Megatron-LM CI Failure Triage
NVIDIA/Megatron-LM
Investigates a failing GitHub Actions run or job for Megatron-LM, finds the root cause plus the PR and test author involved, and files a structured bug issue.
Discover and run Holoscan Sensor Bridge example applications on a connected devkit.
$ npx skills add NVIDIA/skills --skill hsb-app -a claude-codeProject install by default; add -g for ~/.claude/skills/.
$ gh skill install NVIDIA/skills hsb-app --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-app .claude/skills/hsb-app && 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-app" agent skill from https://github.com/NVIDIA/skills/tree/main/skills/hsb-app into .claude/skills/hsb-app/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "hsb-app", 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-appType 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-app -a codexProject install goes to .agents/skills/; add -g for ~/.codex/skills/.
$ gh skill install NVIDIA/skills hsb-app --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-app .agents/skills/hsb-app && 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-app" agent skill from https://github.com/NVIDIA/skills/tree/main/skills/hsb-app into .agents/skills/hsb-app/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "hsb-app", 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-app -a cursorProject install goes to .agents/skills/; add -g for ~/.cursor/skills/.
$ gh skill install NVIDIA/skills hsb-app --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-app .cursor/skills/hsb-app && 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-app" agent skill from https://github.com/NVIDIA/skills/tree/main/skills/hsb-app into .cursor/skills/hsb-app/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "hsb-app", 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-app--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-app -a gemini-cliProject install goes to .agents/skills/; add -g for ~/.gemini/skills/.
$ gh skill install NVIDIA/skills hsb-app --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-app .gemini/skills/hsb-app && 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-app" agent skill from https://github.com/NVIDIA/skills/tree/main/skills/hsb-app into .gemini/skills/hsb-app/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "hsb-app", 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-appInstalls 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-app -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-app .github/skills/hsb-app && 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-app" agent skill from https://github.com/NVIDIA/skills/tree/main/skills/hsb-app into .github/skills/hsb-app/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "hsb-app", 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-app -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-app --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-app .opencode/skills/hsb-app && 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-app" agent skill from https://github.com/NVIDIA/skills/tree/main/skills/hsb-app into .opencode/skills/hsb-app/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "hsb-app", 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-appDiscover and run Holoscan Sensor Bridge example applications on a connected devkit.
Hsb App is an agent skill from NVIDIA/skills, published by the product's own GitHub organization. Discover and run Holoscan Sensor Bridge example applications on a connected devkit. Filters available apps by the user's platform, HSB software version, board type, and sensors. Supports timed execution, failure analysis, code-edit suggestions, and iterative re-runs.
Its SKILL.md is about 4.1k tokens, which your agent loads only when the skill is triggered. The skill folder holds 7 other files, including reference files (for example `BENCHMARK.md`, `evals/evals.json` and `references/phase-details.md`).
It sits in Development, covering Root cause analysis. 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 these tools, so the agent can use them without asking each time:
ReadWriteEditMultiEditGrepGlobBashFrom allowed-tools in the SKILL.md frontmatter.
Shell commands in SKILL.md call:
sshdockerFrom the folder's file list and the shell code blocks in SKILL.md.
No URLs in SKILL.md. Its commands use ssh and docker, which can reach the network depending on how they are called.
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.
Hsb App loads about 4.1k tokens when it runs, and up to ~9.7k if it reads all its reference files. Until then it costs about 69 tokens; SKILL.md has 1,586 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 noted patterns worth knowing about, such as sudo or a known installer.
REMOTE_SUDO sudo / sudo -n / "" — default to "sudo" if not set.allowed-tools: Read, Write, Edit, MultiEdit, Grep, Glob, BashAutomated 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). 1,586 words, ~4,125 tokens.
.claude/skills/hsb-app/SKILL.md (or your agent's skills folder). This skill also uses 5 other files; get the full folder from GitHub.Use this skill when the user wants to discover, select, and run Holoscan Sensor Bridge example applications on a devkit with a connected HSB board.
This skill assumes the devkit is already set up (SSH, demo container built, host configured, board connected). If setup is not complete, instruct the user to run /hsb-setup first.
This workflow runs applications inside the demo container. Only run it when the user explicitly invokes it.
Gate 1 — Read environment variables. Before doing anything else, check these variables and print their resolved values to the user:
SSH_TARGET Remote devkit login (e.g. nvidia@192.168.1.50). Ask the user if not set.
REMOTE_ROOT Remote working directory (e.g. /home/nvidia). Ask the user if not set.
REMOTE_SUDO sudo / sudo -n / "" — default to "sudo" if not set.
REMOTE_SSH_OPTS Additional SSH options (optional).
HSB_PLATFORM Platform hint (optional).SSH_TARGET and REMOTE_ROOT are required. Stop and ask the user for them if either is missing.
Gate 2 — Present the phase plan and get confirmation. Before taking any action:
If the user's request already includes platform, board type, and sensors, also state upfront:
examples/ and filter apps by the user's sensor type and platform--headless automatically — only if the user explicitly requests itdocker run, using python3 for Python-based examplesShow the phase plan:
HSB App — Phase Plan
Phase 0: Verify board connectivity and demo container readiness
Phase 1: Discover user setup and select application to run
Phase 2: Run application with monitoring, failure analysis, and iterative debugging
Phase 3: Generate session report (with option to save)Then ask explicitly: Shall I proceed with Phase 0? [Y/n] — do not start Phase 0 until the user confirms.
Gate 3 — Fast path check. After the user confirms in Gate 2, run this check before executing any Phase 0 commands:
ssh -o BatchMode=yes $REMOTE_SSH_OPTS $SSH_TARGET \
"grep _SESSION_VERIFIED /tmp/.claude_hsb_app_session/state.sh 2>/dev/null || echo 'no session'"If the output contains _SESSION_VERIFIED=true, skip Phase 0 and Phase 1 setup discovery — go directly to app selection and inform the user.
examples/ directory to build a list of applications compatible with the user's setup. Present the list and let the user choose an app to run.Reuse the same environment variables from the hsb-setup and hsb-flash skills:
SSH_TARGET for the remote login target (e.g. nvidia@agx-thor-host)REMOTE_ROOT for the remote working directoryREMOTE_SUDO for privileged commandsREMOTE_SSH_OPTS for additional SSH optionsHSB_PLATFORM as an optional platform hintIf these are set, notify the user of these settings and use them without re-asking.
Before Phase 0, print the resolved remote execution settings.
When no valid session state exists, show the full phase plan:
Then execute one phase at a time.
When the session state file (/tmp/.claude_hsb_app_session/state.sh) exists and contains _SESSION_VERIFIED=true, the skill skips Phase 0 and Phase 1 setup discovery because connectivity and hardware were already verified. Instead, inform the user and jump directly to app selection:
Session already verified — skipping connectivity checks.
SSH target: $SSH_TARGET
Board: HSB Lattice | FPGA: XXXX
Platform: AGX Thor | HSB version: X.X.X
Sensors: Dual IMX274
Proceeding directly to application selection.Then execute:
Phase 0 must be re-run (ignoring the fast path) when:
No such device errors), clear _SESSION_VERIFIED from the session state and re-run Phase 0 before retrying./hsb-app --full, run Phase 0 from scratch.See ## Phase gate below for the full confirmation protocol.
If something fails, do not just dump raw logs. Summarize:
See references/phase-details.md for full step-by-step phase instructions.
Use the same persistent SSH session model as hsb-setup and hsb-flash. Each phase runs as a single SSH heredoc block:
ssh -o BatchMode=yes $REMOTE_SSH_OPTS $SSH_TARGET bash -s <<'REMOTE'
set -e
# restore state from previous phase
source /tmp/.claude_hsb_app_session/state.sh 2>/dev/null || true
cd "${_CLAUDE_CWD:-__REMOTE_ROOT__}"
# phase commands
echo "=== Phase N: description ==="
command1
command2
# save state for next phase (preserves _SESSION_VERIFIED if already set)
_PREV_VERIFIED="${_SESSION_VERIFIED:-}"
mkdir -p /tmp/.claude_hsb_app_session
{
echo "export _CLAUDE_CWD=\"$(pwd)\""
echo "export PATH=\"$PATH\""
echo "export REPO_DIR=\"$REPO_DIR\""
echo "export VERSION=\"$VERSION\""
echo "export HSB_PLATFORM=\"$HSB_PLATFORM\""
echo "export BOARD_TYPE=\"$BOARD_TYPE\""
echo "export SENSORS=\"$SENSORS\""
echo "export FPGA_VERSION=\"$FPGA_VERSION\""
echo "export SELECTED_APP=\"$SELECTED_APP\""
echo "export APP_OPTIONS=\"$APP_OPTIONS\""
echo "export APP_TIMEOUT=\"$APP_TIMEOUT\""
[ "$_PREV_VERIFIED" = "true" ] && echo "export _SESSION_VERIFIED=true"
} > /tmp/.claude_hsb_app_session/state.sh
REMOTEReplace __REMOTE_ROOT__ with the literal value of $REMOTE_ROOT when composing the heredoc.
Application commands run inside the demo container. Use the detached pattern with a named container.
For apps with --timeout, use the watchdog pattern. For indefinite-run apps, stream logs and wait for the user to request a stop.
After every app run, stop and remove the container. See references/phase-details.md for the cleanup pattern.
After Phase 3 (or on any failure that stops the workflow):
docker ps --filter "name=hsb_app_" --format '{{.Names}}' | xargs -r docker stop -t 2 2>/dev/null || true
ssh -o BatchMode=yes $REMOTE_SSH_OPTS $SSH_TARGET "rm -rf /tmp/.claude_hsb_app_session"After completing each phase (Phases 0–2), always prompt the user for confirmation before starting the next phase.
Exception: When --y (auto-approve mode) is active, phase gates are skipped. See "Auto-approve mode (--y)" section.
Proceed to Phase <N+1> (<phase description>)? [Y/n]All prompts in this skill require explicit typed responses. Never treat a blank or Enter-only input as a selection — re-prompt the user instead.
App workflow paused after Phase N.
You can resume by re-invoking the skill.--help)If $ARGUMENTS contains --help or -h, print the following and stop:
HSB Application Runner Skill
USAGE
/hsb-app [OPTIONS]
OPTIONS
--help, -h Show this help message and exit
--verbose Show full raw command output for every phase
--y Auto-approve all phase gates (skip user confirmation
between phases). Not recommended — a confirmation
warning is shown before proceeding. All output is
saved to a timestamped log file.
--timeout N Set app runtime in seconds (default: no timeout,
app runs until user asks to stop)
--full Force full verification from Phase 0, even if the
session was already verified
ENVIRONMENT VARIABLES (set before invoking the skill)
SSH_TARGET Remote login target (e.g. ubuntu@10.0.0.1)
REMOTE_ROOT Remote working directory
REMOTE_SUDO Privilege escalation: 'sudo', 'sudo -n', or ''
REMOTE_SSH_OPTS Additional SSH options
HSB_PLATFORM Platform hint
HSB_REPO_DIR Repo directory name under REMOTE_ROOT (default: holoscan-sensor-bridge)
Example: HSB_REPO_DIR=hololink → repo at $REMOTE_ROOT/hololink
WORKFLOW PHASES
Phase 0 Verify board connectivity and demo container readiness
(skipped on repeat runs in the same session)
Phase 1 Discover user setup, scan examples, select application
(setup discovery skipped on repeat runs)
Phase 2 Run application with monitoring and iterative debugging
Phase 3 Generate and optionally save session report
EXAMPLES
/hsb-app
/hsb-app --verbose
/hsb-app --timeout 60
/hsb-app --timeout 30 --verbose
/hsb-app --y
/hsb-app --y --timeout 120
/hsb-app --full
/hsb-app --help/hsb-app/hsb-app --verbose/hsb-app --timeout 60/hsb-app --timeout 30 --verbose/hsb-app --y/hsb-app --y --timeout 120/hsb-app --full/hsb-app --full --verbose/hsb-app --help--verbose)The skill supports a --verbose flag:
Check whether $ARGUMENTS (the text after the slash command) contains any of: --help / -h, --verbose, --y, --timeout N, or --full (case-insensitive). Strip all flags (and their values) from arguments before further parsing.
When --full is present, ignore any cached session state and run Phase 0 from scratch.
--verbose)--y)The skill supports a --y flag that skips all phase gates and runs the entire workflow from start to finish without waiting for user confirmation between phases. This is not recommended for normal use.
When --y is detected, display a warning and ask the user to confirm:
⚠ WARNING: Auto-approve mode (--y) is enabled.
This is NOT RECOMMENDED. All phase gates will be skipped and the entire
workflow will run without pausing for your confirmation between phases.
You will not be able to review intermediate results, ask questions, or
abort between phases. All output will be saved to a timestamped log file.
NOTE: In auto-approve mode, the app selection in Phase 1 will still
require your input (you must choose which app to run), but the app will
run with default settings automatically. Debug iterations in Phase 2
will be skipped — the app runs once and the result is reported.
Type 'yes' to confirm auto-approve mode, or anything else to cancel:--y is active--timeout was specified on the command line (to avoid indefinite hangs).hsb-app-log-YYYY-MM-DD-HHMMSS.md in $REMOTE_ROOT/ or current directory.--y --verbose: Auto-approve with full raw output.--y --timeout N: Auto-approve with a fixed app runtime.--y alone: Auto-approve with concise output and no timeout (app runs for a default 30 seconds in auto-approve mode to avoid indefinite hangs).--timeout)The skill supports a --timeout N flag where N is the number of seconds to run the application.
Match --timeout followed by a whitespace-separated integer in $ARGUMENTS. Example: --timeout 60.
docker stop. The output collected during that window is shown to the user.© 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 5 other files (references) in skills/hsb-app of NVIDIA/skills.
Open the folder on GitHubat commit 0e0d506
Hsb App 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 App this skillNVIDIA/skills | 3.5k | — | ~4.1k | Automated safety check: Notes | Apache-2.0 | |
| Megatron-LM CI Failure TriageNVIDIA/Megatron-LM | 18k | — | ~1.6k | Automated safety check: Pass | Apache-2.0 | |
| Code Design Rationale Investigatorcursor/plugins | 10k | 9 repos | ~2.6k | Automated safety check: Pass | None | |
| LLM Torch Profiler Analysissgl-project/sglang | 37k | 2 repos | ~6.4k | Automated safety check: Pass | Apache-2.0 | |
| OpenLogi macOS Permissions TriageAprilNEA/OpenLogi | 23k | — | ~2.5k | Automated safety check: Notes | Apache-2.0 | |
| Bug Finder for daisyUIsaadeghi/daisyui | 43k | — | ~2.3k | Automated safety check: Pass | MIT |
NVIDIA/Megatron-LM
Investigates a failing GitHub Actions run or job for Megatron-LM, finds the root cause plus the PR and test author involved, and files a structured bug issue.
cursor/plugins
Digs into why code is shaped the way it is by checking git history, pull requests and connected tools in parallel, then reporting a cited read on the tradeoffs.
sgl-project/sglang
Unified LLM torch-profiler triage skill for sglang, vllm, TensorRT-LLM, and TokenSpeed.
AprilNEA/OpenLogi
Decides whether an OpenLogi device problem on macOS is a privacy-permission (TCC) problem, using agent log lines, and says which identity needs which grant.
saadeghi/daisyui
Investigates suspected bugs in the daisyUI monorepo through read-only analysis, then writes a decision-ready fix plan in tmp/bugs without changing any product code.
garrytan/gstack
Investigates bugs, errors and stack traces in phases and requires a root-cause hypothesis to be confirmed before any fix is written.
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
Categories
Discover and run Holoscan Sensor Bridge example applications on a connected devkit. Hsb App is an agent skill from NVIDIA/skills, published by the product's own GitHub organization. Discover and run Holoscan Sensor Bridge example applications on a connected devkit.
Hsb App fits situations like: tasks that involve Root cause analysis.
Run `npx skills add NVIDIA/skills --skill hsb-app -a claude-code`. Or copy the skill folder (skills/hsb-app in NVIDIA/skills) into .claude/skills/hsb-app in your project. Claude Code loads it when a task matches its description.
Run `npx skills add NVIDIA/skills --skill hsb-app -a codex`. Or copy the skill folder (skills/hsb-app in NVIDIA/skills) into .agents/skills/hsb-app 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-app -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-app, .gemini/skills/hsb-app, .github/skills/hsb-app and .opencode/skills/hsb-app in your project.
Going by SKILL.md and its folder, Hsb App needs the command-line tools its instructions call (ssh and docker). Our summary lists: Docker. Its frontmatter pre-approves these tools: Read, Write, Edit, MultiEdit, Grep, Glob, Bash.
SKILL.md contains no URLs. Its commands use ssh and docker, which can reach the network depending on how they are called. This is read from the text; nothing was executed.
Our automated static check of SKILL.md found notes only (runs commands with sudo; pre-approves every shell command (allowed-tools: bash)), nothing it rates as a warning. It is not a guarantee. Review the folder before installing.
Hsb App 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 4.1k tokens (SKILL.md is roughly 17k 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 5.6k tokens, read only when the agent opens those files.
Skills that share tags, products or a category with Hsb App: Megatron-LM CI Failure Triage (NVIDIA/Megatron-LM, 18k stars), Code Design Rationale Investigator (cursor/plugins, 10k stars), LLM Torch Profiler Analysis (sgl-project/sglang, 37k stars) and OpenLogi macOS Permissions Triage (AprilNEA/OpenLogi, 23k 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.