tooluniverse-antibody-engineering
Pass
Audited by Gen Agent Trust Hub on Apr 16, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill demonstrates an indirect prompt injection surface by processing untrusted user-provided antibody sequences and target antigens without employing boundary markers or sanitization steps before tool interpolation.
- Ingestion points: 'vh_sequence', 'vl_sequence', and 'target_antigen' parameters in functions described in SKILL.md (e.g., identify_germline, search_clinical_precedents).
- Boundary markers: Absent in instructions and report templates.
- Capability inventory: File system write operations for reports and sequence files, and network-enabled tool calls via scientific databases (IMGT, IEDB, PubMed, UniProt).
- Sanitization: Absent; the skill assumes biological sequence integrity without verification.
Audit Metadata