ClawHub
Back to skill

Security audit

Aitun Tunnel

Security checks across malware telemetry and agentic risk

Overview

The skill does what it says by creating public tunnels, but it also documents exposing powerful non-web services like SSH, RDP, and databases without enough safety guardrails.

Install only if you intend to publish a local service to the internet. Use it for temporary, non-sensitive web previews or test webhooks, prefer pip/uv over remote shell installers, avoid --tcp-ports for SSH/RDP/databases unless you fully understand the risk, and stop the tunnel when finished.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (3)

Description-Behavior Mismatch

Medium
Confidence
95% confidence
Finding
The skill is presented as a tool for exposing local web pages and HTTP services, but its documented CLI supports forwarding arbitrary TCP ports such as SSH or databases. That materially expands the attack surface from web previewing to remote access and infrastructure exposure, which could let an agent unintentionally publish sensitive internal services to the internet.

Context-Inappropriate Capability

Medium
Confidence
93% confidence
Finding
Documenting SSH/RDP-style TCP forwarding is not aligned with the stated purpose of sharing local web pages, and it enables exposure of powerful administrative services. In an agent context, that mismatch increases the chance of unsafe use because a user may invoke a seemingly benign web-sharing skill to expose remote login or other sensitive endpoints.

Missing User Warnings

Medium
Confidence
97% confidence
Finding
The skill repeatedly instructs users to expose local services to the public internet but provides no substantive warning about authentication, sensitive data exposure, or unintended access. Even if the tunnel is encrypted, publishing local endpoints without safety guidance can lead to data leakage, unauthorized access, or accidental exposure of development/admin interfaces.

VirusTotal

63/63 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.