What it is

Russian profanity mode for AI agents, aimed at using a compact linguistic core (four roots) plus prefixes/suffixes to produce idiomatic output. The project is described as faster and more economical in root usage, with emphasis on generating natural-sounding speech rather than literal translations.

How it works

• Core idea: four roots plus affixes to form idiomatic expressions.
• Levels of calibration for agents: ordinary agent, lite, full (default), ultra. Switching level via commands like "/pohuy <level>". Each level affects how the internal object references are handled (via useMemo in examples).
• Activation commands: "(/pohuy)" to enable and commands to disable, e.g., "нормальный режим" to turn off.
• Internal emphasis on code quality: claims about clean commits/docs, and avoidance of security-focused shifters.

Getting started

The README provides installation steps for different toolchains:

• Claude Code plugin path: "```bash

  Claude Code — плагин

  claude plugin marketplace add smixs/pohuy && claude plugin install pohuy@pohuy

• Skills registry path: "```bash npx skills add smixs/pohuy

It also shows how to enable/disable and how to select levels, including a table with level descriptions and guidance on useMemo for reducing re-render churn.

Recent releases

• There are no releases listed in the repository metadata or releases section.

Traction

• Stars: 25 (no 7d or 1d data provided in FACTS, so no ongoing traction metrics beyond the current count are shown).

Caveats

• License: MIT
• Created: 2026-07-17
• Last push: 2026-07-18
• Language: not specified (language field is "-") and repository description is in Russian; README is Russian with some English phrases.

What it is not

• The README contains placeholder sections (e.g., "???” in How it works) and marketing-like phrasing not backed by external references.