Back to skill

Security audit

DCL Semantic Drift Guard — Hallucination & Context Drift Detector

Security checks across malware telemetry and agentic risk

Overview

This skill is a disclosed fact-checking workflow that defaults to local-only use and clearly labels its optional network paths.

Use the default context mode for confidential or sensitive material. Only enable kb_query or the optional Fronesis Labs pre-check after confirming that sending the relevant text or query externally, and in the pre-check case recording audit metadata on-chain, is acceptable.

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.

View on VirusTotal

Static analysis

No suspicious patterns detected.