ClawHub

Cloud Architecture Diagrams

Security checks across malware telemetry and agentic risk

Overview

This diagramming skill is mostly coherent, but its renderer can fetch and embed SVGs from arbitrary user-provided URLs without clear opt-in or host limits.

Review before installing if you handle private architecture details or run specs from other people. Prefer bundled icons and avoid icon_url unless you trust the URL; the maintainer should make remote fetching opt-in, restrict it to trusted HTTPS hosts, and block localhost/private-network targets.

SkillSpector

By NVIDIA
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
90% confidence
Finding
The instructions say the build process may fetch icons and also document refresh/indexing workflows that download data from external sources, but they do not clearly require user notice or consent before network activity. In an agent setting, this can leak user-provided architecture details, service names, or spec contents to third-party hosts and violates the expectation that a diagramming skill with bundled offline assets operates locally.

Missing User Warnings

Medium
Confidence
94% confidence
Finding
The script performs outbound network requests to fetch SVG icons from user-influenced URLs (`node['icon_url']` or search-index URLs) using `urllib.request.urlopen` with no allowlist, prompt, or clear disclosure. In an agent/skill context, this creates SSRF and unintended data egress risk: a crafted spec can cause connections to arbitrary hosts, including internal services or sensitive metadata endpoints, and leak that network access through timing/errors or cached content.

VirusTotal

63/63 vendors flagged this skill as clean.

View on VirusTotal