ClawHub
Back to skill

Security audit

每日安全巡检

Security checks across malware telemetry and agentic risk

Overview

This is a disclosed read-only OpenClaw security checklist that creates a local report and should be used only where report contents can be kept private.

Install only in OpenClaw workspaces where the agent is allowed to inspect security configuration and logs. Keep generated reports private, and only forward them to Telegram/Feishu or similar channels if those channels are trusted, because reports may identify where sensitive configuration issues exist even when secret values are redacted.

SkillSpector

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

Missing User Warnings

Low
Confidence
84% confidence
Finding
The template explicitly instructs the agent to save a generated report into a workspace path without mentioning user confirmation or a dry-run option. Even though the write target is predictable and low-risk, this still authorizes state-changing file output that could occur without the user's explicit consent.

Credential Access

High
Category
Privilege Escalation
Content
### 2. 防火墙
（提醒用户在本机确认已开启）

### 3. API 密钥在 .env
（✅ 正常 / ❌ 发现明文 + 位置与建议）

### 4. SOUL.md 安全监控规则
Confidence
92% confidence
Finding
.env

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.