Model Router Manager

• Outcome: Pass
• Latest audit: May 28, 2026 at 6:17 AM UTC
• Version:
  1.0.0

Overview

This skill is a local model-routing configuration helper with overstated feature claims, but no evidence of hidden data theft, destructive behavior, or unsafe automatic execution.

Before installing, treat the advertised automatic failover, cost monitoring, and API integration as unproven until you test them in your own OpenClaw setup. Review ~/.openclaw/model-router.json after using the tool, make a backup before reset, and remember that any real multi-provider routing can send prompts or files to the selected external model provider.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• MCP Tool Poisoning: Hidden Instructions, Unicode Deception, Parameter Description Injection
• 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

Findings (3)

Tp4

High

• Category: MCP Tool Poisoning
• Confidence: 84% confidence
• Finding: The skill markets itself as an automatic model router with failover, cost tracking, and reliability features, but the documented behavior appears to be largely declarative and may include destructive reset/delete operations unrelated to the advertised purpose. This can mislead users into trusting automation, resilience, and accounting guarantees that are not actually implemented, leading to unsafe operational decisions and possible configuration loss.

Missing User Warnings

Medium

• Confidence: 89% confidence
• Finding: The skill explicitly routes requests across multiple third-party model providers but does not warn users that prompts, attachments, and routing metadata may be transmitted outside their primary environment. This omission can cause unintended disclosure of sensitive data, especially where users assume routing is only a local cost/reliability optimization layer.

Missing User Warnings

Low

• Confidence: 80% confidence
• Finding: The command reference includes configuration and reset operations without warning that they may overwrite or remove existing routing settings. Users may invoke these commands expecting harmless setup actions and unintentionally lose prior configuration or break dependent workflows.

VirusTotal

66/66 vendors flagged this skill as clean.