Security audit
Protein Mining
Security checks across malware telemetry and agentic risk
Overview
This is a non-executable bioinformatics workflow guide with no evidence of hidden actions, credential access, persistence, or data exfiltration.
Before installing, verify the referenced bioinformatics tools and any separate protein-mining command from trusted sources. Treat generated reports, figures, and candidate proteins as research assistance that still requires expert review and reproducibility checks.
SkillSpector
By NVIDIA
Vulnerability Patterns
- Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
- Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
- Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
- Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
- Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
VirusTotal
65/65 vendors flagged this skill as clean.