Rate Limiting
Security checks across malware telemetry and agentic risk
Overview
This is a documentation-only rate-limiting guidance skill with no code execution, credential access, persistence, or hidden behavior.
This appears safe to install as a guidance skill. Users should still adapt the rate-limiting recommendations to their own architecture and product policies, but the artifacts do not request access to accounts, files, credentials, or runtime systems.
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.