LLM Torch Profiler Analysis
sgl-project/sglang
Unified LLM torch-profiler triage skill for sglang, vllm, TensorRT-LLM, and TokenSpeed.
Choose or explain HSB Sensor RX packetizer fields for HOLOLINKdef.svh.
$ npx skills add NVIDIA/skills --skill hsb-ip-packetizer -a claude-codeProject install by default; add -g for ~/.claude/skills/.
$ gh skill install NVIDIA/skills hsb-ip-packetizer --agent claude-codeProject scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).
$ git clone --depth 1 https://github.com/NVIDIA/skills.git skills-src && mkdir -p .claude/skills && cp -r skills-src/skills/hsb-ip-packetizer .claude/skills/hsb-ip-packetizer && rm -rf skills-srcUse ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.
Claude Code skills documentation · loads skills from .claude/skills/
Install the "hsb-ip-packetizer" agent skill from https://github.com/NVIDIA/skills/tree/main/skills/hsb-ip-packetizer into .claude/skills/hsb-ip-packetizer/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "hsb-ip-packetizer", then confirm the skill loads.Claude Code copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$skill-installer install https://github.com/NVIDIA/skills/tree/main/skills/hsb-ip-packetizerType this inside Codex. $skill-installer <name> installs a curated skill from openai/skills. The installer writes to $CODEX_HOME/skills (default ~/.codex/skills). Restart Codex if the skill does not show up.
$ npx skills add NVIDIA/skills --skill hsb-ip-packetizer -a codexProject install goes to .agents/skills/; add -g for ~/.codex/skills/.
$ gh skill install NVIDIA/skills hsb-ip-packetizer --agent codexProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/NVIDIA/skills.git skills-src && mkdir -p .agents/skills && cp -r skills-src/skills/hsb-ip-packetizer .agents/skills/hsb-ip-packetizer && rm -rf skills-srcUse ~/.agents/skills/ instead of .agents/skills for a personal install.
Codex skills documentation · loads skills from .agents/skills/
Install the "hsb-ip-packetizer" agent skill from https://github.com/NVIDIA/skills/tree/main/skills/hsb-ip-packetizer into .agents/skills/hsb-ip-packetizer/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "hsb-ip-packetizer", then confirm the skill loads.Codex copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ npx skills add NVIDIA/skills --skill hsb-ip-packetizer -a cursorProject install goes to .agents/skills/; add -g for ~/.cursor/skills/.
$ gh skill install NVIDIA/skills hsb-ip-packetizer --agent cursorProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/NVIDIA/skills.git skills-src && mkdir -p .cursor/skills && cp -r skills-src/skills/hsb-ip-packetizer .cursor/skills/hsb-ip-packetizer && rm -rf skills-srcUse ~/.cursor/skills/ instead of .cursor/skills for a personal install.
Cursor skills documentation · loads skills from .cursor/skills/, .agents/skills/, .claude/skills/, .codex/skills/
Install the "hsb-ip-packetizer" agent skill from https://github.com/NVIDIA/skills/tree/main/skills/hsb-ip-packetizer into .cursor/skills/hsb-ip-packetizer/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "hsb-ip-packetizer", then confirm the skill loads.Cursor copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ gemini skills install https://github.com/NVIDIA/skills.git --path skills/hsb-ip-packetizer--scope user (default) or --scope workspace; --path is the subfolder of the repo that holds the skill; --consent skips the security confirmation prompt.
$ npx skills add NVIDIA/skills --skill hsb-ip-packetizer -a gemini-cliProject install goes to .agents/skills/; add -g for ~/.gemini/skills/.
$ gh skill install NVIDIA/skills hsb-ip-packetizer --agent gemini-cliProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/NVIDIA/skills.git skills-src && mkdir -p .gemini/skills && cp -r skills-src/skills/hsb-ip-packetizer .gemini/skills/hsb-ip-packetizer && rm -rf skills-srcUse ~/.gemini/skills/ instead of .gemini/skills for a personal install, then run /skills reload.
Gemini CLI skills documentation · loads skills from .gemini/skills/, .agents/skills/
Install the "hsb-ip-packetizer" agent skill from https://github.com/NVIDIA/skills/tree/main/skills/hsb-ip-packetizer into .gemini/skills/hsb-ip-packetizer/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "hsb-ip-packetizer", then confirm the skill loads.Gemini CLI copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ gh skill install NVIDIA/skills hsb-ip-packetizerInstalls for Copilot at project scope by default; add --scope user for a personal install. Preview a skill first with gh skill preview. Needs GitHub CLI 2.90.0 or later (public preview).
$ npx skills add NVIDIA/skills --skill hsb-ip-packetizer -a github-copilotProject install goes to .agents/skills/; add -g for ~/.copilot/skills/.
$ git clone --depth 1 https://github.com/NVIDIA/skills.git skills-src && mkdir -p .github/skills && cp -r skills-src/skills/hsb-ip-packetizer .github/skills/hsb-ip-packetizer && rm -rf skills-srcUse ~/.copilot/skills/ instead of .github/skills for a personal install. Commit .github/skills so cloud agent and code review can use it.
GitHub Copilot skills documentation · loads skills from .github/skills/, .claude/skills/, .agents/skills/
Install the "hsb-ip-packetizer" agent skill from https://github.com/NVIDIA/skills/tree/main/skills/hsb-ip-packetizer into .github/skills/hsb-ip-packetizer/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "hsb-ip-packetizer", then confirm the skill loads.GitHub Copilot copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ npx skills add NVIDIA/skills --skill hsb-ip-packetizer -a opencodeOpenCode documents no install command of its own. Project install goes to .agents/skills/; add -g for ~/.config/opencode/skills/.
$ gh skill install NVIDIA/skills hsb-ip-packetizer --agent opencodeProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/NVIDIA/skills.git skills-src && mkdir -p .opencode/skills && cp -r skills-src/skills/hsb-ip-packetizer .opencode/skills/hsb-ip-packetizer && rm -rf skills-srcUse ~/.config/opencode/skills/ instead of .opencode/skills for a personal install.
OpenCode skills documentation · loads skills from .opencode/skills/, .claude/skills/, .agents/skills/
Install the "hsb-ip-packetizer" agent skill from https://github.com/NVIDIA/skills/tree/main/skills/hsb-ip-packetizer into .opencode/skills/hsb-ip-packetizer/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "hsb-ip-packetizer", then confirm the skill loads.OpenCode copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
hsb-ip-packetizerChoose or explain HSB Sensor RX packetizer fields for HOLOLINKdef.svh.
Hsb Ip Packetizer is an agent skill from NVIDIA/skills, published by the product's own GitHub organization. Choose or explain HSB Sensor RX packetizer fields for HOLOLINKdef.svh. Do not use for full defs, validation, or runtime APB programming.
Its SKILL.md is about 3.6k tokens, which your agent loads only when the skill is triggered. The skill folder holds 9 other files, including reference files (for example `BENCHMARK.md`, `evals/evals.json` and `references/handoff-contract.md`). Compatibility notes: Targets HSB IP rev 16'h2604; backward-compatible with 16'h2603. Prefer live HSB IP packetizer source when available; warn on unknown revisions. Designed to…
It works with NVIDIA AI Platform. The repository describes itself as: Agent Skills for NVIDIA products — install into Claude Code, Codex, and other coding agents to run Physical AI, robotics, simulation, CUDA, and RAG workflows end to end. The licence is Apache-2.0.
4 steps, taken from the first numbered list in SKILL.md.
Read from SKILL.md and the folder at commit 0e0d506. It shows what the files ask for, not the result of running them.
Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.
From allowed-tools in the SKILL.md frontmatter.
No scripts in the folder and no shell commands in SKILL.md (its code samples are yaml).
From the folder's file list and the shell code blocks in SKILL.md.
No URLs in SKILL.md.
From URLs in SKILL.md, links to its own repository left out.
Names no API keys, tokens, secrets or passwords.
From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.
Targets HSB IP rev 16'h2604; backward-compatible with 16'h2603. Prefer live HSB IP packetizer source when available; warn on unknown revisions. Designed to work standalone or as a companion to hsb-ip-def.
From compatibility in the SKILL.md frontmatter.
Hsb Ip Packetizer loads about 3.6k tokens when it runs, and up to ~7.1k if it reads all its reference files. Until then it costs about 39 tokens; SKILL.md has 1,708 words of instructions outside code blocks.
Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.
The automated check found no risky patterns in SKILL.md.
Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.
The full file from NVIDIA/skills at commit 0e0d506, republished under its Apache-2.0 licence (© NVIDIA). 1,708 words, ~3,585 tokens.
.claude/skills/hsb-ip-packetizer/SKILL.md (or your agent's skills folder). This skill also uses 7 other files; get the full folder from GitHub.Choose packetizer-related HOLOLINK_def.svh fields for the NVIDIA Holoscan Sensor Bridge Sensor RX path.
This skill owns only the packetizer slice of HOLOLINK_def.svh:
SIF_RX_PACKETIZER_EN[]SIF_RX_VP_COUNT[]SIF_RX_SORT_RESOLUTION[]SIF_RX_VP_SIZE[]SIF_RX_NUM_CYCLES[]Use hsb-ip-def for whole-file generation, validation, comparison, and non-packetizer macros.
SENSOR_RX_IF_INST must be defined and nonzero before packetizer fields are relevant.SIF_RX_WIDTH[] values and the user's data manipulation intent are required before deriving enabled packetizer values.16'h2604 and compatible rev 16'h2603.hsb-ip-def.SENSOR_RX_IF_INST is defined. If Sensor RX is disabled, say no packetizer fields are needed.references/handoff-contract.md and make packetizer_profile_overlay the first non-empty output after any required clarification.HOLOLINK_def.svh enables/sizes hardware; runtime software still controls pattern RAM, virtual-port selection, sort controls, bypass, padding, duplication, and latency.hsb-ip-def, use facts already known by that def flow. If the prompt gives RX count, enabled/pass-through intent, lane or stream width, virtual stream count, and cycle window, emit the overlay without re-asking for macro names or whole-file details. Leave final merge and full validation to hsb-ip-def.Prefer live HSB IP source over bundled packetizer references whenever the user's workspace provides it. The bundled references describe known rev 16'h2604 and backward-compatible rev 16'h2603; live source is the authority for the checked-out IP.
When source is available:
HOLOLINK_top.sv plus the packetizer RTL files:<hsb-ip-root>/top/HOLOLINK_top.sv<hsb-ip-root>/packetizer/packetizer_top.sv<hsb-ip-root>/packetizer/packetizer.sv<hsb-ip-root>/packetizer/virtual_port.sv<hsb-ip-root>/packetizer/odd_even_gen.svHOLOLINK_REV and HOLOLINK_BACKWARD_COMPAT_REV from HOLOLINK_top.sv when present.HOLOLINK_top still maps the five def fields into packetizer_top as documented before making source-sensitive claims.When no source is available, use the bundled references as known-revision guidance and say that the answer is based on the skill's supported HSB IP rev.
Load only the reference needed for the user's task:
| File | When to load |
|---|---|
references/packetizer-def-fields.md | Choosing or explaining HOLOLINK_def.svh packetizer fields, array length rules, bypass placeholders, RTL-derived constraints, or standalone SVH output formatting |
references/packetizer-architecture.md | Explaining packetizer blocks, data flow, clocks, APB registers, runtime-vs-def split, or why a field exists |
references/handoff-contract.md | Producing packetizer_profile_overlay YAML for hsb-ip-def or enforcing output-mode separation |
HOLOLINK_def.svh generation flow, use Def-Skill Handoff.HOLOLINK_def.svh, use Standalone SVH.references/packetizer-architecture.md for the register context.Collect these facts in order, skipping facts the user already supplied. Stop asking as soon as the data manipulation description is specific enough to derive the profile.
SENSOR_RX_IF_INST: number of Sensor RX interfaces. If undefined or zero, stop.SIF_RX_WIDTH[]: per-RX-interface input width. This is DIN_WIDTH for each packetizer instance.Skip per-macro confirmation prompts such as "Confirm SIF_RX_VP_COUNT?", "Confirm SIF_RX_SORT_RESOLUTION?", "Confirm SIF_RX_VP_SIZE?", or "Confirm SIF_RX_NUM_CYCLES?" Ask for missing data-shape facts, such as "How many virtual streams should RX0 split into?" or "What bit granularity is being rearranged?"
After the required facts are known, derive all packetizer arrays together:
SIF_RX_PACKETIZER_EN[i] = 1 for RX interfaces that need packetizer hardware for rearrangement, split, replication, or other runtime-controlled data manipulation; 0 for pass-through interfaces.SIF_RX_VP_COUNT[i] equals the number of virtual output streams/ports required by the user's manipulation for that RX interface.SIF_RX_VP_SIZE[i] is the per-virtual-port data size/width in bits. For an equal contiguous split of one SIF_RX_WIDTH[i] input word into N virtual ports, use SIF_RX_WIDTH[i] / N if it is a positive power-of-two divisor and matches the user's intent.SIF_RX_SORT_RESOLUTION[i] is the bit granularity of the data rearrangement operation. If the user describes fixed-size lanes, use that lane size when it passes the RTL-derived constraints. If no ordering manipulation is needed and the packetizer is used only for contiguous split/replication, use SIF_RX_WIDTH[i] to disable the sort network internally when legal.SIF_RX_NUM_CYCLES[i] is the number of input sensor-data cycles that participate in the manipulation. Use 1 for single-cycle rearrange/split/replication; use the stated cycle window when the manipulation spans multiple input cycles.If several profiles remain plausible after reading the data description, ask one clarifying question about the behavior or data shape that distinguishes them. Do not fall back to "typical" packetizer values.
For disabled entries in a mixed enabled/disabled design, fill the peer arrays with ignored placeholder values matching the RTL bypass constants:
SIF_RX_VP_COUNT[i] = 1SIF_RX_SORT_RESOLUTION[i] = 2SIF_RX_VP_SIZE[i] = 32SIF_RX_NUM_CYCLES[i] = 1SENSOR_RX_IF_INST entries.1, emit all four peer arrays at exactly SENSOR_RX_IF_INST length.0, emit only SIF_RX_PACKETIZER_EN[]; the four peer arrays are not required for the packetizer decision.'{default:<value>} only when every entry is identical and that form improves readability.SENSOR_RX_IF_INST, SIF_RX_WIDTH[], DATAPATH_WIDTH, host fields, clocks, or any non-packetizer macros from this skill.Use this when the user asks for packetizer lines to add to a HOLOLINK_def.svh.
Steps:
references/packetizer-def-fields.md for choosing values, checking constraints, and using the canonical standalone output formatting.systemverilog block containing only the packetizer localparams.hsb-ip-def or its validate_def.py.Use this when invoked by hsb-ip-def while that skill is generating a full HOLOLINK_def.svh.
Steps:
references/handoff-contract.md.packetizer_profile_overlay label once requirements are known. The def skill consumes the flat keys directly.Required label:
packetizer_profile_overlay
YAML keys:
sif_rx_packetizer_ensif_rx_vp_count only when any enable is 1sif_rx_sort_resolution only when any enable is 1sif_rx_vp_size only when any enable is 1sif_rx_num_cycles only when any enable is 1Example shape:
packetizer_profile_overlay
sif_rx_packetizer_en: [1, 0]
sif_rx_vp_count: [4, 1]
sif_rx_sort_resolution: [16, 2]
sif_rx_vp_size: [128, 32]
sif_rx_num_cycles: [1, 1]The def skill remains responsible for merging this overlay into its profile, generating the final SVH, and running full validation.
Use this when the user asks what a packetizer field does, why a value matters, or whether a packetizer combination is suitable.
Steps:
references/packetizer-def-fields.md for field-level answers.references/packetizer-architecture.md for architecture, data flow, clock domains, or runtime APB behavior.top/HOLOLINK_top.sv:1356-1362 or packetizer/packetizer.sv:69-85 when explaining RTL behavior.HOLOLINK_def.svh, state the packetizer-local concern and tell the user to validate the full file with hsb-ip-def.references/packetizer-architecture.md only for context.hsb-ip-def.Use hsb-ip-packetizer to choose packetizer fields for two Sensor RX interfaces. RX0 is 512 bits and must split each input word into four 128-bit virtual streams; RX1 is pass-through. Treat as Standalone SVH and emit only packetizer localparams: enable [1, 0], virtual-port count [4, 1], sort resolution [512, 2] for a contiguous split without lane reordering, virtual-port size [128, 32], and cycles [1, 1]. Tell the user to run full-file validation with hsb-ip-def.hsb-ip-def is generating my HOLOLINK_def.svh and needs packetizer values. RX0 should rearrange 16-bit lanes into four virtual streams; RX1 should stay pass-through. Treat as Def-Skill Handoff and put the YAML overlay first:packetizer_profile_overlay
sif_rx_packetizer_en: [1, 0]
sif_rx_vp_count: [4, 1]
sif_rx_sort_resolution: [16, 2]
sif_rx_vp_size: [16, 32]
sif_rx_num_cycles: [1, 1] Then state that hsb-ip-def remains responsible for merging this overlay, generating the full HOLOLINK_def.svh, and running full validation.
Do I need hsb-ip-packetizer if SENSOR_RX_IF_INST is disabled in my HOLOLINK_def.svh? Explain that packetizer fields apply only to Sensor RX interfaces, emit no packetizer arrays, and avoid runtime APB pattern-RAM configuration.HOLOLINK_def.svh files.HOLOLINK_def.svh files.© NVIDIA, Apache-2.0. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file
SKILL.md and 7 other files (references) in skills/hsb-ip-packetizer of NVIDIA/skills.
Open the folder on GitHubat commit 0e0d506
Hsb Ip Packetizer next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.
| Skill | Stars | Used in | Tokens | Auto-check | Licence | Repo updated |
|---|---|---|---|---|---|---|
| Hsb Ip Packetizer this skillNVIDIA/skills | 3.5k | — | ~3.6k | Automated safety check: Pass | Apache-2.0 | |
| LLM Torch Profiler Analysissgl-project/sglang | 37k | 2 repos | ~6.4k | Automated safety check: Pass | Apache-2.0 | |
| Skill InspectorNVIDIA/SkillSpector | 20k | — | ~1.8k | Automated safety check: Pass | Apache-2.0 | |
| Embeddings via 9Routerdecolua/9router | 30k | — | ~604 | Automated safety check: Pass | MIT | |
| NEAR AI Cloud Private Inferenceinternet-court/internet-court-skill | 6.4k | 2 repos | ~1.3k | Automated safety check: Pass | Custom licence | |
| Nemoclaw Maintainer Normalize Title TagsNVIDIA/NemoClaw | 23k | — | ~693 | Automated safety check: Pass | Apache-2.0 |
sgl-project/sglang
Unified LLM torch-profiler triage skill for sglang, vllm, TensorRT-LLM, and TokenSpeed.
NVIDIA/SkillSpector
Decides whether an agent skill is safe to install by combining a SkillSpector static scan with the agent's own source review, ending in APPROVE, CAUTION or REJECT.
decolua/9router
Generates vector embeddings through the 9Router /v1/embeddings endpoint, using models from providers such as OpenAI, Gemini, Mistral and Voyage for RAG and semantic search.
internet-court/internet-court-skill
Shows how to call NEAR AI Cloud through an OpenAI-compatible API and verify that inference ran in a TEE, using attestation checks and signed chat responses.
NVIDIA/NemoClaw
Remove bracketed NemoClaw tags from GitHub issue and PR titles.
NVIDIA/Megatron-LM
Walks an agent through working inside the Megatron-LM CI container and changing dependencies with uv, so lock files resolve the same locally and in CI.
NVIDIA/skills
A skill your agent uses when the user wants to deploy, run, debug, tear down, or call the REST API of the RTVI-CV 2D detection / tracking microservice.
NVIDIA/skills
Generates, validates, compares and explains HOLOLINK_def.svh macro files for the HSB IP, using bundled Python scripts and asking before it writes anything.
NVIDIA/skills
Runs and validates an end-to-end Mission Control demo in a locally installed Isaac Sim, with a Nova Carter robot driven through a Python server.
NVIDIA/skills
Orchestrates defect image generation for PCBA, metal surface and glass inspection with NVIDIA Cosmos AnomalyGen on OSMO, from cold-start Day 0 to real-photo Day 1 labeling.
NVIDIA/skills
Orchestrates video data augmentation and auto-labeling workflows on OSMO, from flow selection and preflight checks to submission, monitoring and output download.
NVIDIA/skills
Runs NVIDIA TAO Data Services KPI analysis on object detection results, comparing predictions to ground truth and writing per-class precision, recall and AP to a CSV.
Works with
Choose or explain HSB Sensor RX packetizer fields for HOLOLINKdef.svh. Hsb Ip Packetizer is an agent skill from NVIDIA/skills, published by the product's own GitHub organization.svh.
Hsb Ip Packetizer fits situations like: runtime APB programming.
Run `npx skills add NVIDIA/skills --skill hsb-ip-packetizer -a claude-code`. Or copy the skill folder (skills/hsb-ip-packetizer in NVIDIA/skills) into .claude/skills/hsb-ip-packetizer in your project. Claude Code loads it when a task matches its description.
Run `npx skills add NVIDIA/skills --skill hsb-ip-packetizer -a codex`. Or copy the skill folder (skills/hsb-ip-packetizer in NVIDIA/skills) into .agents/skills/hsb-ip-packetizer in your project. Codex loads it when a task matches its description.
Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add NVIDIA/skills --skill hsb-ip-packetizer -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/hsb-ip-packetizer, .gemini/skills/hsb-ip-packetizer, .github/skills/hsb-ip-packetizer and .opencode/skills/hsb-ip-packetizer in your project.
SKILL.md names no scripts, command-line tools or credentials: Hsb Ip Packetizer is instructions for the agent only. Compatibility (from SKILL.md): Targets HSB IP rev 16'h2604; backward-compatible with 16'h2603. Prefer live HSB IP packetizer source when available; warn on unknown revisions. Designed to work standalone or as a companion to hsb-ip-def..
SKILL.md contains no URLs. Any network use would come from the scripts or tools the agent runs. This is read from the text; nothing was executed.
Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. Review the folder before installing.
Hsb Ip Packetizer is published under the Apache-2.0 licence (declared in SKILL.md). It allows redistribution, so the full SKILL.md is shown on this page.
About 3.6k tokens (SKILL.md is roughly 14k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full. Its references folder adds about 3.5k tokens, read only when the agent opens those files.
Skills that share tags, products or a category with Hsb Ip Packetizer: LLM Torch Profiler Analysis (sgl-project/sglang, 37k stars), Skill Inspector (NVIDIA/SkillSpector, 20k stars), Embeddings via 9Router (decolua/9router, 30k stars) and NEAR AI Cloud Private Inference (internet-court/internet-court-skill, 6.4k stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.
NVIDIA (a GitHub organization, an official publisher) maintains it in NVIDIA/skills, which has 3,534 GitHub stars. The repository holds 380 skills in this directory. The repository was last updated on October 7, 2026.
Source: NVIDIA/skills on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.