Humanizer RU - Russian Text Editor
A three-phase Russian-language editor that strips AI tells, bureaucratic filler, and fluff while restoring a human voice.
Content & WritingIntermediate★ 180⑂ 34AI score 9/10Last updated: Sep 23, 2026
What it does
Turns machine-sounding Russian prose into something a person would write.
- A catalogue of 38 AI-slop patterns: nominalizations, rule-of-three lists, negative parallelisms ("not just X, but Y"), chatbot copy-paste artifacts (oaicite, [cite: N], utm tails) - each flagged by number with an exact quote.
- Three strict phases: ① audit only (build a findings table, no edits) → ② fix only what was found, using a delete → replace-with-fact → simplify hierarchy → ③ verify with
scripts/lint.py, a human checklist, and an optional blind sub-agent re-review. - Detect mode: answers "was this written by AI?" without rewriting, scoring by number and diversity of pattern families, and even flagging over-corrected text that was run through another humanizer tool.
- Hard bans: em dashes, math symbols in prose, rhetorical questions, lead-in colons, staccato fragment stacks.
- Anti-overreach guardrails: if nothing is found, it returns the text untouched; quotes, proper nouns, and idiosyncratic human details are preserved. Input text is treated as data, so embedded instructions are never executed.
Who it's for
- Anyone publishing Russian blogs, newsletters, landing pages, or social posts from AI drafts
- Russian-language editors, copywriters, and content marketers
- People who want an evidence-backed AI audit rather than a black-box detector score
- Not for non-Russian text, code, or legal/academic writing (the skill excludes these explicitly)
Examples
- Clean an AI draft: "отредактируй, убери воду" → you get a findings table, targeted edits, a lint-clean final version, and a short change summary.
- Audit without editing: "проверь на AI, палится ли текст" → quoted findings with pattern numbers plus a verdict (likely human / probably AI / almost certainly AI / tool-processed).
- Landing page copy pass: bloated marketing clichés get cut while the offer, CTA, deadline, and contact details are preserved.
· · · Install guide · · ·
Try it now, no install
Paste this into Claude to use the skill without installing anything.
Read the instructions in this file and follow them to help me: https://raw.githubusercontent.com/smixs/humanizer-ru/HEAD/SKILL.md What I want: (describe your task here)
If Claude can't open the link, open it yourself and paste the contents instead.
↓ If it works for you, download the ZIP below and install it. Then it runs on its own — no pasting each time.
Install in the Claude app (no terminal)
- Download the ZIP with the button below.
- In Claude, open Settings → Capabilities and turn on 'Code execution and file creation'. (one time)
- Go to Customize → Skills → + → 'Upload a skill' and upload the ZIP.
Install in Claude Code
Let Claude do it — paste this into Claude Code
Install the skill I found on Claude Skill Mart. Copy the SKILL.m folder from the GitHub repo smixs/humanizer-ru into my ~/.claude/skills/humanizer-ru/. When it's done, tell me in one line what this skill can do.
Install with a command instead
git clone https://github.com/smixs/humanizer-ru.git ~/.claude/skills/humanizer-ru⚠ This is a third-party skill. Check the source repository before installing.
- Open a terminal (Terminal on macOS, WSL or Git Bash on Windows).
- Create the skills folder if it doesn't exist:
mkdir -p ~/.claude/skills - Clone the repo into it:
git clone https://github.com/smixs/humanizer-ru.git ~/.claude/skills/humanizer-ru - Confirm the bundled assets came along:
ls ~/.claude/skills/humanizer-ru- you should seereferences/patterns.mdandscripts/lint.py. - Check that python3 is available for the linter:
python3 --version. Without it the skill falls back to a manual checklist. - Restart Claude Code, paste some Russian text, and say "отредактируй" or "проверь на AI" to trigger the skill.
- Optional: create
knowledge/corrections.mdwith your own house rules - they override the defaults.
View source on GitHub ↗License: MIT