CLIP vision-language model for image-text retrieval, zero-shot classification, embedding extraction, ONNX export, and TensorRT deployment. Use when fine-tuning or training CLIP, running zero-shot classification, computing image embeddings, or deploying CLIP to ONNX/TensorRT. This is a single-action model skill; do not use it for an iterative weak-attribute improvement loop that keeps retraining and evaluating until progress stops, which belongs to tao-run-deft-pas.

Iteratively optimize cuTile kernel performance through systematic profiling, bottleneck analysis, IR comparison, and targeted tuning. Covers tile sizes, occupancy, autotune configs, TMA, latency hints, persistent scheduling, num_ctas, flush_to_zero, and IR-level debugging. Use when asked to "optimize cutile kernel", "improve kernel perf", "tune cutile performance", "make kernel faster", or iteratively benchmark and refine a cuTile GPU kernel in the TileGym project.

NVIDIA RAG Blueprint — 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).

Use this skill when building or packaging reusable NeMo Relay runtime behavior as an embedded configuration component or a manifest-backed `rust_dynamic` native or `worker` gRPC plugin, with deterministic validation and rollback-safe registration.

Use this skill when first-time NeMo Relay users want to try Relay, choose the least-complex supported quick start, or verify initial value through the CLI, a maintained integration, or direct Python, Node.js, or Rust instrumentation before production setup.

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.

Operational guide for enabling hierarchical context parallelism in Megatron-Bridge, including config knobs, code anchors, pitfalls, and verification.

Use when you need to add, remove, edit, list, or change the boot default of an nvpmodel power mode on a Jetson/Tegra (Orin, Thor) target. Triggers: edit power mode, tune frequency caps.

Use when you need to add, remove, edit, list, or change the boot default of an nvfancontrol fan profile on a Jetson/Tegra (Orin, Thor) target. Triggers: edit fan profile, tune fan curve.

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.

Use this skill when the user needs to locate authoritative information about NVIDIA DOCA without access to the source tree — finding the right docs.nvidia.com page for a library/service/tool, identifying which DOCA libraries are installed and at what version, locating a sample on disk or its public GitHub source, decoding an on-disk path under /opt/mellanox/doca, or recovering from a 404'd or renamed doc URL. Trigger even when the user does not explicitly mention 'DOCA' or 'docs.nvidia.com' — typical implicit phrasings include 'where can I read about this library', 'which version do I have installed', 'where is the sample for X', 'this NVIDIA URL is broken what is the new one', 'what is in /opt/mellanox/doca', or 'where can I ask NVIDIA about this'. Refuse and route elsewhere for hands-on programming patterns, env prep and install verification, library API tutorials, or hardware/firmware mutation — those belong to doca-programming-guide, doca-setup, the per-library skills, and doca-hardware-safety.

Use this skill for hands-on DOCA Arg Parser CLI work on a shipped sample or new DOCA-using app — adding / removing / renaming flags; wiring `doca_argp_init` → register params → `doca_argp_start` → `doca_argp_destroy` in order; picking a parameter type from the full public enum (`DOCA_ARGP_TYPE_STRING`, `_INT`, `_BOOLEAN`, `_DEVICE`, `_DEVICE_REP`, `_DOUBLE` — six values, not three); preserving the standard `--device` / `--representor` / `--json` (`-j`; real flag is `--json`, NOT `--json-config`) / `--sdk-log-level` surface; or debugging `DOCA_ERROR_BAD_STATE` / `INVALID_VALUE` / `NOT_SUPPORTED` / `IO_FAILED` from `doca_argp_*`. Trigger on implicit phrasings: "add a custom flag to a DOCA sample", "should I use getopt here", "BAD_STATE registering a new param", "my JSON config key is rejected", or "my sample's --json is ignored". Refuse and route elsewhere for variadic-flag / subcommand / shell-completion features, DOCA Core context, or DOCA Log internals.

NVIDIA DeepStream SDK 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.

Use the `http` package to execute GET, POST, PUT, or DELETE requests. Use when you need to fetch from or send data to a REST API.

WPA/WPA2/WPA3-Enterprise (802.1X / EAP) attack methodology — EAP method identification (PEAP-MSCHAPv2, EAP-TTLS, EAP-TLS, EAP-GTC, EAP-PWD, EAP-FAST), evil-twin RADIUS attacks with eaphammer for credential capture, MSCHAPv2 challenge-response cracking, EAP-TLS client certificate theft paths (DPAPI, NDES, AD CS auto-enrollment), supplicant validation bypass (missing server cert validation, missing CN pinning, BYOD misconfigurations), and post-capture pivots into AD via cracked domain credentials. Use for corporate Wi-Fi engagements where the network is 802.1X authenticated.

WPA/WPA2-PSK attack methodology — four-way handshake capture via targeted deauthentication, PMKID attacks (no client required), hcxdumptool / hcxpcapngtool conversion to hashcat hc22000 format, GPU-accelerated cracking with dictionary, mask, and rule-based attacks, vendor default-PSK generators (UPC, Sky, BT, etc.), 802.11r FT key cracking, opportunistic key cache analysis, and signal-level optimization. Use when the in-scope network is WPA/WPA2 Personal — the most common consumer/SMB encryption mode.

Wi-Fi reconnaissance methodology — adapter selection, monitor mode and packet injection setup, regulatory domain handling, multi-band airspace mapping, hidden SSID discovery, BSSID/ESSID/channel/PMF/encryption fingerprinting, client probe analysis, vendor OUI lookup, war-driving with Kismet/airodump-ng/Wigle, and structured airspace data capture for downstream attacks. Use at the start of any wireless engagement to build the target map before active attacks; covers 2.4 GHz, 5 GHz, and 6 GHz (Wi-Fi 6E) bands and adapter compatibility for each.

Bluetooth Low Energy (BLE) attack methodology — GATT enumeration, characteristic read/write without auth, pairing downgrade (Just Works forced), LE Secure Connections bypass, MITM via active relay, sniffing with Sniffle (TI CC1352) / Ubertooth / Frontline, encryption key extraction (LE Legacy Pairing crackable, LE Secure Connections strong), proximity authentication abuse (cars, locks), and companion-app trust analysis. Use for IoT BLE devices, smart locks, fitness trackers, medical devices, BLE beacons, or any device pairing over BLE.

This skill helps an LLM build correct native Model Context Protocol integrations with @ax-llm/ax. Use when the user asks about AxMCPClient, MCP transports, tools, prompts, resources, subscriptions, tasks, sampling, elicitation, roots, authentication, OAuth, MCP Apps, recording/replay, or MCP integration with AxGen, AxAgent, AxFlow, chat, optimization, and AxEventRuntime.

Use when changing Ax website language docs, language-specific snippets, examples, API symbol mappings, generated package capabilities, or adding a new website language route. Keeps the markdown-only Hugo site source-audited and generated from repo truth.