可信 — 风险评分 5/100
上次扫描:22 小时前 重新扫描
5 /100
humanizer-zh
去除文本中的 AI 生成痕迹,让文字更自然、更有人味
Purely instructional text-processing skill with no executable code, scripts, or dependencies — all declared capabilities are appropriate for the stated text-editing purpose.
技能名称humanizer-zh
分析耗时24.7s
引擎pi
可以安装
Skill is safe to use as-is. No security concerns identified.
资源类型声明权限推断权限状态证据
文件系统 READ READ ✓ 一致 SKILL.md: allowed-tools includes Read; references ~/.openclaw/workspace/skills/n…
文件系统 WRITE WRITE ✓ 一致 SKILL.md: allowed-tools includes Write and Edit
网络访问 NONE NONE SKILL.md line 292 references an external Wikipedia URL only as documentation cit…
命令执行 NONE NONE No shell or subprocess usage detected
环境变量 NONE NONE No environment variable access found
技能调用 NONE NONE No cross-skill invocation detected, only documentation reference to narrative-vo…
1 项发现
🔗
中危 外部 URL 外部 URL
https://en.wikipedia.org/wiki/Wikipedia:Signs_of_AI_writing
SKILL.md:292

目录结构

3 文件 · 12.0 KB · 317 行
Markdown 2f · 310L YAML 1f · 7L
├─ 📋 clawhub.yml YAML 7L · 314 B
├─ 📝 README.md Markdown 17L · 329 B
└─ 📝 SKILL.md Markdown 293L · 11.4 KB

安全亮点

✓ No executable scripts, code files, or binaries present
✓ Allowed tools (Read, Write, Edit, AskUserQuestion) are appropriate and minimal for a text-editing skill
✓ No credential harvesting, data exfiltration, or credential theft indicators
✓ No obfuscation, base64 payloads, or anti-analysis techniques
✓ No network egress or C2 communication
✓ No supply-chain risks (no dependencies, no package files)
✓ Skill is purely instructional markdown — no attack surface from code execution
✓ The external Wikipedia URL is only a static documentation citation, not a dynamically fetched resource