Security audit

Rag

Security checks across malware telemetry and agentic risk

Overview

This is a documentation-only RAG guidance skill whose security-relevant content is disclosed and aligned with its stated purpose.

Installers should understand this is guidance, not executable code. If using its production-log advice, avoid copying raw user queries or retrieved chunks into evaluation datasets unless you have approval, redaction, access controls, and retention rules in place.

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

Findings (1)

Missing User Warnings

Medium

Confidence: 93% confidence
Finding: The guidance recommends collecting production queries, retrieved chunks, and user feedback for dataset creation, but it omits any privacy, consent, retention, redaction, or access-control considerations. In a RAG system, production queries and feedback commonly contain personal, confidential, or proprietary data, so copying them into evaluation datasets can create unnecessary data exposure and compliance risk.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

Detected: suspicious.prompt_injection_instructions

Prompt-injection style instruction pattern detected.

Warn

Code: suspicious.prompt_injection_instructions
Location: security.md:41