ClawHub
Back to skill

Security audit

Smart Router Publish

Security checks for vulnerabilities and agentic risk

Overview

This skill is a disclosed local model-routing helper, with some broad triggers users should be aware of but no evidence of hidden, destructive, or data-exfiltrating behavior.

Install this if you want an agent to consult a local router for model-tier decisions. Expect broad routing-related phrases to activate it, and expect local Ollama setup or model downloads if the router backend is actually installed.

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 (3)

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The trigger list contains generic phrases such as 'switch to', 'routing', 'what model', 'which model', and 'speed up' that are likely to appear in normal conversation outside explicit skill invocation. This can cause unintended activation of the router skill and make the agent follow the skill's routing instructions in contexts where the user did not actually request model selection, increasing the chance of inappropriate behavior or prompt-scope interference.

Content

No source excerpt is available for this finding.

Shadow Command Trigger

Medium
Category
Trigger Abuse
Confidence
91% confidence
Finding

The trigger phrase 'save cost' can collide with built-in or reserved command patterns beginning with 'save', creating ambiguity about whether the user intended a native command or this skill. Such shadowing can divert command handling to the skill unexpectedly, causing unintended activation and interfering with trusted command behavior.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
88% confidence
Finding

The manifest includes Chinese trigger phrases alongside English ones, which establishes language-specific behavior but does not explain how language selection is determined or whether users can opt in or out. Under the policy, locale or language behavior should be user-selectable or explicitly justified when constrained.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.