Chinese Humanizer
PassAudited by VirusTotal on May 11, 2026.
Overview
Type: OpenClaw Skill Name: zh-humanizer Version: 1.0.0 The OpenClaw AgentSkills bundle 'zh-humanizer' is classified as benign. The `_meta.json` file contains standard metadata. The `SKILL.md` provides detailed instructions for an AI agent to 'humanize' Chinese text by removing AI-style writing traces. While the `allowed-tools` (Read, Write, Edit, Grep, Glob, AskUserQuestion) grant significant file system access, the instructions within `SKILL.md` are entirely focused on text transformation and do not contain any directives for data exfiltration, malicious execution, persistence, or any other harmful behavior. There is no evidence of prompt injection attempts designed to subvert the agent for malicious purposes; instead, the markdown is used to define the agent's persona and task for its stated purpose.
Findings (0)
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.
The skill may help polish text naturally, but it could also be used to make AI-generated text look human-authored.
The stated editing goal is to make text appear human-written rather than AI-generated. This is disclosed and central to the skill, but it can affect how downstream readers perceive authorship.
任务是把文本改写得像真人自然写成，而不是 AI 输出
Use it for legitimate editing and style improvement, and avoid using the output to falsely claim human authorship where disclosure is expected.
If used on files, the agent may change document contents.
The skill can read, search, write, and edit files. These tools are expected for a text-editing skill, but they can modify local documents if the agent applies them.
allowed-tools: Read Write Edit Grep Glob AskUserQuestion
Invoke it on specific intended text or files, and review edits before accepting them.