Lyra — Cognitive Architect
Security checks across malware telemetry and agentic risk
Overview
This is a prompt-building skill with no executable code, install scripts, credential handling, network activity, or persistence.
Installers should understand this skill changes how the agent conducts prompt-building conversations. It is low-risk as packaged, but users should still independently verify outputs for high-stakes areas such as legal, financial, or accuracy-critical work.
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.