ClawHub
Back to skill

Security audit

Smart Model Switching

Security checks across malware telemetry and agentic risk

Overview

This skill is a model-routing guide that helps choose between Claude models and shows no evidence of hidden execution, data collection, persistence, or privileged access.

Install this if you want lightweight guidance for routing tasks to cheaper or stronger Claude models. For high-stakes workflows, review or tighten the routing heuristics so sensitive, production-critical, legal, medical, financial, or security tasks consistently use an appropriate higher-capability model.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (1)

Vague Triggers

Medium
Confidence
88% confidence
Finding
The routing rules use subjective thresholds such as 'if a human would need more than 30 seconds of focused thinking' and broad categories like 'most user conversations,' which can cause inconsistent or unnecessary escalation to higher-tier models. In practice, this can increase cost exposure, widen access to stronger model capabilities for tasks that do not need them, and create unpredictable behavior across agents or automation.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.