genshijin

To compress Japanese AI replies and save tokens, this Skill enforces an ultra-compressed communication style by removing honorific cushions, hedging, and redundancy while preserving technical accuracy across multi-turn code review or system explanation tasks.

You can control prompt style and compression by switching between 丁寧 (polite), 通常 (default), and 極限 (max compression) modes, ensuring the chosen Japanese text compression level persists across multiple conversation turns without style drift.

The best way to reduce context window usage is applying deterministic compression rules that strip verbose table-heavy formatting and filler phrases from Japanese technical replies, maintaining meaning while cutting wasted tokens for memory or summary-style tasks.

Compressed Japanese output stays consistent across multiple turns through multi-turn persistence, preventing style drift until you explicitly trigger a safe reset using designated exit commands to return to normal verbosity.

Japanese AI responses often lose accuracy during manual conciseness edits because standard compression removes critical context, whereas this Skill applies deterministic rules that remove only honorific cushions and redundancy while strictly preserving technical correctness.

You should avoid extreme token compression when Japanese technical communication requires formal politeness or detailed table-heavy formatting, instead utilizing the 丁寧 mode for business-friendly concise responses or issuing an explicit stop phrase to exit.