by 安全扫描
OpenClaw
安全
high confidenceThe skill's code, instructions, and packaging are internally consistent with its stated purpose (rewriting text to evade AI detectors); it requests no credentials or external installs and operates only on local files, but it explicitly enables detector evasion which may be unethical or violate policies.
评估建议
This skill is technically coherent: the scripts implement exactly what the README says and operate only on local files (no network calls or credentials). Before installing, consider: 1) Ethical/legal risk — the tool explicitly helps evade AI-detectors (could violate school/employer policies or service TOS). 2) Source verification — no homepage and the metadata ownerId mismatch is a minor red flag; prefer code from a known, trusted publisher. 3) Audit patterns.json and the scripts yourself (they'...详细分析 ▾
ℹ 用途与能力
The name/description (humanize AI text to bypass detectors) match the included scripts (detect.py, transform.py, compare.py) and patterns.json. The required surface (no env vars, no external binaries) is proportional. Minor packaging inconsistency: _meta.json.ownerId differs from the registry owner ID in the metadata — likely benign but worth verifying the publisher.
ℹ 指令范围
SKILL.md and the scripts instruct the agent to read/write local text files, scan for patterns, and rewrite content to reduce detector signals. The instructions do not request unrelated files, system credentials, or network endpoints. However, the explicit runtime objective is to bypass AI detectors — a clear misuse vector (academic dishonesty, terms-of-service evasion) even though it stays within the technical scope described.
✓ 安装机制
No install spec or external downloads. All code is packaged in the skill and operates locally; there are no remote fetches, archive extraction, or registry installs to review.
✓ 凭证需求
No environment variables, credentials, or config paths are required. The scripts only read local files (text and patterns.json) and write outputs; requested access is minimal and proportional.
✓ 持久化与权限
The skill does not request always:true and has no mechanism to persist or modify other skills or global agent settings. It runs on demand and only interacts with files the user provides.
安全有层次，运行前请审查代码。
运行时依赖
无特殊依赖
安装命令
点击复制官方npx clawhub@latest install humanize-ai-text-1-0-0-1
镜像加速npx clawhub@latest install humanize-ai-text-1-0-0-1 --registry https://cn.longxiaskill.com