防止AI幻觉
Security checks across malware telemetry and agentic risk
Overview
This is a single markdown skill that gives agents a fact-checking checklist and does not include code, installers, network targets, or hidden high-impact behavior.
This skill is suitable for users who want stricter fact-checking discipline from their agent. Before installing, note that it may make the agent slower or more conservative on factual tasks, and review any persistent-memory use if your conversations include sensitive information.
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.