Security audit

TG Cam Claw

Security checks for vulnerabilities and agentic risk

Overview

This is a privacy-sensitive camera skill, but its camera access and credential use are disclosed, scoped to the stated purpose, and not backed by hidden code or unrelated behavior.

Install only if you are authorized to access the linked cameras. Treat TIVS_API_KEY like a password, understand it may be saved locally in openclaw.json, and avoid using broad requests like “what happened today” unless you intend the agent to check camera events.

SkillSpector

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 (2)

Vague Triggers

Medium

Confidence: 90% confidence
Finding: The skill advertises broad trigger phrases like “今天发生了什么” and “看看有哪些设备”, which can overlap with ordinary conversation and cause unintended invocation of a surveillance-capable skill. In this context, accidental activation is more sensitive than usual because the skill can retrieve camera snapshots, event histories, and device status, exposing private household monitoring data without clear user intent.

Missing User Warnings

Medium

Confidence: 92% confidence
Finding: The skill description and usage guidance do not warn users that it accesses highly sensitive surveillance data, including current images, event images, and behavioral history. Because the skill handles monitoring content from cameras in private spaces, the absence of explicit privacy notice, consent expectations, and caution around third-party visibility materially increases the risk of inappropriate disclosure or use.

Static analysis

No suspicious patterns detected.