Address security findings from K-Dense-AI audit

- Pin idc-index version in auto-upgrade subprocess call to REQUIRED_VERSION
- Document expected network endpoints in Overview (GCS, S3, BigQuery, DICOMweb)
- Add tested-with comment for optional dependencies

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

Author: fedorov

CHANGELOG.md

@@ -11,6 +11,12 @@ and this project adheres to [Semantic Versioning](https://semver.org/).
 
 - `tests/test_bq_snippets.py`: BigQuery snippet validation using `bq query --dry_run` — 33 tests covering all SQL examples in `references/bigquery_guide.md` (dicom_all, original_collections_metadata, segmentations, quantitative_measurements, qualitative_measurements, private elements, and clinical tables); skips automatically when `bq` CLI is unavailable or unauthenticated
 
+### Security
+
+- Fixed auto-upgrade subprocess call to pin `idc-index` to `REQUIRED_VERSION` (was `"idc-index"`, now `f"idc-index=={REQUIRED_VERSION}"`), ensuring the installed version always matches the tested version declared in the frontmatter
+- Added network access transparency note to Overview documenting expected external endpoints (GCS, S3, BigQuery, DICOMweb proxy, Google Healthcare API) and clarifying that no credentials or environment variables are accessed by the skill
+- Added tested-with version comment to optional dependency install block (`pandas>=1.5, numpy>=1.23, pydicom>=2.3`)
+
 ### Changed
 
 - Updated to idc-index 0.12.1 (idc-index-data 24.0.4, IDC data version v24)

SKILL.md

@@ -16,6 +16,8 @@ metadata:
 
 Use the `idc-index` Python package to query and download public cancer imaging data from the National Cancer Institute Imaging Data Commons (IDC). No authentication required for data access.
 
+**Expected network access:** `idc-index` queries a local DuckDB index (no network for metadata). File downloads use public GCS (`storage.googleapis.com`) and AWS S3 (`s3.amazonaws.com`) — no authentication required. DICOMweb access uses either the public IDC proxy (`proxy.imaging.datacommons.cancer.gov`, no auth) or the Google Cloud Healthcare API (`healthcare.googleapis.com`, requires GCP authentication). Optional BigQuery queries (`bigquery.googleapis.com`) also require GCP authentication. No credentials or environment variables are accessed by this skill.
+
 **Current IDC Data Version: v24** (always verify with `IDCClient().get_idc_version()`)
 
 **Primary tool:** `idc-index` ([GitHub](https://github.com/imagingdatacommons/idc-index))
@@ -31,7 +33,7 @@ installed = idc_index.__version__
 if installed < REQUIRED_VERSION:
     print(f"Upgrading idc-index from {installed} to {REQUIRED_VERSION}...")
     import subprocess
-    subprocess.run(["pip3", "install", "--upgrade", "--break-system-packages", "idc-index"], check=True)
+    subprocess.run(["pip3", "install", "--upgrade", "--break-system-packages", f"idc-index=={REQUIRED_VERSION}"], check=True)
     print("Upgrade complete. Restart Python to use new version.")
 else:
     print(f"idc-index {installed} meets requirement ({REQUIRED_VERSION})")
@@ -245,6 +247,7 @@ If you see an older version, upgrade with: `pip install --upgrade idc-index`
 
 **Optional (for data analysis):**
 ```bash
+# Tested with: pandas>=1.5, numpy>=1.23, pydicom>=2.3
 pip install pandas numpy pydicom
 ```