Lutron Caseta

Security checks across malware telemetry and agentic risk

Overview

This Lutron smart-home plugin does what it claims, but it can immediately control whole-home lighting and other devices without built-in confirmation safeguards.

Install only if you are comfortable giving an agent direct control over your Lutron bridge. Require explicit confirmation before whole-home actions like all-off, scene activation, shade movement, Smart Away changes, or Pico button taps, and treat exported bridge snapshots as sensitive home-layout data.

SkillSpector

By NVIDIA

Vulnerability Patterns

Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
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

Findings (3)

Missing User Warnings

Medium

Confidence: 95% confidence
Finding: The skill exposes high-impact physical control actions, including `lutron_all_off`, scene activation, cover control, and button tap simulation, without any built-in confirmation, warning, or policy gating for destructive or disruptive operations. In an agent setting, a mistaken model action, prompt injection from upstream content, or ambiguous user request could immediately turn off devices across a home or manipulate occupancy-related behavior, causing safety, availability, or privacy issues.

Missing User Warnings

Medium

Confidence: 92% confidence
Finding: The skill documents a whole-house `lutron_all_off` capability including a 'Kill every light in the house' workflow without requiring an explicit confirmation or warning before executing it. In a home-automation context, bulk shutdown actions can disrupt occupants, create safety issues in dark environments, or be triggered from ambiguous user requests, so the lack of a confirmation guard is a real safety/security weakness.

Missing User Warnings

Medium

Confidence: 86% confidence
Finding: The skill exposes `lutron_export` as a full JSON snapshot of areas, devices, scenes, occupancy groups, and buttons, but provides no privacy warning about the sensitivity of that data. In context, this export can reveal detailed home layout, device inventory, occupancy-related metadata, and automation structure, which materially increases privacy and recon risk if requested casually, stored insecurely, or shared with other tools.

VirusTotal

62/62 vendors flagged this plugin as clean.

View on VirusTotal