alterlab-tiledbvcf

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

• Third-party content exposure detected (high risk: 0.80). The skill's required workflows explicitly ingest and read VCF/BCF data from open cloud storage and URLs (e.g., ds.ingest_samples([...]) and the TileDB-Cloud example source="s3://my-bucket/vcf-files/" / tiledb:// URIs in SKILL.md), which are untrusted, user-provided third‑party files that the agent reads and whose contents can influence subsequent queries and operations.