ClawHub
Back to skill

Security audit

jf-open-pro-ai-outdoor

Security checks for vulnerabilities and agentic risk

Overview

This skill appears purpose-built for a JF outdoor security system, but it can make lasting security changes and handle device credentials without strong confirmation safeguards.

Review this skill carefully before installing. Use it only with an account intended to administer the target JF device, keep the app secret/JWT/device password protected, and require explicit confirmation before any delete, disable, credential-sync, or configuration-change action.

SkillSpector

Vulnerability Patterns
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • 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
Findings (4)

Lp3

Medium
Category
MCP Least Privilege
Confidence
86% confidence
Finding
The skill documentation directs use of environment variables, shell commands, and networked API calls, but no explicit permissions are declared. This creates a transparency and governance gap: users or platforms may authorize a skill without understanding it can access credentials and perform authenticated remote operations against security devices.

Tp4

High
Category
MCP Tool Poisoning
Confidence
91% confidence
Finding
The declared description emphasizes outdoor security features, but the documented behavior also includes higher-risk capabilities such as vehicle CRUD, credential/device-auth synchronization, service switching, and analytics/alarm-history access. This mismatch can mislead users and reviewers about the true scope of sensitive operations, increasing the chance of overbroad trust and unintended execution of actions affecting devices, data, and credentials.

Vague Triggers

Medium
Confidence
80% confidence
Finding
The trigger phrases are broad, generic operational terms such as '异常告警', '智能检测', and '车辆管理', which are likely to overlap with normal monitoring conversations. In a security-device skill with state-changing actions, accidental invocation can lead to unintended configuration changes or exposure of sensitive operational data.

Missing User Warnings

Medium
Confidence
88% confidence
Finding
The skill documents destructive and security-relevant actions—such as deleting vehicles, disabling services, and updating device behavior—without requiring explicit confirmation or warning users about consequences. In the context of outdoor security systems, such actions can degrade monitoring coverage, suppress alerts, or remove trusted-vehicle data, directly weakening physical security operations.

Static analysis

No suspicious patterns detected.