Subagent Isolation Guard
Security checks across malware telemetry and agentic risk
Overview
This is a small instruction-only skill for isolating subagents, with disclosed routing and workspace rules and no evidence of hidden execution or data access.
Install this only if you intentionally want subagent sessions marked with :subagent: to bypass semantic routing and use separate workspaces. Review any AGENTS.md or semantic-webhook-server.py changes before applying them, because those changes can persist and affect future subagent behavior.
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
66/66 vendors flagged this skill as clean.