Humanizer Zh Cn
v1.0.0去除文本中的 AI 生成痕迹。适用于编辑或审阅文本，使其听起来更自然、更像人类书写。 基于维基百科的"AI 写作特征"综合指南。检测并修复以下模式：夸大的象征意义、 宣传性语言、以 -ing 结尾的肤浅分析、模糊的归因、破折号过度使用、三段式法则、 AI 词汇、否定式排比、过多的连接性短语。
⭐ 0· 376·3 current·3 all-time
by聿歆@v585
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
OpenClaw
Benign
high confidencePurpose & Capability
Name and description match the SKILL.md and README: the skill is an instruction-only text-editing helper to remove AI-writing patterns. It requests no credentials, no binaries, and no install-time behavior beyond what is typical for a text editing skill.
Instruction Scope
SKILL.md instructs the agent to identify and rewrite AI-like patterns and to preserve tone/meaning. It lists allowed tools (Read/Write/Edit/AskUserQuestion) and shows examples of processing files (e.g., article.md). Allowing read/write access is coherent for a text-editing skill, but users should understand that providing files/text gives the skill access to that content for processing.
Install Mechanism
The skill is instruction-only and has no formal install spec in the registry. README suggests installing via npx or git clone from a GitHub repo. That is typical, but installing from an external repo executes user-side operations (network fetch/clone) and users should verify the repository before running install commands.
Credentials
No environment variables, credentials, or config paths are requested. This is proportional for a text transformation skill.
Persistence & Privilege
always:false and standard autonomous invocation settings are used. The skill does not request persistent privileges or modify other skills or system settings.
Scan Findings in Context
[no-findings] expected: The static regex scanner found nothing — expected because the package is instruction-only (SKILL.md/README) and contains no executable code files to analyze.
Assessment
This skill appears internally consistent and focused on rewriting text to remove AI-style patterns. Before installing or using it: (1) Verify the source repository referenced in README (the registry lists no homepage and the README points to a GitHub URL) to ensure it matches the published package and is trustworthy. (2) Be cautious when processing sensitive or confidential text — the skill will read and rewrite whatever you give it. (3) When installing via npx/git, review the remote repo contents locally before running any install commands. (4) Test the skill on non-sensitive samples to confirm output quality and that it preserves meaning/tone as you expect.Like a lobster shell, security has layers — review code before you run it.
latestvk973j17r99y36nsc1hvhepkezd83ayvy
License
MIT-0
Free to use, modify, and redistribute. No attribution required.