AI模型切换器

• Outcome: Pass
• Latest audit: May 29, 2026 at 8:30 PM UTC
• Version:
  1.0.0

Overview

This skill is a disclosed model-switching helper, but users should understand that cloud model modes may send future work to third-party providers.

Install only if you want help switching OpenClaw between local and cloud models. Use local mode for confidential prompts, code, or documents unless you trust the selected cloud provider, and review or delete the local switch log/stats files if model usage history is sensitive.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access
• Supply Chain: Unpinned Dependencies, External Script Fetching, Obfuscated Code
• Excessive Agency: Unrestricted Tool Access, Autonomous Decision Making, Scope Creep

Findings (3)

Missing User Warnings

Medium

• Confidence: 89% confidence
• Finding: The skill explicitly promotes automatic routing of user tasks to cloud models and also states that detailed usage and switching history are recorded, but it does not warn users that prompts, documents, or metadata may leave the local environment and be retained by third-party providers. In a model-switching skill, that omission is security-relevant because users may assume local/private handling while sensitive content is silently sent to external services.

Missing User Warnings

Low

• Confidence: 78% confidence
• Finding: The troubleshooting section tells users to check API keys but provides no guidance on secure storage, least-privilege handling, or the fact that those credentials enable transmission to external cloud services. While this is documentation-only and not direct secret exfiltration, it normalizes cloud credential use without basic security context, increasing the chance of unsafe configuration or accidental exposure.

Missing User Warnings

Medium

• Confidence: 92% confidence
• Finding: The README explicitly promotes automatic switching to cloud models for research, coding, and long-document tasks, but it does not clearly warn users that their prompts and potentially sensitive data may be transmitted off-device to third-party providers. In a tool framed as cost/performance optimization, this omission can cause users to unknowingly expose confidential content, especially because model selection is described as automatic and seamless.

VirusTotal

65/65 vendors flagged this skill as clean.