Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
git clone --depth 1 https://github.com/ilyautov/humanizer-ruWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/plugins/ilyautov/humanizer-ru/gemini-extension)<a href="https://agentmods.dev/plugins/ilyautov/humanizer-ru/gemini-extension"><img src="https://agentmods.dev/badge/plugins/ilyautov/humanizer-ru/gemini-extension/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/plugins/ilyautov/humanizer-ru/gemini-extension"><img src="https://agentmods.dev/badge/plugins/ilyautov/humanizer-ru/gemini-extension.svg" alt="Reviewed on agentmods" width="80" height="20"></a>Grade A, and why
humanizer-ru scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 6d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
What it actually says
{
"name": "humanizer-ru",
"version": "3.32.0",
"description": "Редактура русского текста: убирает канцелярит, кальки и нейросетевые штампы, не меняя факты и голос автора. 67 паттернов, 21 жёсткий запрет, 3 режима, сканер с баллом чистоты 0-100 и факт-замок. Russian text quality editor: removes bureaucratese, calques and AI-style clichés without changing facts or the author's voice. 67 patterns, 21 hard bans, 3 modes (edit / audit / targeted fix), deterministic scanner with a 0-100 cleanliness score and a fact-lock."
}
What ships with it
30 files beside gemini-extension.json in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
- assets/icon-128.png 4.8 KB
- assets/icon-256.png 9.8 KB
- assets/icon-512.png 9.8 KB
- assets/readme-banner.jpg 480 KB
- scripts/build_extension.py 7.4 KB runs code
- scripts/build_release_zip.py 1.8 KB runs code
- scripts/build_sitemap.py 3.6 KB runs code
- scripts/bump_release.py 19 KB runs code
- scripts/check_examples.py 21 KB runs code
- scripts/check_live_distribution.py 13 KB runs code
- scripts/check_workshop_build.py 3.2 KB runs code
- scripts/export_web_rules.py 9.9 KB runs code
- scripts/install_smoke.py 6.9 KB runs code
- scripts/lint_skill.py 15 KB runs code
- scripts/ru_plural.py 5.6 KB runs code
- scripts/scan_gate.py 6.7 KB runs code
- scripts/self_scan.py 12 KB runs code
- scripts/test_facts_numbers.py 3.8 KB runs code
- scripts/test_gates.py 15 KB runs code
- scripts/test_install_smoke.py 1.4 KB runs code
- scripts/test_install_surface.py 1.2 KB runs code
- scripts/test_markers.py 54 KB runs code
- scripts/test_mcp_server.py 7.8 KB runs code
- scripts/test_mining.py 5.4 KB runs code
- scripts/test_modern_slop.py 6.4 KB runs code
- scripts/test_scan_gate.py 7.8 KB runs code
- scripts/test_scan.py 16 KB runs code
- scripts/test_score.py 23 KB runs code
- scripts/test_web_parity.py 16 KB runs code
- scripts/web_scan_cli.mjs 1.5 KB runs code
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 6d ago Changed ec2290f68fbb
- 9d ago Changed cde811135d72
- 13d ago Changed 87e53c1fc698
- 26d ago Changed a927cadb66d4
- 27d ago Changed 7907c91ec172
- 28d ago Changed 70c16d9570dd
- 1mo ago Changed 3b899599a2c6
- 1mo ago Changed 463e8a39d495
- 1mo ago Changed 7559123227e4
- 1mo ago First seen · 6 lines scan A b55133250ae1
humanizer-ru is a plugin published in the GitHub repository ilyautov/humanizer-ru (489 stars, last pushed 2d ago), licensed MIT. Its token cost is not measured: this kind of file is read by the harness, not the model. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-07.
Other plugins, from other repositories
humanize-korean
A Korean-language text editor that detects patterns commonly associated with AI-written or translation-like writing and rewrites them to sound more human.
im-not-ai
A Gemini CLI extension that revises Korean text written by AI to sound more like human writing. It detects over 40 common AI-writing patterns and rewrites them.
humanizer-ru
A plugin for reviewing Russian-language text for signs of machine-generated writing and, when explicitly requested, rewriting it to sound more natural.
humanizer-zh-tw
A Claude Code plugin for making Traditional Chinese writing sound more natural and less machine-generated.
patina
Detect and rewrite AI writing patterns in Korean, English, Chinese, and Japanese so text reads as if a human wrote it. Meaning-preservation (MPS) verified, audit-friendly. The root SKILL.md is loaded as the /patina skill.
humanizer-fr
Réécrit un texte français qui sonne IA pour qu'il se lise comme son auteur, sans changer ce qu'il dit.