Claude Skill MartBrowse skillsWhat is a Skill?
 
← Back to list

PathML Computational Pathology Workflow

A PathML 3.0.5 playbook for loading and tiling whole-slide images, building preprocessing/QC pipelines, validating spatial graphs, and planning bounded local inference.

Data & AnalyticsAdvanced★ 33,030⑂ 3,248AI score 8/10Last updated: Aug 9, 2026

What it does

Guides local, research-only computational pathology work with PathML 3.0.5:

  • Slide loading (OpenSlide / Bio-Formats backends), tiling, pyramid levels, and (i, j) coordinate bookkeeping
  • Preprocessing and QC pipelines via Pipeline + SlideData.run() (tissue detection, blur, empty-mask handling)
  • .h5path data management, manifest validation, and patient-level splits to prevent leakage
  • Multiplex (CODEX/Vectra) quantification and KNN/RAG/HACT spatial graph schema validation
  • Bounded inference planning: tile counts, RAM, output size, batching
  • Bundled offline CLIs in scripts/ for manifest validation, pipeline planning, image QC, schema checks, and inference planning

It also sets firm guardrails: de-identify pixels and metadata, keep re-identification keys out of the workspace, never load untrusted pickle/.pt checkpoints, verify ONNX SHA-256, and gate any remote model download behind explicit consent.

Who it's for

  • Digital/computational pathology researchers, grad students, and data scientists
  • Teams that need reproducible, auditable WSI tiling pipelines
  • Labs handling slides that may embed patient identifiers

Not a medical device: outputs must not be used to diagnose, grade, stage, or treat patients.

Example uses

  1. First pipeline: "Detect tissue on an H&E .svs slide and write 512px tiles to .h5path" → you get correct HESlide + TissueDetectionHE code plus an 8-tile smoke test before a full run.
  2. Capacity planning: "How many tiles and how much RAM for a 100000×80000 slide at 512px?" → run python scripts/plan_pipeline.py --width 100000 --height 80000 --tile-size 512 --stride 512.
  3. Spatial graph QA: "Check channel order and node-feature alignment in my graph JSON" → python scripts/validate_spatial_schema.py graph --input graph.json --root .

· · · Install guide · · ·

Install in the Claude app (no terminal)
  1. Download the ZIP with the button below.
  2. In Claude, open Settings → Capabilities and turn on 'Code execution and file creation'. (one time)
  3. Go to Customize → Skills → + → 'Upload a skill' and upload the ZIP.
↓ Download ZIP
Install in Claude Code

Let Claude do it — paste this into Claude Code

Install the skill I found on Claude Skill Mart.
Copy the skills/pathml folder from the GitHub repo K-Dense-AI/scientific-agent-skills into my ~/.claude/skills/pathml/.
When it's done, tell me in one line what this skill can do.

Install with a command instead

git clone https://github.com/K-Dense-AI/scientific-agent-skills.git && mkdir -p ~/.claude/skills && cp -r scientific-agent-skills/skills/pathml ~/.claude/skills/

⚠ This is a third-party skill. Check the source repository before installing.

  1. Open a terminal.
  2. Clone the repository: git clone https://github.com/K-Dense-AI/scientific-agent-skills.git
  3. Create the skills directory: mkdir -p ~/.claude/skills
  4. Copy just this skill: cp -r scientific-agent-skills/skills/pathml ~/.claude/skills/
  5. Install native prerequisites first:
    • Ubuntu/Debian: sudo apt-get install openslide-tools gcc g++ libblas-dev liblapack-dev openjdk-17-jdk
    • macOS: brew install openslide openjdk@17
  6. Create a Python 3.11 environment and install PathML: uv venv --python 3.11 && source .venv/bin/activate && uv pip install "pathml==3.0.5"
  7. Verify: python -c "import importlib.metadata as m; print(m.version('pathml'))"
  8. Restart Claude Code, then ask something like "Build a PathML preprocessing pipeline for my H&E slides" to trigger the skill.
View source on GitHub ↗License: MIT