Alerting
Security checks across malware telemetry and agentic risk
Overview
This is a simple alerting workflow guide with no executable code or requests for sensitive access.
Safe to install as an operational guidance skill. Review any alerting or on-call policy changes before applying them in production, but the skill itself does not add code, persistence, credentials, or privileged 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
63/63 vendors flagged this skill as clean.