Baichuan Ai
Security checks across malware telemetry and agentic risk
Overview
This is an instruction-only skill for summarizing public Baichuan AI pages, with no code, credentials, account actions, or persistence requested.
This appears safe to install for lightweight summaries of public Baichuan AI information. Verify pricing, quotas, and model capability details against the original Baichuan pages before relying on them, especially because public product information can change.
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
64/64 vendors flagged this skill as clean.