ClawHub
Back to skill

Security audit

Auto Model Switcher

Security checks for vulnerabilities and agentic risk

Overview

This skill is a disclosed model-routing guide with broad automatic switching behavior but no hidden code, persistence, file access, or destructive actions.

Install only if you are comfortable with the agent automatically choosing among your configured models/providers. Review your provider settings, costs, and data-handling expectations, and specify a model manually for sensitive tasks.

A.I.G

Vulnerability Patterns
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs

SkillSpector

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 (2)

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

The skill advertises very broad invocation criteria such as coding, analysis, writing, research, and long-context work, which overlap with a large share of normal user requests. This can cause the router to trigger frequently and take over model-selection decisions in situations where the user did not explicitly ask for switching, increasing the chance of unintended routing, privacy exposure to alternate providers, or degraded task handling.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
91% confidence
Finding

The task-detection logic relies on generic keywords like 'write', 'research', 'analyze', 'code', and common file extensions or commands, which are easy to match in benign conversation and may misclassify many requests. In an auto-switching skill, this creates unsafe ambiguity: attackers or accidental phrasing can steer requests to different models/providers, potentially affecting confidentiality, cost, reliability, or policy enforcement.

Content

No source excerpt is available for this finding.