BrainNetworkTransformer (BNT) (Claude Skill)

Run BrainNetworkTransformer for fMRI phenotype prediction, including data loading, training, and evaluation.

Type	Claude Skill
Supplier	CUHK-AIM-Group (community OSS, MIT)
Availability	GA — part of the NeuroClaw neuroimaging skill library
Pricing	Free / OSS (MIT)
Capabilities	Read/Write — Claude runs the skill’s Python locally (Bash), not as an MCP tool

How to install

• Claude Code — clone and copy the skill into your skills directory:

  git clone https://github.com/CUHK-AIM-Group/NeuroClaw
  cp -r NeuroClaw/skills/bnt ~/.claude/skills/
  

  Project-scoped alternative: copy into .claude/skills/ instead. NeuroClaw skills assume the collection’s shared helpers (claw-shell, modality tool skills) and the upstream neuroimaging stack (FreeSurfer/FSL/fMRIPrep/etc.) — install those dependencies, or run the bundled installer for the full environment:

  cd NeuroClaw && python installer/setup.py
  

  which configures the Python env, CUDA/GPU, and the neuroimaging tools.

What it does

Use this model doc whenever the user wants to run BrainNetworkTransformer for fMRI phenotype prediction, including data loading, training, and evaluation. BNT uses dense FC matrices (no PyG dependency) with DEC pooling + interpretable transformer encoder.

Primary use cases: Run BrainNetworkTransformer for fMRI phenotype prediction, including data loading, training, and evaluation.

Notes

Distributed as a SKILL.md (plus code examples) in the NeuroClaw skill library — Claude executes it locally via Bash/Python rather than as an MCP server. Upstream license: MIT. The skill directory upstream is skills/bnt.

Sources

• CUHK-AIM-Group/NeuroClaw
• skills/bnt/SKILL.md

---

Installed this tool?

Share feedback — install path, OS, errors, workarounds. The form opens with this tool pre-selected and a link back to this page.