SkillFoxx

Type a name, a task or a service

humanizer-ru (chat paste hygiene)

humanizer-ru

A CLI, MCP server and skill that find and safely strip chatbot copy-paste artifacts in Russian text and check facts after editing

CLIMedium riskRussian stack

vladimir-human/humanizer-ruAuthor: Vladimir-Human

Install

pip install humanizer-ru
humanizer-markers --scan input.txt

For English input add --language en.

This is third-party code. Review the repository files before installing.

What it does

The project catches traces of pasting from chat interfaces: citation markers, utm tails, invisible characters and hidden markup. humanizer-markers reports findings with positions and class, humanizer-clean strips supported artifacts while keeping a backup and leaving code, URLs and frontmatter alone. humanizer-facts compares numbers, dates, names, quotes and negations before and after an edit, and humanizer-report produces a report. The same checks are available to agents through an MCP server and a text skill; the tool gives no authorship verdicts.

Who it is for. For editors, teachers and teams who need a verifiable gate for chatbot paste artifacts.

Good fit when

  • Text still has markers like contentReference or utm_source after copying
  • You need to confirm editing did not lose numbers or names
  • You want a CI check for pasted content in markdown docs

Not a fit when

  • You want to determine whether a text was written by AI
  • You need a full stylistic rewrite

Example request

Check the article for ChatGPT copy traces, remove them and compare facts before and after

Limitations

It finds paste artifacts, not machine style itself: rewritten or smooth text without artifacts is not detected. Do not run polish on Markdown without --preserve-markup. The English profile covers only artifacts and facts.

How to disable. Run pip uninstall humanizer-ru and remove the humanizer-ru block from the MCP client config.

Security check

  • The clean command with --in-place overwrites files, keeping the original as .bak

README in short

The README explains who needs the tool and gives a short find, strip and verify path via pip. The full workflow covers finding, safe cleaning with a backup, diff, fact comparison and reporting. It includes an MCP config via uvx, a matrix of verified capabilities and an explicit list of what the project does not do. Marker classes, a browser demo and same-named projects are described separately. MIT licensed.

FAQ

How is it different from ilyautov/humanizer-ru?

It is a separate project with the same name. It focuses on paste artifacts and fact checks rather than stylistic editing.

What does exit code 1 mean?

Markers were found. 0 means no traces, 2 means the input could not be read.

Editors’ pick

Microsoft's utility and MCP server that turn PDF, Word, Excel, PowerPoint, HTML and other files into Markdown for models

MCP serverMedium risk184K
Editors’ pick

Anthropic's official skill examples: docx, pdf, pptx and xlsx documents, design, MCP builder and the Agent Skills spec

SkillMedium risk176.3K
Editors’ pick

A skill that builds editable PPTX decks from PDF, DOCX, URLs and Markdown with native shapes, charts and animations

SkillMedium risk54.4K
Editors’ pick

Anthropic's legal plugin suite: contracts, privacy, employment, IP, litigation, AI governance and legal education

PluginMedium risk9.4K