ClawHub
Back to skill

Security audit

chromecast-control-free

Security checks for vulnerabilities and agentic risk

Overview

This skill mostly matches Chromecast control, but it grants command/write capability and contains overbroad, inconsistent instructions that could lead to LAN device control or unrelated file/API actions without clear user confirmation.

Review before installing. Use this only when you want the agent to control Chromecast-compatible devices on your local network, and prefer giving an explicit device name or IP. Avoid sharing credentials, do not use it on public or shared networks, and be aware the skill text includes unrelated API/file-processing boilerplate that should not be treated as permission for broader actions.

SkillSpector

Vulnerability Patterns
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • 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)

Intent-Code Divergence

Medium
Confidence
81% confidence
Finding
The skill is presented as a local Chromecast-control tool, yet its feature list broadly claims file processing, API integration, and command execution capabilities. In an agent environment with read/write/exec permissions, this kind of overbroad capability description can cause the agent to perform actions outside the intended trust boundary, increasing the chance of unsafe file access, unintended external requests, or risky command use.

Vague Triggers

Medium
Confidence
88% confidence
Finding
An overly broad activation trigger can cause the skill to be invoked for loosely related requests, including contexts where users did not intend network scanning or device control. Because this skill has exec capability and can discover/control LAN devices, accidental activation meaningfully raises the risk of unintended actions against local systems.

Missing User Warnings

Medium
Confidence
90% confidence
Finding
The documentation instructs the agent to scan the local network and control devices without clearly requiring user confirmation or warning about impact on nearby devices. In this skill context, those actions directly affect real LAN assets and media endpoints, so missing consent and safety boundaries makes misuse or accidental disruption more likely.

Static analysis

No suspicious patterns detected.