ClawHub
Back to skill

Security audit

TG Cam Skill

Security checks across malware telemetry and agentic risk

Overview

This camera skill appears purpose-built, but it needs review because it can access private camera data and trigger snapshots with broad invocation wording and persistent credential storage guidance.

Install only if you intend to let the agent access this camera service. Before using it, confirm where TIVS_CLI_ID and TIVS_API_KEY will be stored, protect or rotate those credentials as needed, and use explicit camera/device wording when requesting snapshots or event images.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • 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
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (2)

Vague Triggers

Medium
Confidence
89% confidence
Finding
The skill advertises broad natural-language triggers such as viewing devices, checking what happened today, or checking whether a device is online, which overlap with common household conversation and can cause unintended invocation. In a camera-control skill, accidental activation is more sensitive than usual because it may expose device lists, event history, images, or trigger a fresh snapshot capture without the user realizing they invoked surveillance-related functionality.

Missing User Warnings

Medium
Confidence
95% confidence
Finding
The skill description and summary do not prominently warn that the snapshot feature performs an active device-side capture operation rather than merely retrieving an already-available image. Because this is a camera skill, users may interpret '看看当前画面' or '拍一张' as passive viewing, while the API actually sends a command to the device, creating privacy, consent, and surprise-activation risks.

VirusTotal

66/66 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.