humanizer-zh
PassAudited by ClawScan on May 1, 2026.
Overview
This is a no-code text-editing skill that is coherent with its stated purpose, but users should review file edits and avoid using it to misrepresent AI-written content.
This skill appears safe for normal text editing. Before installing, verify the GitHub source if using the README’s install command. When using it, only point it at files you intend to edit, review the resulting changes, and avoid using it to conceal AI authorship where disclosure is expected.
Findings (3)
Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.
If asked to process a file, the skill may read and edit that file’s contents.
These permissions allow the agent to read and modify files, which fits the text-editing purpose but still affects user documents.
allowed-tools: - Read - Write - Edit - AskUserQuestion
Use it on intended documents only, keep backups for important files, and review diffs before accepting changes.
Installing from the wrong or changed remote repository could install content different from what was reviewed here.
The README recommends installing from a remote GitHub URL. The reviewed package contains only markdown instructions, but remote installation should still be verified by the user.
npx skills add https://github.com/op7418/Humanizer-zh.git
Verify the repository URL and contents before installing, especially if using npx or git clone.
In contexts that require disclosure of AI assistance, using this skill could make the text appear more human-authored than it is.
The skill openly aims to make AI-written text sound more human. That is the disclosed purpose, but it can affect reader trust if used to hide AI authorship.
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。
Use it to improve clarity and style, and disclose AI assistance when required by policy, school, workplace, or publication rules.