mcpbeat

I4h Catheter Navigation

nvidia/i4h-catheter-navigation

Overview of `workflows/catheter_navigation/` (fluorosim DRR, XPBD physics, vasculature digital twin). Use when the user asks what the catheter navigation workflow is, what's supported, or where to start.

6k tokens
context cost
the whole folder, loaded on every use
5
files
instructions only
0
copies elsewhere
how many repositories repackaged it
2778
stars on the repo
on the repository, not the skill itself

Install

one command, takes just this skill from the repository
npx skills add https://github.com/NVIDIA/skills --skill i4h-catheter-navigation

What comes with it

17 748 bytes besides the instruction
BENCHMARK.md
evals/evals.json
skill-card.md
skill.oms.sig

The instruction itself

13 sections, as written by the author

i4h Catheter Navigation Workflow

Purpose

Orient on the endovascular catheter navigation workflow before touching a specific stage: which CLI modes exist, how fluorosim / physics / digital-twin pieces fit together, and which per-stage skill to invoke next. This skill routes - it runs no pipeline stage itself.

Base Code

These steps drive the i4h-workflows base code (the workflows/catheter_navigation/ tree). To reuse an existing checkout, set I4H_WORKFLOWS to its path (no clone happens). Otherwise this resolves the current repo, or clones to ~/i4h-workflows - pick that default without prompting. Run every command below from the resolved root:

# Resolve the i4h-workflows base code (provides workflows/catheter_navigation/).
ROOT="${I4H_WORKFLOWS:-$(git rev-parse --show-toplevel 2>/dev/null)}"
if [ ! -d "$ROOT/workflows/catheter_navigation" ]; then
  ROOT="${I4H_WORKFLOWS:-$HOME/i4h-workflows}"
  [ -d "$ROOT/workflows/catheter_navigation" ] || git clone https://github.com/isaac-for-healthcare/i4h-workflows "$ROOT"
fi
export I4H_WORKFLOWS="$ROOT"; cd "$ROOT"

Basics

  • Mode registry (source of truth): workflows/catheter_navigation/metadata.json - lists ./i4h modes and their runtime commands.
  • Human docs: workflows/catheter_navigation/README.md - data download, digital-twin build, viewport controls, troubleshooting.
  • Each pipeline stage has its own skill. Compose them or use [[i4h-catheter-navigation-e2e]] for a full smoke run.
  • No CT data is shipped; users bring their own NIfTI/DICOM (e.g. TotalSegmentator subset).

Supported Modes

Fallback when ./i4h modes catheter_navigation fails (setup incomplete):

| Mode | Purpose |

|---|---|

| preprocess_ct | CT -> attenuation volume (mu_volume.npy) + optional HU cache |

| segment_vessels | Arterial tree mask + centerline (vasculature digital twin) |

| render_drr | Single DRR frame (synthetic phantom if no cache) |

| interactive_viewport | Live fluoroscopy viewport with XPBD catheter physics |

Subprojects

| Directory | Purpose |

|---|---|

| workflows/catheter_navigation/scripts/simulation/fluorosim/examples/interactive_catheter_slang_viewport.py | Interactive viewport glue script (invoked by interactive_viewport mode) |

| workflows/catheter_navigation/scripts/simulation/fluorosim/rendering/ | Local DRR rendering helpers used by viewport fallback path |

| workflows/catheter_navigation/metadata.json | ./i4h mode registry and command wiring |

| Installed packages | fluorosim, vasculature_digital_twin, catheter_vasculature_solver provide renderer + digital twin + solver runtime |

| workflows/catheter_navigation/tests/ | CPU smoke tests (no GPU required) |

Run

Run the step below before answering. Each step is a separate bash call; variables persist in the local agent's tmux session.

Step 1 - list supported modes

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 the command fails, use the Supported Modes fallback table above.

Skill Index

| Skill | Purpose |

|---|---|

| i4h-catheter-navigation-setup | Host/GPU preflight, PYTHONPATH, smoke-test verify |

| i4h-catheter-navigation-digital-twin | Download CT data, preprocess, segment vessels |

| i4h-catheter-navigation-render-drr | Render a single DRR frame |

| i4h-catheter-navigation-viewport | Launch interactive fluoroscopy + catheter sim |

| i4h-catheter-navigation-smoke | Run CPU unittest smoke suite |

| i4h-catheter-navigation-e2e | End-to-end smoke (setup -> digital twin -> render -> tests) |

Prerequisites

  • NVIDIA GPU with CUDA >= 12.8 and Vulkan-capable driver for GPU modes (render, viewport).
  • For any hands-on stage, run [[i4h-catheter-navigation-setup]] first.
  • For patient-specific runs, run [[i4h-catheter-navigation-digital-twin]] before viewport or cache-based DRR.

Limitations

  • Overview/routing only - each stage has its own skill; this one performs no preprocessing, rendering, or simulation.
  • RL/IL training integration with Isaac Lab is not yet a first-class skill stage in v0.7; use the workflow README and fluorosim API for custom data generation.

Troubleshooting

  • Error: mode not found - Cause: typo or ./i4h not run from repo root. Fix: cd to repo root and re-run Step 1; inspect workflows/catheter_navigation/metadata.json.
  • Error: import errors for fluorosim / vasculature_digital_twin / catheter_vasculature_solver - Cause: dependencies not installed in active env. Fix: run [[i4h-catheter-navigation-setup]] and install README requirements.

Final Response

The response is incomplete unless it includes all three parts below, each separated by a blank line.

  • Orientation - one short paragraph: fluorosim stack, digital twin, ./i4h modes.
  • Supported modes - heading plus markdown table from Step 1 or the Supported Modes fallback table.
  • Available skills - heading plus the full Skill Index table copied into the response.

Then recommend [[i4h-catheter-navigation-setup]] if setup is not done, and name one next stage skill matched to the user's goal.

How to use it

Copy the folder

Take nvidia/i4h-catheter-navigation from the repository into ~/.claude/skills for personal use, or into .claude/skills inside a project.

Check the name does not clash

The agent identifies a skill by the name field in its header. Two skills with the same name cannot sit side by side — one of them will be ignored.