheduling, portfolio, least squares.", + "metadata": { + "product.primary": "cuOpt", + "classification.category.primary": "accelerated_computing", + "catalog.subdomain": "decision-optimization", + "audience": "developer,application_developer", + "discovery.activity_tags": "configure,optimize,debug,validate" + } + }, + { + "path": "skills/cuopt-server-api-python", + "name": "cuopt-server-api-python", + "description": "cuOpt REST server \u2014 start server, endpoints, Python/curl client examples. Use when the user is deploying or calling the REST API.", + "metadata": { + "product.primary": "cuOpt", + "classification.category.primary": "accelerated_computing", + "catalog.subdomain": "decision-optimization", + "audience": "developer,application_developer,devops_engineer", + "discovery.activity_tags": "deploy,configure,integrate,validate,debug" + } + }, + { + "path": "skills/cuopt-numerical-optimization-formulation", + "name": "cuopt-numerical-optimization-formulation", + "description": "LP, MILP, QP \u2014 concepts, problem-text parsing, and formulation patterns (parameters, constraints, decisions, objective). Concepts only; no API.", + "metadata": { + "product.primary": "cuOpt", + "classification.category.primary": "accelerated_computing", + "catalog.subdomain": "decision-optimization", + "audience": "developer,application_developer,data_scientist", + "discovery.activity_tags": "assess,select,optimize,validate" + } + }, + { + "path": "skills/cuopt-routing-formulation", + "name": "cuopt-routing-formulation", + "description": "Vehicle routing (VRP, TSP, PDP) \u2014 problem types and data requirements. Domain concepts; no API or interface.", + "metadata": { + "product.primary": "cuOpt", + "classification.category.primary": "accelerated_computing", + "catalog.subdomain": "decision-optimization", + "audience": "developer,application_developer,data_scientist", + "discovery.activity_tags": "assess,select,optimize,validate" + } + }, + { + "path": "skills/cuopt-skill-evolution", + "name": "cuopt-skill-evolution", + "description": "After solving a non-trivial problem, detect generalizable learnings and propose skill updates. Always active \u2014 applies to every interaction.", + "metadata": { + "product.primary": "cuOpt", + "classification.category.primary": "developer_tools", + "catalog.subdomain": "decision-optimization", + "audience": "developer", + "discovery.activity_tags": "evaluate,extend,contribute" + } + }, + { + "path": "skills/cuopt-numerical-optimization-api-c", + "name": "cuopt-numerical-optimization-api-c", + "description": "LP, MILP, and QP (beta) with cuOpt \u2014 C API only. Use when the user is embedding LP, MILP, or QP in C/C++.", + "metadata": { + "product.primary": "cuOpt", + "classification.category.primary": "accelerated_computing", + "catalog.subdomain": "decision-optimization", + "audience": "developer,application_developer,hpc_developer", + "discovery.activity_tags": "integrate,optimize,debug,validate" + } + }, + { + "path": "skills/cuopt-server-common", + "name": "cuopt-server-common", + "description": "cuOpt REST server \u2014 what it does and how requests flow. Domain concepts; no deploy or client code.", + "metadata": { + "product.primary": "cuOpt", + "classification.category.primary": "accelerated_computing", + "catalog.subdomain": "decision-optimization", + "audience": "developer,application_developer,devops_engineer,solutions_architect", + "discovery.activity_tags": "assess,select,configure,validate" + } + }, + { + "path": "skills/aiq-deploy", + "name": "aiq-deploy", + "description": "Use when asked to install, deploy, run, validate, troubleshoot, or stop NVIDIA AI-Q Blueprint infrastructure.\n", + "metadata": { + "product.primary": "NeMo Agent Toolkit", + "classification.category.primary": "ai_and_machine_learning", + "catalog.subdomain": "agentic-ai", + "audience": "developer,devops_engineer,platform_engineer", + "discovery.activity_tags": "deploy,configure,validate,troubleshoot,operate" + } + }, + { + "path": "skills/aiq-research", + "name": "aiq-research", + "description": "Use when asked to run deep research or AI-Q research through a reachable NVIDIA AI-Q Blueprint backend.\n", + "metadata": { + "product.primary": "NeMo Agent Toolkit", + "classification.category.primary": "ai_and_machine_learning", + "catalog.subdomain": "agentic-ai", + "audience": "developer,ai_engineer", + "discovery.activity_tags": "generate,summarize,validate" + } + }, + { + "path": "skills/deepstream-dev", + "name": "deepstream-dev", + "description": "NVIDIA DeepStream SDK 9.0 development with Python pyservicemaker API. Use when building video analytics pipelines, GStreamer-based video processing, TensorRT inference integration, object detection/tracking, or Kafka/message broker integration.", + "metadata": { + "product.primary": "DeepStream SDK", + "classification.category.primary": "ai_and_machine_learning", + "catalog.subdomain": "vision-ai", + "audience": "developer,application_developer,ai_engineer", + "discovery.activity_tags": "configure,integrate,debug,troubleshoot,deploy" + } + }, + { + "path": "skills/deepstream-import-vision-model", + "name": "deepstream-import-vision-model", + "description": "Use this skill to bring any vision model from HuggingFace or NVIDIA NGC into an NVIDIA DeepStream pipeline with end-to-end automation: ONNX download, SafeTensors export, TRT engine build, custom nvinfer bbox parser, multi-stream benchmark, and PDF report. Object detection models only.\n", + "metadata": { + "product.primary": "DeepStream SDK", + "classification.category.primary": "ai_and_machine_learning", + "catalog.subdomain": "vision-ai", + "audience": "developer,application_developer,ai_engineer,ml_engineer", + "discovery.activity_tags": "integrate,convert,benchmark,validate,deploy" + } + }, + { + "path": "skills/nemo-automodel-distributed-training", + "name": "nemo-automodel-distributed-training", + "description": "Guide for selecting and configuring distributed training strategies in NeMo AutoModel, including FSDP2, Megatron FSDP, DDP, and parallelism settings.", + "metadata": { + "product.primary": "NeMo Framework", + "classification.category.primary": "ai_and_machine_learning", + "catalog.subdomain": "training-ai", + "audience": "developer,ai_engineer,ml_engineer,hpc_developer", + "discovery.activity_tags": "train,configure,scale,optimize,validate" + } + }, + { + "path": "skills/nemo-automodel-launcher-config", + "name": "nemo-automodel-launcher-config", + "description": "Configure NeMo AutoModel job launches for interactive runs, Slurm clusters, and SkyPilot cloud execution.", + "metadata": { + "product.primary": "NeMo Framework", + "classification.category.primary": "ai_and_machine_learning", + "catalog.subdomain": "training-ai", + "audience": "developer,ai_engineer,ml_engineer,devops_engineer", + "discovery.activity_tags": "configure,deploy,orchestrate,scale,validate" + } + }, + { + "path": "skills/nemo-automodel-model-onboarding", + "name": "nemo-automodel-model-onboarding", + "description": "Guide for onboarding new model architectures into NeMo AutoModel, including architecture discovery, implementation patterns, registration, and validation.", + "metadata": { + "product.primary": "NeMo Framework", + "classification.category.primary": "ai_and_machine_learning", + "catalog.subdomain": "training-ai", + "audience": "developer,ai_engineer,ml_engineer", + "discovery.activity_tags": "extend,integrate,test,validate,debug" + } + }, + { + "path": "skills/nemo-automodel-recipe-development", + "name": "nemo-automodel-recipe-development", + "description": "Create and modify NeMo AutoModel training and evaluation recipes, including YAML structure, builders, and execution flow.", + "metadata": { + "product.primary": "NeMo Framework", + "classification.category.primary": "ai_and_machine_learning", + "catalog.subdomain": "training-ai", + "audience": "developer,ai_engineer,ml_engineer", + "discovery.activity_tags": "configure,train,evaluate,extend,validate" + } + }, + { + "path": "skills/nemoclaw-user-agent-skills", + "name": "nemoclaw-user-agent-skills", + "description": "Describes the agent skills shipped with NemoClaw and how to access them by cloning the repository. Use when users ask about AI agent support, coding assistant integration, or the .agents/skills/ directory. Trigger keywords - nemoclaw agent skills, ai coding assistant, cursor, claude code, copilot.", + "metadata": { + "product.primary": "NeMoClaw", + "classification.category.primary": "ai_and_machine_learning", + "catalog.subdomain": "agentic-ai", + "audience": "developer,application_developer,ai_engineer", + "discovery.activity_tags": "get_started,integrate,select,configure" + } + }, + { + "path": "skills/nemoclaw-user-configure-inference", + "name": "nemoclaw-user-configure-inference", + "description": "Connects NemoClaw to a local inference server. Use when setting up Ollama, vLLM, TensorRT-LLM, NIM, or any OpenAI-compatible local model server with NemoClaw. Trigger keywords - nemoclaw local inference, ollama nemoclaw, vllm nemoclaw, local model server, openai compatible endpoint, switch nemoclaw inference model, change inference runtime, nemoclaw additional model, nemoclaw sub-agent model, openclaw sub-agent, agents.list, sessions_spawn, vlm-demo, nemoclaw tool calling, ollama tool calls, vllm tool-call-parser, raw json in tui, nemoclaw inference options, nemoclaw onboarding providers, nemoclaw inference routing.", + "metadata": { + "product.primary": "NeMoClaw", + "classification.category.primary": "ai_and_machine_learning", + "catalog.subdomain": "agentic-ai", + "audience": "developer,application_developer,ai_engineer,platform_engineer", + "discovery.activity_tags": "configure,inference,integrate,validate,troubleshoot" + } + }, + { + "path": "skills/nemoclaw-user-configure-security", + "name": "nemoclaw-user-configure-security", + "description": "Presents a risk framework for every configurable security control in NemoClaw. Use when evaluating security posture, reviewing sandbox security defaults, or assessing control trade-offs. Trigger keywords - nemoclaw security best practices, sandbox security controls risk framework, nemoclaw credential storage, openshell provider, api key security, openclaw security controls, nemoclaw security boundary, prompt injection, tool access control.", + "metadata": { + "product.primary": "NeMoClaw", + "classification.category.primary": "ai_and_machine_learning", + "catalog.subdomain": "agentic-ai", + "audience": "developer,platform_engineer,security_engineer,devops_engineer", + "discovery.activity_tags": "assess,configure,validate,troubleshoot" + } + }, + { + "path": "skills/nemoclaw-user-deploy-remote", + "name": "nemoclaw-user-deploy-remote", + "description": "Explains how to run NemoClaw on a remote GPU instance, including the deprecated Brev compatibility path and the preferred installer plus onboard flow. Use when deploying NemoClaw to a remote VM, onboarding a Brev instance, or migrating away from the legacy `nemoclaw deploy` wrapper. Trigger keywords - deploy nemoclaw remote gpu, nemoclaw brev cloud deployment, nemoclaw plugins, openclaw plugins, install openclaw plugin, nemoclaw onboard from dockerfile, nemoclaw brev web ui, nemoclaw getting started, brev quickstart, nvidia nemotron agent, nemoclaw sandbox hardening, container security, docker capabilities, process limits.", + "metadata": { + "product.primary": "NeMoClaw", + "classification.category.primary": "ai_and_machine_learning", + "catalog.subdomain": "agentic-ai", + "audience": "developer,devops_engineer,platform_engineer", + "discovery.activity_tags": "deploy,configure,validate,troubleshoot,operate" + } + }, + { + "path": "skills/nemoclaw-user-get-started", + "name": "nemoclaw-user-get-started", + "description": "Installs NemoClaw, launches a sandbox, and runs the first agent prompt. Use when onboarding, installing, or launching a NemoClaw sandbox for the first time. Trigger keywords - nemoclaw quickstart, install nemoclaw openclaw sandbox, nemohermes quickstart, hermes agent nemoclaw, run hermes openshell sandbox, nemoclaw prerequisites, nemoclaw supported platforms, nemoclaw hardware software, nemoclaw windows wsl2 setup, nemoclaw install windows docker desktop.", + "metadata": { + "product.primary": "NeMoClaw", + "classification.category.primary": "ai_and_machine_learning", + "catalog.subdomain": "agentic-ai", + "audience": "developer,application_developer,ai_engineer", + "discovery.activity_tags": "get_started,deploy,configure,validate" + } + }, + { + "path": "skills/nemoclaw-user-manage-policy", + "name": "nemoclaw-user-manage-policy", + "description": "Adds, removes, or modifies allowed endpoints in the sandbox policy. Use when customizing network policy, changing egress rules, or configuring sandbox endpoint access. Trigger keywords - customize nemoclaw network policy, sandbox egress policy configuration, nemoclaw integration policy examples, post-install policy setup, openshell approval workflow, policy preset, nemoclaw approve network requests, sandbox egress approval tui.", + "metadata": { + "product.primary": "NeMoClaw", + "classification.category.primary": "ai_and_machine_learning", + "catalog.subdomain": "agentic-ai", + "audience": "developer,platform_engineer,security_engineer,devops_engineer", + "discovery.activity_tags": "configure,operate,troubleshoot,validate" + } + }, + { + "path": "skills/nemoclaw-user-manage-sandboxes", + "name": "nemoclaw-user-manage-sandboxes", + "description": "Explains operational tasks after the quickstart: listing sandboxes, status and health checks, logs, diagnostics, port forwards, multiple sandboxes, credential reset, rebuilds, network presets, upgrades, and uninstall. Trigger keywords - manage nemoclaw sandboxes, nemoclaw status, nemoclaw list, nemoclaw dashboard port, nemoclaw rebuild, nemoclaw upgrade sandboxes, nemoclaw uninstall, sandbox mutability, sandbox runtime configuration, sandbox rebuild, nemoclaw backup, nemoclaw restore, workspace backup, openshell sandbox download upload, nemoclaw messaging channels, nemoclaw telegram, nemoclaw discord, nemoclaw slack, nemoclaw wechat, nemoclaw whatsapp, openshell channel messaging, nemoclaw workspace files, soul.md, user.md, identity.md, agents.md, sandbox persistence.", + "metadata": { + "product.primary": "NeMoClaw", + "classification.category.primary": "ai_and_machine_learning", + "catalog.subdomain": "agentic-ai", + "audience": "developer,devops_engineer,platform_engineer", + "discovery.activity_tags": "operate,monitor,troubleshoot,configure,recover" + } + }, + { + "path": "skills/nemoclaw-user-monitor-sandbox", + "name": "nemoclaw-user-monitor-sandbox", + "description": "Inspects sandbox health, traces agent behavior, and diagnoses problems. Use when monitoring a running sandbox, debugging agent issues, or checking sandbox logs. Trigger keywords - monitor nemoclaw sandbox, debug nemoclaw agent issues.", + "metadata": { + "product.primary": "NeMoClaw", + "classification.category.primary": "ai_and_machine_learning", + "catalog.subdomain": "agentic-ai", + "audience": "developer,devops_engineer,platform_engineer", + "discovery.activity_tags": "monitor,debug,troubleshoot,inspect" + } + }, + { + "path": "skills/nemoclaw-user-overview", + "name": "nemoclaw-user-overview", + "description": "Explains how OpenClaw, OpenShell, and NemoClaw form the ecosystem, NemoClaw's position in the stack, what NemoClaw adds beyond the community sandbox, and when to prefer NemoClaw versus integrating OpenShell and OpenClaw directly. Use when users ask about the relationship between OpenClaw, OpenShell, and NemoClaw, or when to use NemoClaw versus OpenShell. Trigger keywords - nemoclaw ecosystem, openclaw openshell, nemoclaw vs openshell, sandboxed openclaw, how nemoclaw works, nemoclaw sandbox lifecycle blueprint, nemoclaw overview, openclaw always-on assistants, nvidia openshell, nvidia nemotron, nemoclaw release notes, nemoclaw changelog.", + "metadata": { + "product.primary": "NeMoClaw", + "classification.category.primary": "ai_and_machine_learning", + "catalog.subdomain": "agentic-ai", + "audience": "developer,application_developer,solutions_architect", + "discovery.activity_tags": "assess,select,get_started,summarize" + } + }, + { + "path": "skills/nemoclaw-user-reference", + "name": "nemoclaw-user-reference", + "description": "Describes the NemoClaw plugin and blueprint architecture and how they orchestrate the OpenClaw sandbox. Use when looking up architecture, plugin structure, or blueprint design. Trigger keywords - nemoclaw architecture, nemoclaw plugin blueprint structure, nemoclaw vs openshell, which cli, nemoclaw cli, openshell cli, sandbox commands, nemoclaw cli commands, nemoclaw command reference, nemoclaw network policy, sandbox egress control operator approval, nemoclaw troubleshooting, nemoclaw debug sandbox issues.", + "metadata": { + "product.primary": "NeMoClaw", + "classification.category.primary": "ai_and_machine_learning", + "catalog.subdomain": "agentic-ai", + "audience": "developer,application_developer,platform_engineer,solutions_architect", + "discovery.activity_tags": "assess,select,configure,troubleshoot" + } + }, + { + "path": "skills/skill-card-generator", + "name": "skill-card-generator", + "description": "Use only to generate or update a governance skill card for a specified existing agent skill directory. Do not use for explaining, listing, comparing, or discussing skill capabilities.", + "metadata": { + "product.primary": "Trustworthy AI", + "classification.category.primary": "developer_tools", + "catalog.subdomain": "agentic-ai", + "audience": "developer,platform_engineer,solutions_architect", + "discovery.activity_tags": "generate,validate,inspect,assess" + } + }, + { + "path": "skills/nemo-mbridge-recipe-recommender", + "name": "nemo-mbridge-recipe-recommender", + "description": "Recommend and customize Megatron Bridge recipes for a user's model, GPU count, and training goal. Indexes library recipes (pretrain/SFT/PEFT) and performance recipes.", + "metadata": { + "product.primary": "NeMo Megatron Bridge", + "classification.category.primary": "ai_and_machine_learning", + "catalog.subdomain": "training-ai", + "audience": "developer,ai_engineer,ml_engineer,hpc_developer", + "discovery.activity_tags": "select,configure,train,optimize,validate" + } + }, + { + "path": "skills/nemo-mbridge-mlm-bridge-training", + "name": "nemo-mbridge-mlm-bridge-training", + "description": "Run Megatron-LM (MLM) and Megatron Bridge training with mock or real data. Covers correlation testing, available recipes, and multi-GPU examples.", + "metadata": { + "product.primary": "NeMo Megatron Bridge", + "classification.category.primary": "ai_and_machine_learning", + "catalog.subdomain": "training-ai", + "audience": "developer,ai_engineer,ml_engineer,hpc_developer", + "discovery.activity_tags": "train,configure,validate,debug" + } + }, + { + "path": "skills/nemo-mbridge-multi-node-slurm", + "name": "nemo-mbridge-multi-node-slurm", + "description": "Convert single-node scripts to multi-node Slurm sbatch jobs and debug common multi-node failures. Covers srun-native vs uv run torch.distributed approaches, container setup, NCCL timeouts, OOM sizing for MoE models, and interactive allocation.", + "metadata": { + "product.primary": "NeMo Megatron Bridge", + "classification.category.primary": "ai_and_machine_learning", + "catalog.subdomain": "training-ai", + "audience": "developer,ai_engineer,ml_engineer,hpc_developer,devops_engineer", + "discovery.activity_tags": "configure,orchestrate,scale,troubleshoot,debug" + } + }, + { + "path": "skills/nemo-mbridge-perf-activation-recompute", + "name": "nemo-mbridge-perf-activation-recompute", + "description": "Validate and use selective and full activation recompute in Megatron Bridge to reduce GPU memory usage at the cost of extra compute.", + "metadata": { + "product.primary": "NeMo Megatron Bridge", + "classification.category.primary": "ai_and_machine_learning", + "catalog.subdomain": "training-ai", + "audience": "developer,ai_engineer,ml_engineer,hpc_developer", + "discovery.activity_tags": "optimize,configure,validate,measure" + } + }, + { + "path": "skills/nemo-mbridge-perf-cpu-offloading", + "name": "nemo-mbridge-perf-cpu-offloading", + "description": "Validate and use CPU offloading in Megatron Bridge, including layer-level activation offloading and fractional optimizer state offloading with HybridDeviceOptimizer.", + "metadata": { + "product.primary": "NeMo Megatron Bridge", + "classification.category.primary": "ai_and_machine_learning", + "catalog.subdomain": "training-ai", + "audience": "developer,ai_engineer,ml_engineer,hpc_developer", + "discovery.activity_tags": "optimize,configure,validate,measure" + } + }, + { + "path": "skills/nemo-mbridge-perf-cuda-graphs", + "name": "nemo-mbridge-perf-cuda-graphs", + "description": "Validate and use CUDA graph capture in Megatron Bridge, including local full-iteration graphs and Transformer Engine scoped graphs for attention, MLP, and MoE modules.", + "metadata": { + "product.primary": "NeMo Megatron Bridge", + "classification.category.primary": "ai_and_machine_learning", + "catalog.subdomain": "training-ai", + "audience": "developer,ai_engineer,ml_engineer,hpc_developer", + "discovery.activity_tags": "optimize,configure,validate,benchmark" + } + }, + { + "path": "skills/nemo-mbridge-perf-expert-parallel-overlap", + "name": "nemo-mbridge-perf-expert-parallel-overlap", + "description": "Validate and use MoE expert-parallel communication overlap in Megatron-Bridge, including overlap_moe_expert_parallel_comm, delay_wgrad_compute, and flex dispatcher backends such as DeepEP and HybridEP.", + "metadata": { + "product.primary": "NeMo Megatron Bridge", + "classification.category.primary": "ai_and_machine_learning", + "catalog.subdomain": "training-ai", + "audience": "developer,ai_engineer,ml_engineer,hpc_developer", + "discovery.activity_tags": "optimize,configure,scale,validate" + } + }, + { + "path": "skills/nemo-mbridge-perf-hierarchical-context-parallel", + "name": "nemo-mbridge-perf-hierarchical-context-parallel", + "description": "Operational guide for enabling hierarchical context parallelism in Megatron-Bridge, including config knobs, code anchors, pitfalls, and verification.", + "metadata": { + "product.primary": "NeMo Megatron Bridge", + "classification.category.primary": "ai_and_machine_learning", + "catalog.subdomain": "training-ai", + "audience": "developer,ai_engineer,ml_engineer,hpc_developer", + "discovery.activity_tags": "configure,scale,optimize,validate" + } + }, + { + "path": "skills/nemo-mbridge-perf-megatron-fsdp", + "name": "nemo-mbridge-perf-megatron-fsdp", + "description": "Operational guide for enabling Megatron FSDP in Megatron-Bridge, including config knobs, code anchors, pitfalls, and verification.", + "metadata": { + "product.primary": "NeMo Megatron Bridge", + "classification.category.primary": "ai_and_machine_learning", + "catalog.subdomain": "training-ai", + "audience": "developer,ai_engineer,ml_engineer,hpc_developer", + "discovery.activity_tags": "configure,scale,optimize,validate" + } + }, + { + "path": "skills/nemo-mbridge-perf-memory-tuning", + "name": "nemo-mbridge-perf-memory-tuning", + "description": "Techniques for reducing peak GPU memory in Megatron Bridge \u2014 expandable segments, parallelism resizing, activation recompute, CPU offloading constraints, and common OOM fixes.", + "metadata": { + "product.primary": "NeMo Megatron Bridge", + "classification.category.primary": "ai_and_machine_learning", + "catalog.subdomain": "training-ai", + "audience": "developer,ai_engineer,ml_engineer,hpc_developer", + "discovery.activity_tags": "optimize,troubleshoot,debug,validate" + } + }, + { + "path": "skills/nemo-mbridge-perf-moe-comm-overlap", + "name": "nemo-mbridge-perf-moe-comm-overlap", + "description": "MoE expert-parallel communication overlap in Megatron Bridge. Covers dispatch/combine overlap, flex dispatcher backends, and expert wgrad scheduling.", + "metadata": { + "product.primary": "NeMo Megatron Bridge", + "classification.category.primary": "ai_and_machine_learning", + "catalog.subdomain": "training-ai", + "audience": "developer,ai_engineer,ml_engineer,hpc_developer", + "discovery.activity_tags": "optimize,configure,scale,validate" + } + }, + { + "path": "skills/nemo-mbridge-perf-moe-dispatcher-selection", + "name": "nemo-mbridge-perf-moe-dispatcher-selection", + "description": "Choose the right MoE token dispatcher (`alltoall`, DeepEP, or HybridEP) for the hardware, EP degree, and optimization stage. Summarizes patterns from DSV3, Qwen3, Qwen3-Next, and VLM bring-up work.", + "metadata": { + "product.primary": "NeMo Megatron Bridge", + "classification.category.primary": "ai_and_machine_learning", + "catalog.subdomain": "training-ai", + "audience": "developer,ai_engineer,ml_engineer,hpc_developer", + "discovery.activity_tags": "select,assess,optimize,configure" + } + }, + { + "path": "skills/nemo-mbridge-perf-moe-hardware-configs", + "name": "nemo-mbridge-perf-moe-hardware-configs", + "description": "Representative MoE training playbooks by hardware platform and model family. Summarizes rounded throughput bands, parallelism patterns, and common tuning stacks.", + "metadata": { + "product.primary": "NeMo Megatron Bridge", + "classification.category.primary": "ai_and_machine_learning", + "catalog.subdomain": "training-ai", + "audience": "developer,ai_engineer,ml_engineer,hpc_developer", + "discovery.activity_tags": "select,assess,optimize,configure" + } + }, + { + "path": "skills/nemo-mbridge-perf-moe-long-context", + "name": "nemo-mbridge-perf-moe-long-context", + "description": "Long-context MoE training guidance for Megatron Bridge. Covers CP sizing, selective recompute, dispatcher choices, and practical patterns from DSV3, Qwen3, and Qwen3-Next long-context experiments.", + "metadata": { + "product.primary": "NeMo Megatron Bridge", + "classification.category.primary": "ai_and_machine_learning", + "catalog.subdomain": "training-ai", + "audience": "developer,ai_engineer,ml_engineer,hpc_developer", + "discovery.activity_tags": "configure,optimize,scale,validate" + } + }, + { + "path": "skills/nemo-mbridge-perf-moe-optimization-workflow", + "name": "nemo-mbridge-perf-moe-optimization-workflow", + "description": "Systematic workflow for MoE training optimization in Megatron Bridge, based on the Megatron-Core MoE paper. Covers the Three Walls framework, parallel folding, recompute strategy, dispatcher choice, and CUDA-graph bring-up.", + "metadata": { + "product.primary": "NeMo Megatron Bridge", + "classification.category.primary": "ai_and_machine_learning", + "catalog.subdomain": "training-ai", + "audience": "developer,ai_engineer,ml_engineer,hpc_developer", + "discovery.activity_tags": "optimize,benchmark,measure,validate" + } + }, + { + "path": "skills/nemo-mbridge-perf-moe-vlm-training", + "name": "nemo-mbridge-perf-moe-vlm-training", + "description": "Practical guidance for training MoE VLMs in Megatron Bridge. Compares FSDP and 3D-parallel approaches, using rounded lessons from Qwen3-VL, Qwen3-Next, and other multimodal experiments.", + "metadata": { + "product.primary": "NeMo Megatron Bridge", + "classification.category.primary": "ai_and_machine_learning", + "catalog.subdomain": "training-ai", + "audience": "developer,ai_engineer,ml_engineer,hpc_developer", + "discovery.activity_tags": "train,configure,optimize,validate" + } + }, + { + "path": "skills/nemo-mbridge-perf-parallelism-strategies", + "name": "nemo-mbridge-perf-parallelism-strategies", + "description": "Operational guide for choosing and combining parallelism strategies in Megatron Bridge, including sizing rules, hardware topology mapping, and combined parallelism configuration.", + "metadata": { + "product.primary": "NeMo Megatron Bridge", + "classification.category.primary": "ai_and_machine_learning", + "catalog.subdomain": "training-ai", + "audience": "developer,ai_engineer,ml_engineer,hpc_developer", + "discovery.activity_tags": "select,configure,scale,optimize" + } + }, + { + "path": "skills/nemo-mbridge-perf-sequence-packing", + "name": "nemo-mbridge-perf-sequence-packing", + "description": "Validate and use packed sequences and long-context training in Megatron-Bridge, distinguishing offline packed SFT for LLMs from in-batch packing for VLMs, and applying the right CP constraints.", + "metadata": { + "product.primary": "NeMo Megatron Bridge", + "classification.category.primary": "ai_and_machine_learning", + "catalog.subdomain": "training-ai", + "audience": "developer,ai_engineer,ml_engineer,hpc_developer", + "discovery.activity_tags": "configure,optimize,validate,train" + } + }, + { + "path": "skills/nemo-mbridge-perf-tp-dp-comm-overlap", + "name": "nemo-mbridge-perf-tp-dp-comm-overlap", + "description": "Operational guide for enabling TP, DP, and PP communication overlap in Megatron-Bridge, including config knobs, code anchors, pitfalls, and verification.", + "metadata": { + "product.primary": "NeMo Megatron Bridge", + "classification.category.primary": "ai_and_machine_learning", + "catalog.subdomain": "training-ai", + "audience": "developer,ai_engineer,ml_engineer,hpc_developer", + "discovery.activity_tags": "configure,scale,optimize,validate" + } + }, + { + "path": "skills/nemo-mbridge-resiliency", + "name": "nemo-mbridge-resiliency", + "description": "Resiliency features in Megatron Bridge including fault tolerance, straggler detection, in-process restart, preemption, and re-run state machine.", + "metadata": { + "product.primary": "NeMo Megatron Bridge", + "classification.category.primary": "ai_and_machine_learning", + "catalog.subdomain": "training-ai", + "audience": "developer,ai_engineer,ml_engineer,hpc_developer,devops_engineer", + "discovery.activity_tags": "configure,troubleshoot,recover,validate" + } + }, + { + "path": "skills/cupynumeric-install", + "name": "cupynumeric-install", + "description": "Install and verify cuPyNumeric for Python \u2014 requirements, commands, verification. Source builds are out of scope.", + "metadata": { + "product.primary": "cuPyNumeric", + "classification.category.primary": "accelerated_computing", + "catalog.subdomain": "data-science", + "audience": "developer,data_scientist,hpc_developer", + "discovery.activity_tags": "get_started,configure,validate" + } + }, + { + "path": "skills/dali-dynamic-mode", + "name": "dali-dynamic-mode", + "description": "DALI imperative dynamic mode (`nvidia.dali.experimental.dynamic`, ndd): use when working on ndd code or migrating pipelines; skip pipeline-only tasks.", + "metadata": { + "product.primary": "DALI", + "classification.category.primary": "ai_and_machine_learning", + "catalog.subdomain": "training-ai", + "audience": "developer,ai_engineer,ml_engineer", + "discovery.activity_tags": "configure,migrate,debug,validate" + } + }, + { + "path": "skills/dynamo-interconnect-check", + "name": "dynamo-interconnect-check", + "description": "Validate that a Dynamo deployment's NIXL/UCX/NCCL interconnect is ready for disaggregated serving over RDMA/NVLink. Use after recipe-runner brings a deployment up (especially disagg/multi-node) to confirm the KV transport is correct; use troubleshoot for diagnosing already-failed pods.", + "metadata": { + "product.primary": "Dynamo", + "classification.category.primary": "ai_and_machine_learning", + "catalog.subdomain": "inference-ai", + "audience": "developer,devops_engineer,platform_engineer,ai_engineer", + "discovery.activity_tags": "validate,monitor,troubleshoot,debug" + } + }, + { + "path": "skills/dynamo-recipe-runner", + "name": "dynamo-recipe-runner", + "description": "Select, validate, patch, and deploy existing NVIDIA Dynamo Kubernetes recipes. Use for model/backend/GPU/deployment-mode recipe bring-up; use router-starter for router-only mode work and troubleshoot for broken deployments.", + "metadata": { + "product.primary": "Dynamo", + "classification.category.primary": "ai_and_machine_learning", + "catalog.subdomain": "inference-ai", + "audience": "developer,devops_engineer,platform_engineer,ai_engineer", + "discovery.activity_tags": "deploy,configure,orchestrate,scale,validate" + } + }, + { + "path": "skills/dynamo-router-starter", + "name": "dynamo-router-starter", + "description": "Start or patch Dynamo router modes and run router endpoint smoke checks. Use for round-robin, KV-aware, least-loaded, or device-aware routing setup; use recipe-runner for recipe deployment and troubleshoot for failure diagnosis.", + "metadata": { + "product.primary": "Dynamo", + "classification.category.primary": "ai_and_machine_learning", + "catalog.subdomain": "inference-ai", + "audience": "developer,devops_engineer,platform_engineer,ai_engineer", + "discovery.activity_tags": "configure,deploy,integrate,validate,troubleshoot" + } + }, + { + "path": "skills/dynamo-troubleshoot", + "name": "dynamo-troubleshoot", + "description": "Diagnose failed or unhealthy Dynamo deployments. Use when pods, model-cache jobs, PVCs, workers, frontend/router health, endpoints, or benchmark jobs fail; use recipe-runner/router-starter before this for normal bring-up.", + "metadata": { + "product.primary": "Dynamo", + "classification.category.primary": "ai_and_machine_learning", + "catalog.subdomain": "inference-ai", + "audience": "developer,devops_engineer,platform_engineer,ai_engineer", + "discovery.activity_tags": "troubleshoot,debug,inspect,monitor,recover" + } + }, + { + "path": "skills/nemotron-customize", + "name": "nemotron-customize", + "description": "Plan Nemotron customization pipelines from repo steps: SFT, PEFT/LoRA, AutoModel vs Megatron-Bridge, DPO/RLVR/GRPO/RLHF, curate-then-translate, BYOB/MCQ benchmark prep or translation, checkpoint conversion, ModelOpt optimization, and endpoint or checkpoint evaluation.", + "metadata": { + "product.primary": "Nemotron", + "classification.category.primary": "ai_and_machine_learning", + "catalog.subdomain": "training-ai", + "audience": "developer,ai_engineer,ml_engineer,data_scientist", + "discovery.activity_tags": "fine_tune,train,evaluate,configure,optimize" + } + }, + { + "path": "skills/rag-blueprint", + "name": "rag-blueprint", + "description": "NVIDIA RAG Blueprint \u2014 deploy, configure, troubleshoot, and manage. Handles any RAG action: deploy, install, start, enable, disable, toggle, change, configure, troubleshoot, debug, fix, shutdown, stop, or tear down any RAG feature or service (Agentic RAG, VLM, guardrails, query rewriting, models, search, ingestion, observability, summarization, reasoning, and more).", + "metadata": { + "product.primary": "RAG", + "classification.category.primary": "ai_and_machine_learning", + "catalog.subdomain": "agentic-ai", + "audience": "developer,application_developer,ai_engineer,devops_engineer,platform_engineer", + "discovery.activity_tags": "deploy,configure,troubleshoot,debug,operate" + } + }, + { + "path": "skills/rag-eval", + "name": "rag-eval", + "description": "Filesystem RAG benchmarks: corpus/, train.json, evaluate_rag.py (RAGAS quality). Not for prod monitoring, latency/throughput benchmarking (use rag-perf), or evals outside this repo layout.", + "metadata": { + "product.primary": "RAG", + "classification.category.primary": "ai_and_machine_learning", + "catalog.subdomain": "agentic-ai", + "audience": "developer,ai_engineer,ml_engineer,data_scientist", + "discovery.activity_tags": "evaluate,benchmark,validate,measure" + } + }, + { + "path": "skills/rag-perf", + "name": "rag-perf", + "description": "Performance benchmarking for a deployed NVIDIA RAG Blueprint server: profiling pass + aiperf load test driven by a single YAML config. Not for accuracy / RAGAS scoring (use rag-eval) or for deploying / repairing services (use rag-blueprint).", + "metadata": { + "product.primary": "RAG", + "classification.category.primary": "ai_and_machine_learning", + "catalog.subdomain": "agentic-ai", + "audience": "developer,ai_engineer,ml_engineer,devops_engineer,platform_engineer", + "discovery.activity_tags": "benchmark,profile,measure,optimize,validate" + } + }, + { + "path": "skills/tilegym-adding-cutile-kernel", + "name": "tilegym-adding-cutile-kernel", + "description": "Add a new cuTile GPU kernel operator to TileGym. Covers dispatch registration in ops.py, cuTile backend implementation, __init__.py exports, test creation, and benchmark in tests/benchmark. Use when adding, creating, or implementing a new cuTile operator/kernel in TileGym, or when asking how to register a new cuTile op.", + "metadata": { + "product.primary": "CUDA Tile", + "classification.category.primary": "accelerated_computing", + "catalog.subdomain": "gpu-development", + "audience": "developer,hpc_developer", + "discovery.activity_tags": "extend,integrate,test,benchmark,validate" + } + }, + { + "path": "skills/digital-health-clinical-asr-build", + "name": "digital-health-clinical-asr-build", + "description": "Stage 2 of the Clinical ASR Flywheel. Use when curating clinical terms, tagging IPA, and synthesizing a NeMo manifest. NOT for scoring (use /digital-health-clinical-asr-eval).", + "metadata": { + "product.primary": "Nemotron for Digital Health", + "classification.category.primary": "ai_and_machine_learning", + "catalog.subdomain": "conversational-ai", + "audience": "developer,ai_engineer,ml_engineer,data_scientist", + "discovery.activity_tags": "generate,synthesize,configure,validate" + } + }, + { + "path": "skills/digital-health-clinical-asr-eval", + "name": "digital-health-clinical-asr-eval", + "description": "Stage 3 of Clinical ASR Flywheel. Score a NeMo manifest, produce the five-section KER leaderboard (by-ipa_source diagnostic). Not for ASR auth (/riva-asr).", + "metadata": { + "product.primary": "Nemotron for Digital Health", + "classification.category.primary": "ai_and_machine_learning", + "catalog.subdomain": "conversational-ai", + "audience": "developer,ai_engineer,ml_engineer,data_scientist", + "discovery.activity_tags": "evaluate,measure,validate,compare" + } + }, + { + "path": "skills/digital-health-clinical-asr-finetune", + "name": "digital-health-clinical-asr-finetune", + "description": "Stage 4 of the Clinical ASR Flywheel. Use when priority KER is above 0.3 to run stock NeMo SFT on Parakeet TDT v2 and offline cycle N+1 re-eval. NOT for generic word boosting (use /finetune-asr).", + "metadata": { + "product.primary": "Nemotron for Digital Health", + "classification.category.primary": "ai_and_machine_learning", + "catalog.subdomain": "conversational-ai", + "audience": "developer,ai_engineer,ml_engineer,data_scientist", + "discovery.activity_tags": "fine_tune,train,evaluate,validate" + } + }, + { + "path": "skills/digital-health-clinical-asr-setup", + "name": "digital-health-clinical-asr-setup", + "description": "Stage 1 of Clinical ASR Flywheel. Use when bootstrapping a cycle: NVCF+MW disclosure, NVIDIA_API_KEY check, deps install, TTS+ASR smoke test.", + "metadata": { + "product.primary": "Nemotron for Digital Health", + "classification.category.primary": "ai_and_machine_learning", + "catalog.subdomain": "conversational-ai", + "audience": "developer,ai_engineer,ml_engineer", + "discovery.activity_tags": "get_started,configure,validate,troubleshoot" + } + }, + { + "path": "skills/earth2studio-data-fetch", + "name": "earth2studio-data-fetch", + "description": "Fetch weather/climate data via Earth2Studio data sources for specific variables and times. Do NOT use for inference pipelines, model discovery, or installation.", + "metadata": { + "product.primary": "Earth2Studio", + "classification.category.primary": "ai_and_machine_learning", + "catalog.subdomain": "simulation-modeling", + "audience": "developer,data_scientist,research_academic", + "discovery.activity_tags": "extract,transform,configure,validate" + } + }, + { + "path": "skills/earth2studio-deterministic-forecast", + "name": "earth2studio-deterministic-forecast", + "description": "Build deterministic forecast scripts with Earth2Studio (model, data source, IO, inference). Do NOT use for ensemble, diagnostics, data-only fetch, or install.", + "metadata": { + "product.primary": "Earth2Studio", + "classification.category.primary": "ai_and_machine_learning", + "catalog.subdomain": "simulation-modeling", + "audience": "developer,data_scientist,research_academic", + "discovery.activity_tags": "inference,configure,integrate,validate" + } + }, + { + "path": "skills/earth2studio-discover", + "name": "earth2studio-discover", + "description": "Find Earth2Studio models, data sources, and examples for a weather/climate use case. Do NOT use for writing inference code, downloading data, or installation.", + "metadata": { + "product.primary": "Earth2Studio", + "classification.category.primary": "ai_and_machine_learning", + "catalog.subdomain": "simulation-modeling", + "audience": "developer,data_scientist,research_academic", + "discovery.activity_tags": "select,assess,compare,validate" + } + }, + { + "path": "skills/earth2studio-install", + "name": "earth2studio-install", + "description": "Guide installing Earth2Studio via uv or pip, selecting model extras, and configuring the environment. Do NOT use for writing inference code, choosing models, or PhysicsNeMo questions.", + "metadata": { + "product.primary": "Earth2Studio", + "classification.category.primary": "ai_and_machine_learning", + "catalog.subdomain": "simulation-modeling", + "audience": "developer,data_scientist,research_academic", + "discovery.activity_tags": "get_started,configure,validate" + } + } + ] +} diff --git a/.github/scripts/marketplace/skills.sh.json b/.github/scripts/marketplace/skills.sh.json new file mode 100644 index 00000000..e1e545b7 --- /dev/null +++ b/.github/scripts/marketplace/skills.sh.json @@ -0,0 +1,130 @@ +{ + "$schema": "https://skills.sh/schemas/skills.sh.schema.json", + "notGrouped": "bottom", + "groupings": [ + { + "title": "Agentic AI", + "description": "Build and operate agentic systems, including RAG workflows, evaluation harnesses, tool use, policy, sandboxing, agent configuration, and agent workflow automation.", + "skills": [ + "aiq-deploy", + "aiq-research", + "nemoclaw-user-agent-skills", + "nemoclaw-user-configure-inference", + "nemoclaw-user-configure-security", + "nemoclaw-user-deploy-remote", + "nemoclaw-user-get-started", + "nemoclaw-user-manage-policy", + "nemoclaw-user-manage-sandboxes", + "nemoclaw-user-monitor-sandbox", + "nemoclaw-user-overview", + "nemoclaw-user-reference", + "skill-card-generator", + "rag-blueprint", + "rag-eval", + "rag-perf" + ] + }, + { + "title": "Conversational AI", + "description": "Build and deploy speech, voice, dialogue, and multimodal conversational AI workflows, including real-time assistant and voice-agent experiences.", + "skills": [ + "digital-health-clinical-asr-build", + "digital-health-clinical-asr-eval", + "digital-health-clinical-asr-finetune", + "digital-health-clinical-asr-setup" + ] + }, + { + "title": "Data Science", + "description": "Prepare, analyze, explore, and model data for AI and accelerated analytics workflows that are not specific to one training, inference, or agent task.", + "skills": [ + "cupynumeric-install" + ] + }, + { + "title": "Decision Optimization", + "description": "Formulate and solve routing, scheduling, and numerical optimization problems using optimization solvers, APIs, services, and deployment workflows.", + "skills": [ + "cuopt-numerical-optimization-api-cli", + "cuopt-routing-api-python", + "cuopt-user-rules", + "cuopt-developer", + "cuopt-install", + "cuopt-numerical-optimization-api-python", + "cuopt-server-api-python", + "cuopt-numerical-optimization-formulation", + "cuopt-routing-formulation", + "cuopt-skill-evolution", + "cuopt-numerical-optimization-api-c", + "cuopt-server-common" + ] + }, + { + "title": "GPU Development", + "description": "Develop, tune, profile, and integrate GPU kernels and accelerated code, including CUDA-adjacent workflows, autotuning, low-level performance work, and framework integration.", + "skills": [ + "tilegym-adding-cutile-kernel" + ] + }, + { + "title": "Inference AI", + "description": "Deploy, serve, optimize, and evaluate models for production inference, including quantization, runtime performance, and serving configuration.", + "skills": [ + "dynamo-interconnect-check", + "dynamo-recipe-runner", + "dynamo-router-starter", + "dynamo-troubleshoot" + ] + }, + { + "title": "Simulation and Modeling", + "description": "Create, run, validate, and analyze simulation and modeling workflows for systems, environments, processes, and physical phenomena.", + "skills": [ + "earth2studio-data-fetch", + "earth2studio-deterministic-forecast", + "earth2studio-discover", + "earth2studio-install" + ] + }, + { + "title": "Training AI", + "description": "Build, validate, tune, and maintain large-scale model training workflows, including distributed training, model support, training validation, performance tuning, and training-stack CI.", + "skills": [ + "nemo-automodel-distributed-training", + "nemo-automodel-launcher-config", + "nemo-automodel-model-onboarding", + "nemo-automodel-recipe-development", + "nemo-mbridge-recipe-recommender", + "nemo-mbridge-mlm-bridge-training", + "nemo-mbridge-multi-node-slurm", + "nemo-mbridge-perf-activation-recompute", + "nemo-mbridge-perf-cpu-offloading", + "nemo-mbridge-perf-cuda-graphs", + "nemo-mbridge-perf-expert-parallel-overlap", + "nemo-mbridge-perf-hierarchical-context-parallel", + "nemo-mbridge-perf-megatron-fsdp", + "nemo-mbridge-perf-memory-tuning", + "nemo-mbridge-perf-moe-comm-overlap", + "nemo-mbridge-perf-moe-dispatcher-selection", + "nemo-mbridge-perf-moe-hardware-configs", + "nemo-mbridge-perf-moe-long-context", + "nemo-mbridge-perf-moe-optimization-workflow", + "nemo-mbridge-perf-moe-vlm-training", + "nemo-mbridge-perf-parallelism-strategies", + "nemo-mbridge-perf-sequence-packing", + "nemo-mbridge-perf-tp-dp-comm-overlap", + "nemo-mbridge-resiliency", + "dali-dynamic-mode", + "nemotron-customize" + ] + }, + { + "title": "Vision AI", + "description": "Build vision and video AI workflows for analytics, search, summarization, alerts, real-time understanding, and model integration.", + "skills": [ + "deepstream-dev", + "deepstream-import-vision-model" + ] + } + ] +}