mxchip-smart-control

• Outcome: Review
• Latest audit: May 29, 2026 at 9:27 PM UTC
• Version:
  1.0.2

Overview

This is a coherent smart-home control skill, but it gives agents direct authority to change real devices and trigger scenes without documented confirmation safeguards.

Review before installing. Only use this with an MXCHIP account and devices you intend the agent to control, store MXCHIP_OAUTH_TOKEN as a secret, and configure your agent workflow to list devices or scenes first and require explicit confirmation before any device control, AC change, or scene trigger.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Excessive Agency: Unrestricted Tool Access, Autonomous Decision Making, Scope Creep
• Trigger Abuse: Overly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access

Findings (4)

Missing User Warnings

Medium

• Confidence: 89% confidence
• Finding: The README documents tools that can turn devices on/off, change AC settings, and trigger scenes, but it does not clearly warn that these actions affect real physical devices and may activate home automations. In a smart-home control skill, missing safety disclosure increases the chance of unintended or unsafe actions by users or downstream agents, especially where scenes may unlock, power, or otherwise alter the home environment.

Missing User Warnings

Medium

• Confidence: 86% confidence
• Finding: The skill describes impactful operations such as turning devices on or off, changing AC settings, and triggering scenes, but it does not foreground user-safety warnings or confirmation expectations before those actions. In a smart-home context, undocumented or unconfirmed execution can cause physical-world effects such as unlocking routines, powering appliances, or changing environmental conditions.

Missing User Warnings

Medium

• Confidence: 85% confidence
• Finding: The documentation provides direct examples for controlling devices and triggering scenes that can cause real-world physical actions, but it does not warn about unintended activation, occupancy/safety implications, or the need for explicit user confirmation before execution. In a smart-home control skill, this omission increases the risk that an agent or integrator will treat these actions like harmless API calls and perform unsafe or unauthorized operations.

Missing User Warnings

Medium

• Confidence: 91% confidence
• Finding: The client exposes a direct `trigger_scene(scene_id)` capability that can execute smart-home automations without any built-in confirmation, policy check, or safety gating. In a smart-home context, scenes may unlock doors, change HVAC state, or power devices, so a caller mistake, prompt injection, or unauthorized tool invocation could cause real-world effects.

VirusTotal

66/66 vendors flagged this skill as clean.