Security audit
Auto Skill 20260404 063632
Security checks across malware telemetry and agentic risk
Overview
This skill is a lightweight API-debugging workflow guide and does not include executable code, hidden installation behavior, or evidence of data theft.
Before installing, treat this as an API troubleshooting prompt template rather than a runnable tool. Do not paste full API keys or private credentials into the workflow; share redacted formats and endpoint details unless you intentionally want the agent to inspect a live configuration.
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
VirusTotal findings are pending for this skill version.
Static analysis
No suspicious patterns detected.