ClawHub
Back to skill

Security audit

https://openrouter.ezsite.ai - Unified LLM API — one key for Claude, GPT, Gemini. 2x credits, auto-failover, OpenAI-compatible.

Security checks for vulnerabilities and agentic risk

Overview

This is a documentation-only skill for configuring a third-party LLM router, with no hidden code or persistence, but users should understand their prompts and EZRouter API key go through that service.

Before installing or using this skill, treat EZRouter as a third-party intermediary: use a dedicated EZRouter key, avoid sending sensitive code or private data unless you accept that service's privacy and security posture, and review its terms before routing production model traffic through it.

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
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Findings (2)

Missing User Warnings

Medium
Confidence
93% confidence
Finding
The skill instructs users to route API keys and prompts to a third-party proxy service instead of the original model providers, but it does not disclose the security, privacy, logging, retention, or trust implications of doing so. This can expose sensitive prompts, outputs, and credentials to an intermediary that users may incorrectly assume is equivalent to the native provider endpoints.

External Transmission

Medium
Category
Data Exfiltration
Content
### Claude — Messages API

```bash
curl https://openrouter.ezsite.ai/api/claude/v1/messages \
  -H "Authorization: Bearer your_cr_key_here" \
  -H "Content-Type: application/json" \
  -H "anthropic-version: 2023-06-01" \
Confidence
95% confidence
Finding
The provided curl example demonstrates sending an Authorization bearer token and user content to a non-native external endpoint, which normalizes external transmission of secrets and potentially sensitive prompt data. In the context of an agent skill, this is risky because users may copy-paste the example into real environments without understanding that they are delegating trust to a third-party service.

Static analysis

No suspicious patterns detected.