Agent Prompt Patterns
Security checks across malware telemetry and agentic risk
Overview
This is a markdown-only reference skill for designing safer AI-agent operating rules, with no hidden code or automatic actions.
Safe to install as a reference guide. When applying its examples, keep approval gates for email, deletion, financial actions, access-control changes, public posting, and pushes; do not log full secrets, tokens, session data, or unnecessary credential fingerprints.
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
66/66 vendors flagged this skill as clean.