Production Agent Builder
Security checks across malware telemetry and agentic risk
Overview
This skill is a planning guide for designing AI agents and does not include code, installation steps, credential access, or hidden execution behavior.
This appears safe to install as a planning aid. When applying the framework to build real agents, review any separate tools, credentials, memory stores, and action permissions you add, and keep human approval for email, data changes, spending, deletion, and publishing.
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
65/65 vendors flagged this skill as clean.