ToyBridge

Security checks across malware telemetry and agentic risk

Overview

This skill openly controls a local BLE toy, but it gives agents direct physical control commands without clear confirmation, consent, or safety limits.

Review before installing. Use only with a device and ToyBridge server you control, verify the stop command first, and require your agent to ask before starting or changing vibration. Specify intensity and duration explicitly, and inspect the external ToyBridge server separately.

SkillSpector

By NVIDIA

Vulnerability Patterns

Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
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

Findings (1)

Missing User Warnings

High

Confidence: 95% confidence
Finding: The skill clearly issues local HTTP commands that can actuate a physical device, but it does not present an explicit safety warning or consent boundary up front. Because this is physical actuation of intimate hardware, missing warnings materially increases the risk of unsafe, surprising, or coercive use when the agent interprets ambiguous requests.

VirusTotal

66/66 vendors flagged this skill as clean.

View on VirusTotal