ClawHub

China Dev Environment

Security checks across malware telemetry and agentic risk

Overview

This skill is a visible setup guide for China-friendly developer mirrors and proxies, but users should review its global and destructive commands before running them.

Install only if you want an agent to provide China-focused development setup advice. Before running its examples, prefer project-local config where possible, confirm any sudo Docker daemon edits or global Git/npm/pip/Go settings, and back up lockfiles before deleting dependency folders or caches.

SkillSpector

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

Missing User Warnings

Medium
Confidence
93% confidence
Finding
The bootstrap workflow persistently changes user and system configuration, including npm, pip, git, Go, and Docker settings, without an explicit warning that these modifications are long-lived and affect future sessions. In an agent context, this is risky because a user asking for troubleshooting or setup advice may not expect global state changes, and Docker daemon changes require privileged writes to /etc/docker and service restarts.

Missing User Warnings

Medium
Confidence
91% confidence
Finding
The dependency-fix workflow includes destructive commands like removing node_modules and package-lock.json and purging package caches without first warning about loss of cached artifacts, lockfile state, or reproducibility metadata. In an automated agent setting, this can disrupt builds or silently alter dependency resolution, especially when deleting lockfiles causes fresh dependency selection.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal