agentsclimarketplace

I4h catheter navigation e2e

Skill NVIDIA/skills/skills/i4h-catheter-navigation-e2e

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.

Install
npx -y skills add NVIDIA/skills --skill i4h-catheter-navigation-e2e

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

What its author says it does

Copied from the file, not written here

End-to-end smoke for catheter navigation covering setup, digital twin, DRR, and unit tests. Use when asked to run the full catheter workflow smoke or demo the v0.7 pipeline.

The file declares its own license as Apache-2.0. That is the author’s claim about this one file, and it is not the same thing as the license GitHub reports for the repository, which is listed with the other numbers below.

SKILL.md

3.8 KB, as published. Nobody here has run it

i4h Catheter Navigation - End-to-End Smoke

Purpose

Run a v0.7 smoke pipeline: verify setup, build a digital twin from a TotalSegmentator subject, render a DRR from cache, and run CPU unittests. Skips the interactive viewport (requires display).

Base Code

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

Stages map to individual skills; this skill chains them with shared RUN_DIR and CACHE.

StageSkill
setup + CPU smoke[[i4h-catheter-navigation-setup]]
digital twin[[i4h-catheter-navigation-digital-twin]]
GPU DRR[[i4h-catheter-navigation-render-drr]]
unittest[[i4h-catheter-navigation-smoke]]

Run

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

Step 1 - initialize run

REPO_ROOT="${I4H_WORKFLOWS:-$(git rev-parse --show-toplevel 2>/dev/null)}"; [ -d "$REPO_ROOT/workflows/catheter_navigation" ] || REPO_ROOT="$HOME/i4h-workflows"
WF_ROOT="${REPO_ROOT}/workflows/catheter_navigation"
RUN_DIR="${WF_ROOT}/runs/e2e_$(date +%Y%m%d_%H%M%S)"
mkdir -p "${RUN_DIR}/logs"
ln -sfn "${RUN_DIR}" "${WF_ROOT}/runs/.latest"
CACHE="${CACHE:-/tmp/ct_cache}"
SUBJ="${SUBJ:-}"
export I4H_WORKFLOWS="$REPO_ROOT"

Step 2 - setup verify

Follow [[i4h-catheter-navigation-setup]] Steps 1-3; tee to ${RUN_DIR}/logs/setup.log.

Step 3 - digital twin (requires SUBJ)

Set SUBJ to an extracted TotalSegmentator subject, then follow [[i4h-catheter-navigation-digital-twin]] Steps 3-4.

If SUBJ is unset, skip to Step 4 with synthetic DRR only and note partial e2e in the summary.

Step 4 - render DRR

If Step 3 succeeded, set CACHE and follow [[i4h-catheter-navigation-render-drr]] cache variant. Otherwise run synthetic render to ${RUN_DIR}/drr_synthetic.png.

Step 5 - smoke tests

Follow [[i4h-catheter-navigation-smoke]]; log to ${RUN_DIR}/logs/smoke.log.

Step 6 - summary

cat > "${RUN_DIR}/logs/SUMMARY.txt" <<EOF
Catheter navigation e2e smoke
RUN_DIR=${RUN_DIR}
CACHE=${CACHE}
SUBJ=${SUBJ:-<skipped>}
Artifacts:
  $(ls -1 "${RUN_DIR}" 2>/dev/null || true)
EOF
cat "${RUN_DIR}/logs/SUMMARY.txt"

Flags (informal)

  • Skip digital twin when SUBJ unset -> synthetic DRR only.
  • Add viewport manually via [[i4h-catheter-navigation-viewport]] when a display is available.

Outputs

  • RUN_DIR/logs/ - per-stage logs and SUMMARY.txt
  • RUN_DIR/drr.png or drr_synthetic.png - rendered frame
  • CACHE/ - digital twin artifacts when Step 3 ran

Prerequisites

  • [[i4h-catheter-navigation-setup]] host requirements.
  • SUBJ for full e2e (otherwise partial smoke).
  • GPU for Step 4 cache render.

Limitations

  • Does not launch the interactive viewport (display required).
  • Does not cover RL/IL training loops.

Troubleshooting

  • Partial run when no CT data - expected; document which stages ran.
  • GPU render failure - fall back to synthetic DRR and report GPU setup gap.

Final Response

Print RUN_DIR, stages completed/skipped, DRR output path, smoke-test status, and recommend [[i4h-catheter-navigation-viewport]] for interactive demo.

Keep looking

Skills are one crate of 328,083. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.