Skip to content
Get packs

Guide · Agent skill

How to use the BIDS Neuroimaging Data Organizer agent skill

Updated · written with AI from the project's own documentation and checked against our catalogue data

The bids agent skill gives an AI agent working knowledge of the Brain Imaging Data Structure (BIDS), a shared standard for naming and organizing neuroscience files such as MRI, EEG and MEG recordings. This guide is for researchers, lab staff and students who want the agent to help build BIDS folder layouts, query datasets, run the validator or convert scanner data. By the end you will know how to install the skill, phrase requests that trigger it, and what tools and checks you still need to run yourself.

BIDS Neuroimaging Data OrganizerK-Dense-AI/scientific-agent-skills · MIT

Quick start

  1. 1Download and unzip. Download the skill ZIP from this page and unzip the bids folder.
  2. 2Place the folder. Move it into ~/.claude/skills/ for personal use, or into a project's .claude/skills/ folder for one project.
  3. 3Or upload in the apps. In the Claude apps, go to Settings → Capabilities → Skills and upload the ZIP.
  4. 4Install the tools. Install pybids, the validator and any converters you need, plus dcm2niix for DICOM conversion.
  5. 5Make a natural request. Describe your goal and dataset path in plain language, mentioning BIDS.
  6. 6Validate the result. Run the official BIDS validator and review warnings before sharing or analysis.

What is an agent skill and what does this one do?

An agent skill is a folder of instructions and reference files that an AI agent loads when a task matches the skill's description. This skill is built around the BIDS specification. With it installed, the agent knows the BIDS rules and can help build compliant folder layouts, search existing datasets by subject, session or task, check them with the validator, convert scanner DICOM files into BIDS, write JSON metadata sidecars and create derivative outputs.

The skill also includes bundled reference notes on the specification, conversion tools, metadata fields and BEPs (BIDS Extension Proposals), plus an updater script for refreshing the standard's schema.

  • Organize raw imaging, electrophysiology or behavioral data into BIDS directories
  • Query a dataset with PyBIDS by subject, session, task, run or modality
  • Validate a dataset before sharing or submission
  • Convert DICOM to BIDS with HeuDiConv, dcm2bids or BIDScoin
  • Write dataset_description.json, JSON sidecars, events and participants files
  • Configure .bidsignore and create BIDS derivatives

Install the BIDS skill from the ZIP

Download the skill ZIP from this page and unzip it. For a personal skill available in all your projects, place the folder in ~/.claude/skills/. For a skill that only applies to one project, place it in that project's .claude/skills/ folder.

In the Claude apps you can instead upload the ZIP under Settings → Capabilities → Skills.

  • Personal install: unzip into ~/.claude/skills/
  • Project install: unzip into the project's .claude/skills/
  • Claude apps: Settings → Capabilities → Skills, then upload the ZIP

Install the Python tools the skill uses

The skill expects Python 3.10 or newer. The source lists these installation commands for the core libraries. Install only the converters you need.

The validator can be installed through its PyPI wrapper or directly with Deno.

uv pip install pybids uv pip install bids-validator-deno # Alternative: deno install -ERWN -g -n bids-validator jsr:@bids/validator uv pip install heudiconv uv pip install dcm2bids # BIDScoin: uv pip install bidscoin uv pip install nibabel uv pip install pydicom

Trigger the skill with natural requests

You do not need special syntax. Describe the task in plain language and mention BIDS, a modality or a tool name. The agent should load the skill when the request matches its description.

Good requests name the goal and the data. Include the dataset path and what you already have, so the agent can choose between querying, validating and converting.

  • "Organize these MRI and EEG folders into a BIDS layout for OpenNeuro."
  • "Find all functional runs for subject 03 in session 2, task rest."
  • "Validate my dataset in ./study_bids and summarize the issues by code."
  • "Convert these DICOM exports to BIDS using a heuristic file."
  • "Write a dataset_description.json for a new BIDS dataset named Pilot Study."

Examples of what you can ask for

Typical sessions move from raw files to a checked dataset. The agent can draft the layout and sidecars, then point you to the validator for a formal check. For derivatives, the skill covers the derivatives layout and its own dataset_description.json.

  • Build a sidecar JSON for a task fMRI run and list the fields still missing.
  • Use PyBIDS to list every EEG file for a given task and build paths from entities.
  • Explain why a validator warning about inconsistent subjects appeared and how to document missing sessions in participants.tsv.
  • Set up a .bidsignore file to exclude non-BIDS files from validation.

What you need before you start

The skill itself does not replace the tools. It needs Python 3.10+, PyBIDS, the validator (bids-validator-deno or Deno), and dcm2niix for DICOM conversion. Add HeuDiConv, dcm2bids or BIDScoin depending on which converter you use. Installation and schema updates need network access. No API credentials are required.

  • Python 3.10+
  • PyBIDS (pybids)
  • bids-validator-deno or Deno
  • dcm2niix (for DICOM conversion)
  • HeuDiConv, dcm2bids or BIDScoin, as needed
  • Network access for installation and schema updates

Safety notes: scripts and network access

This skill runs scripts and uses the network. The bundled scripts/update_schema.py fetches the current schema and BEP list; run it from the skill directory with python scripts/update_schema.py. The bundled snapshot is BIDS 1.11.2 with schema 2.0.0. BEP proposals are not adopted requirements.

The skill does not address privacy or data sharing approval. Clinical or identifiable patient data needs de-identification review before it is shared or uploaded to a repository. Successful indexing is not a full compliance check, so always run the official validator and inspect warnings yourself.

  • Read scripts before running them, and run them from a trusted copy of the skill.
  • Check network-dependent steps when working with sensitive data.
  • Record the validator and BIDS schema versions with your dataset.
python scripts/update_schema.py

Tips for better results

Give the agent the exact dataset path and tell it whether files are already in BIDS form. If validation fails with "Not a BIDS dataset", the usual cause is a missing dataset_description.json at the root. For missing SliceTiming, the skill advises recovering actual slice offsets from scanner metadata or a verified protocol rather than guessing from slice order.

  • Start with a small subset of subjects to test a heuristic before converting everything.
  • Keep the original DICOM data; conversion should not overwrite it.
  • Review validator issue codes in the JSON report rather than relying on a summary.

What people use it for

  • Preparing an MRI study for deposit on OpenNeuro or DANDI
  • Writing PyBIDS scripts that select files by subject, session and task
  • Converting a new scanner export to BIDS with a HeuDiConv heuristic
  • Teaching students how BIDS entities and directory levels are structured

Frequently asked questions

Is the BIDS agent skill free?

The skill is published under the MIT licence on the K-Dense-AI/scientific-agent-skills repository. The source does not list prices, so check the repository for current terms.

What do I need to install before using it?

You need Python 3.10 or newer, PyBIDS, the BIDS validator (bids-validator-deno or Deno), and dcm2niix for DICOM conversion. Add HeuDiConv, dcm2bids or BIDScoin if you convert scanner data.

Where do I put the skill folder?

Unzip it into ~/.claude/skills/ for personal use or into a project's .claude/skills/ folder. In the Claude apps you can upload the ZIP under Settings → Capabilities → Skills.

Can the skill replace the BIDS validator?

No. The skill warns that successful indexing is not a full compliance check, so you must still run the official validator and inspect its warnings.

Is it safe to use with patient data?

The skill does not address privacy or sharing approval. Clinical or identifiable data should go through de-identification review before you share it or upload it to a repository.

What are the alternatives for converting DICOM to BIDS?

The skill covers three converters: HeuDiConv (Python heuristics), dcm2bids (config-file based) and BIDScoin. Choose based on whether your team prefers heuristic code or config files.

New agent skills, repos and AI tools — every week

The best new skills and open-source projects, the fastest-growing repos, and one free prompt. Five minutes to read. Unsubscribe anytime.

Our verdict

This skill suits neuroscience labs, researchers writing PyBIDS code and students learning BIDS who work with an agent. It is not a substitute for the official validator or a privacy review of identifiable clinical data.

Download the skill