ClawHub
Back to skill

Security audit

wechat-md-publish

Security checks across malware telemetry and agentic risk

Overview

This is a disclosed WeChat publishing helper that needs account credentials and can publish content, but its sensitive behavior matches its stated purpose.

Install only if you are comfortable giving the skill WeChat Official Account API credentials and sending selected Markdown content plus referenced local images to WeChat. Prefer creating drafts first, use --publish only after explicit review, and protect ~/.wechat_publish/config.json and token_cache.json with restrictive permissions or a dedicated environment.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Rogue AgentSelf-Modification, Session Persistence
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (1)

Session Persistence

Medium
Category
Rogue Agent
Content
### 3. 创建配置文件

```bash
mkdir -p ~/.wechat_publish
cat > ~/.wechat_publish/config.json << 'EOF'
{
  "app_id": "你的AppID",
Confidence
82% confidence
Finding
The setup instructions store long-lived WeChat AppID/AppSecret credentials in plaintext under ~/.wechat_publish/config.json, creating persistent local secret storage without any permissions hardening or secret-manager guidance. If the host is multi-user, backed up broadly, or later compromised, those credentials can be recovered and used to access the associated WeChat publishing account.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.