多模型路由器

• Outcome: Pass
• Latest audit: May 29, 2026 at 8:43 PM UTC
• Version:
  1.0.0

Overview

This is a markdown-only model-selection guide with no executable install steps, but users should be mindful of cost and privacy if they implement routing across multiple AI providers.

Before installing, treat this as a routing guide rather than a ready-made secure router. If you implement it with real APIs, make parallel routing opt-in, avoid sending sensitive prompts to multiple providers unnecessarily, and monitor token costs and provider logging policies.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Trigger Abuse: Overly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access
• Supply Chain: Unpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (1)

Vague Triggers

Medium

• Confidence: 89% confidence
• Finding: The activation phrases are generic enough that the skill may trigger in unintended contexts, causing the router behavior to override or interfere with unrelated user requests. In a model-selection skill, accidental invocation can change which model is used, affecting cost, data flow, and response quality without explicit user intent.

VirusTotal

65/65 vendors flagged this skill as clean.