ClawHub

Network Tools

Security checks across malware telemetry and agentic risk

Overview

This is a disclosed network utility skill that makes user-directed web requests, diagnostics, and downloads, with privacy and overwrite cautions but no evidence of hidden or malicious behavior.

Install this only if you want an agent to make outbound network requests and save downloaded files through local tools. Do not pass tokens, cookies, private/internal URLs, or sensitive hostnames unless you intentionally want them sent to the destination or network diagnostic services, and review downloaded files before opening or executing them.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (4)

Vague Triggers

Medium
Confidence
90% confidence
Finding
The skill recommends use for broadly defined scenarios like any external data retrieval, API integration, file downloads, and network troubleshooting. In an agent setting, that kind of permissive activation guidance can cause the skill to be invoked for routine tasks without adequate checks on destination, data sensitivity, or user consent, increasing the chance of unintended outbound requests.

Missing User Warnings

Medium
Confidence
95% confidence
Finding
The examples encourage fetching URLs, sending custom headers, using proxies, downloading content, and querying third-party services, but they do not prominently warn that these actions transmit data to external systems. This is dangerous because agents or users may unknowingly send bearer tokens, request metadata, internal targets, or other sensitive information off-host.

Missing User Warnings

Medium
Confidence
92% confidence
Finding
The script can download arbitrary remote content and write it to local files, including auto-generated filenames derived from untrusted URLs, without any confirmation prompt, safe-directory restriction, or overwrite warning. In an agent context, this increases the risk of unintended file creation, disk pollution, or replacement/corruption of existing files if the tool is invoked on attacker-controlled URLs or paths.

Missing User Warnings

Medium
Confidence
88% confidence
Finding
The skill transmits user-provided URLs, hosts, and domains to remote servers and third-party services using curl, wget, httpie, DNS tools, whois, traceroute, and IP-check endpoints without any privacy notice or trust boundary. In an agent setting, this can leak sensitive internal URLs, hostnames, or investigation targets to external infrastructure, enabling privacy loss and possible SSRF-like misuse when pointed at internal resources.

VirusTotal

66/66 vendors flagged this skill as clean.

View on VirusTotal