imaging-data-commons

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 0.90). The skill's required workflow (SKILL.md) explicitly instructs the agent to query and ingest public third‑party data and metadata—e.g., using client.sql_query() and client.download_from_selection() against public IDC cloud buckets, BigQuery (references/bigquery_guide.md), DICOMweb public proxy, and manifest S3/GS URLs—so the agent will read and act on untrusted public content (IDC/portal/bucket data and forum links) that can materially determine downloads, viewer URLs, and license/processing decisions.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.70). The skill includes a runtime package upgrade/install that will fetch and execute the idc-index package from PyPI (https://pypi.org/project/idc-index/), which is a required external dependency retrieved and executed at runtime.