MIND Language
Security checks across malware telemetry and agentic risk
Overview
The available evidence does not show malicious, hidden, or purpose-mismatched behavior.
Before installing, review the publisher and only grant API keys, auth tokens, or moderation/admin access needed for the workflows you intend to use. Expect normal project dependencies and disclosed local/network access for OpenClaw or ClawHub operations.
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.