中文 EN

扫描报告

扫描新文件

✓ 可信 — 风险评分 0/100
上次扫描：1 天前 重新扫描
0 /100
jcvd
Authorization gatekeeper for OpenClaw agents. Scoped grants, time-bound permissions, skill scanning, prompt injection detection, and full audit trail.
Jean-Claw Van Damme is a legitimate authorization gatekeeper skill that performs security scanning and audit logging. No malicious behavior detected.
技能名称jcvd
分析耗时25.0s
引擎pi
✓
可以安装
This skill is safe to install. It functions as documented: a pure markdown authorization framework with helper scripts for scanning and audit export.
资源类型声明权限推断权限状态证据
文件系统 NONE READ ✓ 一致 SKILL.md (implicit): reads target skill files during scan
命令执行 NONE WRITE ✓ 一致 scan-skill.sh, audit-export.sh: grep/cp operations scoped to target skill direct…
2 项发现
🔗
中危 外部 URL 外部 URL
https://agenticpoa.com
README.md:9
🔗
中危 外部 URL 外部 URL
https://snyk.io/blog/clawhub-malicious-google-skill-openclaw-malware/
README.md:13

目录结构

5 文件 · 27.0 KB · 801 行
Markdown 2f · 431L Shell 2f · 305L JSON 1f · 65L
├─ 🔧 audit-export.sh Shell 68L · 2.1 KB
├─ 📋 policy.json JSON 65L · 1.5 KB
├─ 📝 README.md Markdown 178L · 7.1 KB
├─ 🔧 scan-skill.sh Shell 237L · 6.3 KB
└─ 📝 SKILL.md Markdown 253L · 10.0 KB

安全亮点

✓ Pure markdown skill -- no compiled code, fully readable
✓ No external dependencies or third-party packages
✓ No obfuscation or encoded payloads anywhere
✓ Helper scripts are documented in README.md architecture section
✓ Security-focused tool: prompt injection detection, skill scanning, audit logging
✓ No credential harvesting, data exfiltration, or network exfiltration behavior
✓ No base64-encoded payloads or dynamic code execution
✓ No references to sensitive paths (~/.ssh, ~/.aws, .env)
✓ scan-skill.sh correctly detects its own patterns as a self-referential security scanner
✓ Open source MIT license, author from known organization (agenticpoa)