tiledbvcf
Pass
Audited by Gen Agent Trust Hub on Oct 1, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes genomic variant data from external VCF/BCF files and cloud storage (S3, Azure, GCS). This ingestion of untrusted data represents a potential attack surface where maliciously crafted genomic files could attempt to influence agent behavior.
- Ingestion points:
ds.ingest_samples()andds.read()operations described inSKILL.mdprocess external files and cloud-hosted URIs. - Boundary markers: The instructions do not define explicit delimiters or warnings for the agent to ignore instructions embedded within the processed genomic data.
- Capability inventory: The skill possesses capabilities for file system access (
ds.export,ds.ingest_samples) and network operations via cloud storage integration (S3,GCS,Azure) and thetiledb-cloudAPI. - Sanitization: No explicit sanitization or filtering of the variant data content is mentioned to prevent instructions from reaching the agent context.
Audit Metadata