Agent Eval
Security checks across malware telemetry and agentic risk
Overview
The skill is a coherent agent-evaluation guide, but it asks for recurring access to task history, persistent memory writes, and reporting to a boss without enough scoping or consent controls.
Install only if you want an agent to evaluate historical task logs and maintain persistent improvement records. Before enabling the scheduled workflow, set explicit allowed memory paths, require confirmation before writes or reports, define approved recipients, and avoid sending sensitive task details without review.
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
63/63 vendors flagged this skill as clean.