ClawHub
Back to skill

Security audit

Home Assistant Integration

Security checks across malware telemetry and agentic risk

Overview

This Home Assistant skill is coherent, but it gives an agent broad physical smart-home control with overly broad triggers and limited built-in safety scoping.

Review before installing. Use a dedicated least-privilege Home Assistant account or token, avoid granting access to locks, alarms, garage doors, cameras, and notification services unless intentionally needed, and require explicit user confirmation for state-changing or privacy-sensitive commands. Prefer a CA certificate over disabling SSL verification.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • 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
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (4)

Vague Triggers

Medium
Confidence
91% confidence
Finding
The skill advertises very broad natural-language trigger phrases for powerful home-control actions without defining confirmation requirements, authorization boundaries, or exclusion rules. In a smart-home context, ambiguous phrases can cause unintended actuation of lights, heating, automations, or other devices, especially when routed through chat/voice channels.

Missing User Warnings

Medium
Confidence
88% confidence
Finding
The markdown promotes camera access, notifications, automations, and physical device control but does not warn about privacy exposure or real-world safety consequences. In a home-assistant skill, these capabilities can reveal occupancy patterns, expose camera snapshots, or change environmental controls, so missing warnings materially increases misuse risk.

Vague Triggers

High
Confidence
98% confidence
Finding
The trigger list is extremely broad and includes common words like 'light', 'camera', 'door', 'lock', 'play', 'pause', 'alert', and 'HA'. In a natural-language agent, this can cause accidental invocation of a skill that is capable of controlling physical devices, including locks, climate, automations, and notifications, creating unintended smart-home actions.

Missing User Warnings

Medium
Confidence
92% confidence
Finding
The skill description emphasizes convenience and breadth of control but does not clearly warn that it can perform safety- and security-relevant actions like lock/door control, automation triggering, and outbound notifications. In this context, understated capability disclosure increases the chance that users enable or invoke the skill without understanding the real-world consequences of mistakes or abuse.

VirusTotal

63/63 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.