protein-assembly
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFEPROMPT_INJECTIONDATA_EXFILTRATIONNO_CODE
Full Analysis
- Indirect Prompt Injection (LOW): The skill processes untrusted external sequence data. Ingestion points: local GenBank/FASTA files and PDB/fpbase APIs. Boundary markers: Absent. Capability inventory: Network queries and file-write operations. Sanitization: Absent.
- Data Exposure (LOW): Instructions require querying non-whitelisted domains (PDB API, fpbase) and reading user-provided sequence files in full.
- No Code (SAFE): No executable code or external scripts are included in the skill.
Audit Metadata