Humanize Chinese

PassAudited by ClawScan on May 1, 2026.

Overview

This appears to be a local Chinese text-rewriting tool with no evidence of data theft, but it is explicitly designed to lower AI-detection scores and has minor metadata/provenance inconsistencies.

Before installing, consider whether using a tool to reduce AIGC or AI-detection scores is allowed in your academic, workplace, or publishing context. If you use it, run it only on intended local files, save outputs to safe paths, and review the rewritten text for accuracy.

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.

Note

ASI09: Human-Agent Trust Exploitation

What this means

The tool may help make AI-assisted academic text appear less AI-generated, which could mislead evaluators if used against rules.

Why it was flagged

The README explicitly markets the tool for lowering academic AI-detection/AIGC scores, which is relevant to user trust, academic integrity, and policy compliance.

Skill content

用 ChatGPT 写了论文初稿，担心知网/维普查出来？这个工具就是干这个的。

Recommendation

Use it only where rewriting and AI-assistance disclosure are permitted, and verify that the final text, citations, and claims remain accurate.

Note

ASI02: Tool Misuse and Exploitation

What this means

If invoked on the wrong file or output path, it could read or overwrite text files the user did not intend to process.

Why it was flagged

The skill is allowed to run local commands and read/write files. This is expected for its CLI text-processing purpose, but it is still local execution and file mutation authority.

Skill content

allowed-tools:
  - Read
  - Write
  - Edit
  - exec
...
python scripts/humanize_cn.py text.txt -o clean.txt

Recommendation

Run it only on intended documents, choose safe output paths, and keep backups of important files.

Note

ASI04: Agentic Supply Chain Vulnerabilities

What this means

It may be harder to confirm exactly which release or source revision is being installed.

Why it was flagged

The registry metadata reports version 2.4.0 while package.json and SKILL.md present v2.1.0-era metadata, creating a minor release/provenance inconsistency.

Skill content

"version": "2.1.0"

Recommendation

If provenance matters, verify the package against the claimed repository and release before relying on it.