nvidia-skill-finder
Use for NVIDIA-related requests where an NVIDIA skill might help, even if the user did not ask for a skill. Trigger on NVIDIA products, hardware, software,…
Use when Jetson codec, profile, chroma, bit-depth, dimension, engine-count, or operational support must be reconciled using live SDK APIs, authenticated NVIDIA samples, and NVIDIA documentation. Also use for Jetson questions about Netflix, Widevine, or other DRM-protected
$ npx -y skills add NVIDIA/skills --skill jetson-video-capability --agent claude-codeHow it fires
How this skill gets triggered: by you, by Claude, or both.
/jetson-video-capabilityContext preview
The summary Claude sees to decide when to auto-load this skill.
Use when Jetson codec, profile, chroma, bit-depth, dimension, engine-count, or operational support must be reconciled using live SDK APIs, authenticated NVIDIA samples, and NVIDIA documentation. Also use for Jetson questions about Netflix, Widevine, or other DRM-protected
name: jetson-video-capability license: "Apache-2.0" description: >- Use when Jetson codec, profile, chroma, bit-depth, dimension, engine-count, or operational support must be reconciled using live SDK APIs, authenticated NVIDIA samples, and NVIDIA documentation. Also use for Jetson questions about Netflix, Widevine, or other DRM-protected streaming-service playback to apply the codec-scope boundary. metadata: author: "Vinit Bansal <vinitkumarb@nvidia.com>" tags: [jetson, video-codec-sdk, pynvvideocodec, nvenc, nvdec, capability] languages: [python] data-classification: public
Answer Video Codec SDK and PyNvVideoCodec support questions without confusing an API response, a successful operation, and product documentation. Query the live target first, retain exact operation evidence separately, and publish the final product-support verdict from applicable NVIDIA documentation.
prerequisite. When it is supplied, authenticate and use it exactly. When it is absent, this skill authenticates only the selected installed surface: package-owned native sources and requested report targets, or the exact invoking PyNvVideoCodec interpreter and wheel. It never installs, repairs, scans for a venv, or imports setup code. A fresh environment artifact that the agent obtains from setup's public read-only probe counts as supplied evidence; it need not originate in the customer's prompt.
current availability from an x86 host or a result copied from another target.
route that surface to `jetson-video-setup`. If that skill is not installed, tell the user to install it. Never infer codec or product support from a missing Python import or native build prerequisite.
requested surface unless the user explicitly selects `both`.
setup evidence or the selected local prerequisites described below. An exact operation check also requires `jetson-video-recipe`, which owns recipe resolution, and `jetson-video-pipeline`, which owns authenticated encode-to-independent-decode execution.
[video content policy](../jetson-video-setup/references/video-content.md). Query-only work is media-free; only the documented setup fixture is allowed for the capability-smoke exception. Setup is not required solely for this policy: without it, require one exact user-selected path or URL for any other operation, never substitute catalog or synthetic media, and preserve source URL, license, attribution, path, size, and SHA-256.
Resolve this skill from the installed skills root and set `CAPABILITY_SKILL` to its canonical absolute path. Invoke each owning script directly under isolated Python:
python3 -I "$CAPABILITY_SKILL/scripts/query_native_sample_reports.py" --help python3 -I "$CAPABILITY_SKILL/scripts/query_encoder_caps.py" --help python3 -I "$CAPABILITY_SKILL/scripts/query_decoder_caps.py" --help python3 -I "$CAPABILITY_SKILL/scripts/validate_appenc_av1_ivf.py" --help
If an owning script is missing, stop with `dependency_required`. A setup artifact is optional; a supplied one is never optional to validate. Do not copy modules from another skill or add a fallback import path.
Capability queries and documentation reconciliation require no sibling when the selected SDK prerequisites already exist. Use `jetson-video-setup` for installation, repair, or one read-only readiness handoff when registered PyNvVideoCodec authority is required; use `jetson-video-recipe` plus `jetson-video-pipeline` for an exact requested operation, and `jetson-video-benchmark` for requested throughput. Check the agent's installed skill catalog before each such stage. If the sibling is present, read its `SKILL.md` and invoke its documented public entry point; pass artifacts as data and never import sibling code. If it is absent, preserve every completed query result and say, using the actual names: `I can run <stage>, but it requires <skill>, which is not installed. Install <skill> and retry this stage.` Never require a sibling for an unrequested or optional refinement.
1. **Classify the request first, before any target probe, capability query, or other workflow step.** For a request solely for objective quality metrics, including PSNR or SSIM, state only that this skill does not provide them and that a separately authorized quality workflow is required, then stop. Do not name or recommend an external tool, and do not offer to configure or run the comparison; do not request media, probe, install anything, or launch an operation. For a request solely about Netflix, Widevine, or other DRM-protected streaming-service playback, state only that this skill covers hardware encode/decode of user-supplied non-DRM bitstreams and does not cover, enable, or verify streaming-service or content-DRM playback. Do not claim whether the service will work; do not describe Jetson content-DRM certification; do not recommend or offer to install a browser, a content-DRM module (Widevine, PlayReady), a playback tool, a workaround, or a bypass; do not probe the target or launch an operation. NVDEC decode of supported user-supplied non-DRM bitstreams stays fully in scope for this skill and is never discouraged by this boundary, so you may say so. Then stop. A mixed request that also asks an in-scope codec question is not refused wholesale: answer the in-scope part normally and apply this boundary onl
Official, NVIDIA-verified Agent Skills for Claude Code, Codex, and other coding agents.
Use for NVIDIA-related requests where an NVIDIA skill might help, even if the user did not ask for a skill. Trigger on NVIDIA products, hardware, software,…
Official NVIDIA-authored guidance for NVIDIA cuDF GPU DataFrames, pandas acceleration, dask-cuDF, ETL, joins, groupby, CSV/Parquet I/O, nullable semantics, and…
Use when asked to install, deploy, run, validate, troubleshoot, or stop NVIDIA AI-Q Blueprint infrastructure.
Use when asked to run deep research or AI-Q research through a reachable NVIDIA AI-Q Blueprint backend.
Calibrate a new dataset from live RTSP camera streams via the AutoMagicCalib REST API. Use when the user provides RTSP URLs or asks to calibrate live cameras;…
Run end-to-end calibration on the shipped sample dataset (sdg_08_2_sample_data_010926.zip) against a running AMC microservice. Use when user says 'test sample…