What does this skill do, and when should you use it?
This skill introduces the Isaac for Healthcare endovascular catheter navigation workflow. It explains how the fluorosim stack, XPBD catheter physics, and vasculature digital twin fit together, and lists the available CLI modes. It is an orientation and routing skill, not an execution skill. Hands-on work continues through the stage-specific skills.
The skill locates the i4h-workflows checkout, reads the mode registry in workflows/catheter_navigation/metadata.json, and runs ./i4h modes catheter_navigation to enumerate supported modes. It describes CT preprocessing, vessel segmentation, DRR rendering, and the interactive viewport, identifies relevant scripts, packages, and tests, and routes users to setup, digital-twin, rendering, viewport, smoke-test, or end-to-end skills.
- A researcher wants a high-level map of the catheter navigation workflow before running a stage.
- A developer needs to confirm whether CT preprocessing, vessel segmentation, DRR rendering, or an interactive viewport is supported.
- A medical simulation user is preparing a vasculature digital twin from patient CT data.
- An engineer needs to check the GPU, CUDA, and Vulkan prerequisites for GPU-backed modes.
- A user encounters a missing mode or fluorosim dependency import error and needs the correct troubleshooting path.
How do you install this skill?
- The fallback flow may clone code from GitHub into the user's home directory without confirmation; confirm the destination, network source, and change scope before execution.
- The skill concerns patient-specific CT/DICOM workflows but does not specify de-identification, storage, logging, or access-control requirements; do not send sensitive medical data to unreviewed external services.
- The core path depends on GitHub, CUDA/Vulkan, an NVIDIA GPU, and multiple unpinned packages; initialization may fail on mainland-China or restricted networks.
- BENCHMARK is a skill-supplied evaluation report and is not an independent reproduction performed in this static review.
- Shell / CLI
- Network access
- Local filesystem
Giti4h-workflows repositoryNVIDIA GPUCUDA >= 12.8Vulkan-capable driver
Install the skill from NVIDIA/skills with the skills CLI:
npx skills add nvidia/skills --skill i4h-catheter-navigation --yes
The repository is dual-licensed under Apache License 2.0 and CC BY 4.0; this skill’s SKILL.md declares Apache-2.0. The README does not document a separate client-specific installation directory.
tmp="$(mktemp -d)"
git clone --depth 1 https://github.com/NVIDIA/skills.git "$tmp"
mkdir -p ~/.claude/skills
cp -R "$tmp/skills/i4h-catheter-navigation" ~/.claude/skills/
rm -rf "$tmp"Generated from the source repository and skill path; it copies only this skill's folder. If the author's install steps above differ, follow those first. To scope it to one project, replace ~/.claude/skills with that project's .claude/skills.
How do you use this skill?
Once installed, send your agent any of these to trigger it:
- What is the catheter navigation workflow, which modes are supported, and where should I start?
- ${I4H_WORKFLOWS:-$(git rev-parse --show-toplevel 2>/dev/null)}
After installation, ask the agent: “What is the catheter navigation workflow, which modes are supported, and where should I start?” The skill locates or reuses i4h-workflows and runs:
REPO_ROOT="${I4H_WORKFLOWS:-$(git rev-parse --show-toplevel 2>/dev/null)}"; [ -d "$REPO_ROOT/workflows/catheter_navigation" ] || REPO_ROOT="$HOME/i4h-workflows"
"${REPO_ROOT}/i4h" modes catheter_navigation
If no checkout is available, the skill documents cloning isaac-for-healthcare/i4h-workflows. Exact stage commands are defined by metadata.json and the stage-specific skills; this overview does not execute a pipeline stage.
What are this skill's strengths and limitations?
- Clearly separates mode registration, workflow components, and follow-on stage skills.
- Provides a single orientation point for CT, the vasculature digital twin, DRR, and XPBD catheter physics.
- Includes troubleshooting directions for missing modes and dependency import failures.
- Does not bundle CT data, allowing users to bring their own NIfTI or DICOM data.
- Only orients and routes; it does not preprocess, render, or simulate.
- Patient-specific workflows require the separate digital-twin stage.
- GPU modes require an NVIDIA GPU, CUDA >= 12.8, and a Vulkan-capable driver.
- RL/IL integration with Isaac Lab is not a first-class stage in v0.7.
- Complete stage execution commands are not provided here and depend on metadata.json and other skills.
How does this skill compare with similar options?
Side by side with related skills; every score comes from the same FSRS standard.
| Skill | FS score | Stars | Last updated | License |
|---|---|---|---|---|
| i4h Catheter Navigation Workflow Guide this page ✓ NVIDIA · Official | 52 · Use with care | ★ 3.8k | today | Apache-2.0 |
| Catheter Navigation DRR Renderer ✓ NVIDIA · Official | 52 · Use with care | ★ 3.8k | today | Apache-2.0 |
| Catheter Navigation Fluoroscopy Viewport ✓ NVIDIA · Official | 50 · Use with care | ★ 3.8k | today | Apache-2.0 |
| Catheter Navigation E2E Smoke Test ✓ NVIDIA · Official | 45 · Not recommended | ★ 3.8k | today | Apache-2.0 |
| Catheter Navigation Setup Verification ✓ NVIDIA · Official | 51 · Use with care | ★ 3.8k | today | Apache-2.0 |
How did FollowSkills review this skill?
The skill clearly states that it is routing-only and does not run pipeline stages, and it identifies no credential requirement or covert exfiltration behavior. However, it may clone an external GitHub repository into the user's home directory without confirmation and requires command execution, while omitting data-flow disclosure, patient-data handling, network trust boundaries, rollback, and explicit confirmation; points are deducted for these gaps.
The mode registry, fallback table, prerequisites, and troubleshooting guidance are broadly consistent, and BENCHMARK reports external evaluation results. Static review did not execute the commands, and availability of the checkout, GPU, driver, and multiple packages is not reproducible from the supplied files; abnormal-input coverage and failure diagnostics are limited, so the score remains below the static ceiling.
Trigger scenarios, supported modes, stage skills, and the non-execution boundary are reasonably clear, including the distinction between synthetic rendering and patient-specific flows. Chinese-language support and mainland-China reachability are not addressed, and the core path depends on GitHub cloning and NVIDIA GPU infrastructure; non-fit boundaries are incomplete, so points are deducted.
The material provides a name, version, Apache-2.0 license, owner, dependencies, limitations, troubleshooting, skill index, and benchmark report. However, the benchmark flags misplaced version metadata, recommended Instructions and Examples sections are missing, and changelog, maintenance/update path, and rollback guidance are unclear; full marks are not justified.
As a workflow overview and router, the purpose, modes, stage sequence, and required response structure are clear. The BENCHMARK reports correctness and effectiveness gains on positive tasks, but it covers only four tasks and was not executed in this review; useful output still depends on other skills and external infrastructure, so the static cap applies.
The supplied material gives a revision, references to the mode registry and README, dependency locations, and an evaluation report, providing some auditability. It lacks a committed test suite for this selected skill or reproducible execution logs; benchmark results are largely report-level claims and lack corroboration, so evidence coverage is limited.
Open a dimension to read why it scored that way
Evidence confidence：Low — Mostly static review, author material or a limited demo; useful for discovery, not high-risk decisions.
See the full review method →