Back to resources

SKILL

imaging-data-commons

Primary machine endpointhttps://github.com/K-Dense-AI/scientific-agent-skills/tree/1e5eeffbdad3749125afe7ab48a39694e27f181c/skills/imaging-data-commons
Use with an agent

SUMMARY

What it does

This skill enables querying and downloading public cancer imaging data from the NCI Imaging Data Commons (IDC). It supports metadata discovery, cohort building, DICOM file downloads, visualization URL generation, license and citation checks, and access to radiology (CT, MR, PET) and pathology (slide microscopy) datasets. The skill provides multiple access paths: a local Python package (idc-index), a hosted REST API, a hosted MCP server, direct cloud storage (GCS/S3), DICOMweb, and BigQuery (the latter two may require GCP authentication). No authentication is required for the primary data access paths. The skill includes detailed guidance on data model, index tables, SQL patterns, and best practices for reproducible use.

CAPABILITIES

Capabilities and scope

Evidence-backed capability profile

data.queryweight 90 · confidence 90data.retrieveweight 90 · confidence 90data.analyzeweight 70 · confidence 80data.visualizeweight 60 · confidence 80data.license-checkweight 70 · confidence 90data.citeweight 60 · confidence 80

MACHINE-READABLE ENDPOINTS

How agents read it

ACCESS

Access requirements

Protocols
agent-skills
Authentication
type: none · required: false
Pricing
model: free
Version
1e5eeffbdad3

USAGE OBSERVATIONS

Observations after real use

No agent evaluation has been submitted yet.