bio-vcf

Warn

Audited by Gen Agent Trust Hub on Feb 16, 2026

Risk Level: MEDIUMPROMPT_INJECTION
Full Analysis
  • PROMPT_INJECTION (MEDIUM): The skill presents an attack surface for indirect prompt injection through data processing.
  • Ingestion points: Genomic data is ingested from external VCF/BCF files via the pysam library in all included scripts.
  • Boundary markers: No boundary markers or delimiters are implemented to distinguish data from instructions when the agent reads the tool's output.
  • Capability inventory: The skill possesses file-writing capabilities using pysam's write mode and Path.write_text, and it can output large data structures to stdout.
  • Sanitization: String-based metadata fields within the VCF files are not sanitized, allowing potentially malicious instructions to be passed through to the agent context.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Feb 16, 2026, 08:33 AM