AANA Guardrail Skill

Security checks across malware telemetry and agentic risk

Overview

This looks like a legitimate guardrail skill, but it writes potentially sensitive decision context to local JSON files and calls a helper script that is not included in the package.

Review before installing. Use it only if you already have and trust the referenced AANA CLI, and avoid placing secrets, payment details, or unnecessary personal/account data into event JSON files unless you have clear storage and deletion practices.

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: 87% confidence
Finding: The skill instructs the agent to create an event JSON containing user requests, candidate actions, and potentially private account, billing, payment, or personal data, but it provides no minimization, consent, retention, redaction, or disclosure guidance. In an agent workflow, this can normalize copying sensitive data into local files and secondary tooling, increasing the risk of unnecessary exposure, logging, or downstream mishandling.

VirusTotal

62/62 vendors flagged this skill as clean.

View on VirusTotal