oasis

Security checks across malware telemetry and agentic risk

Overview

This is a disclosed Oasis 3D-world integration skill; its world-editing powers are expected for the purpose, though users should confirm destructive edits.

Install this only if you want an agent to operate inside your Oasis world. Before clearing a world, deleting assets, or removing many objects, ask the agent to summarize the affected scope and wait for explicit confirmation. Also be aware that optional Forge or sculptor features may use API keys or LLM calls configured separately on the Oasis host.

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: 90% confidence
Finding: The skill explicitly exposes destructive capabilities such as `remove_object` and `clear_world` but does not instruct the agent to obtain explicit user confirmation before performing irreversible or broad world modifications. In an agent setting, this increases the risk of accidental deletion, griefing, or irreversible workspace damage from ambiguous prompts, misinterpretation, or prompt injection routed through normal tool use.

VirusTotal

66/66 vendors flagged this skill as clean.

View on VirusTotal