A Yi Turns "Nonviolent Communication" into an Open-Source AI Skill

If human civilization in the future is taken over by the superintelligence Musk talks about, and all legal codes have to be rewritten, then Nonviolent Communication will most likely be the only communication protocol that doesn't need a single line of its logic deleted or changed. It is not at all a bestselling book that teaches you to raise your emotional intelligence; it is the only underlying operating system for modern people in the wasteland of language that can settle interests and needs completely without harming anyone.

And unfortunately, 99% of people who have read Nonviolent Communication, after closing the book, still haven't argued one fewer argument they should have argued, and haven't said one fewer impulsive thing they should have said.

Because everyone understands the principles, but in those few seconds when emotions run high, the human brain simply doesn't have time to perform this complex grammatical transformation.

So I simply took Marshall Rosenberg's entire Nonviolent Communication and turned it into an open-source and fully executable AI Skill.

It is not a lightweight reading note, nor is it the kind of hypocritical praise-group rhetoric that dodges and deflects everywhere.

It is a structured plug-in for Cursor, Claude Code, codex, Hermes, and various other Agents:

Just throw at it the most impulsive sentence, the sentence in which you most want to curse someone out, or the sentence in which you are most at a loss, and it can, at millisecond speed, decompose it into the underlying "observation—feeling—need—request," precisely translating toxin-filled attacking language into a top-tier communication solution that can solve the problem.

The underlying layer of this Skill uses a very clean "routing-type architecture": one main control entry point, paired with 5 hardcore reference modules loaded on demand, engineering and codifying all the most essential psychological mechanisms in the original book:

First, the four-element hard constraint engine: strictly locks down the four core boundaries of "observation vs. comment," "feeling vs. thought," "need vs. strategy," and "request vs. demand," and strictly forbids the AI from using condescending judgments to escalate conflict;

Second, empathic listening and emotional desensitization: specifically deals with "what should I reply when he says this," automatically piercing through the other party's hostile thorns to grasp the real lack behind them;

Third, anger transformation and conflict review: decomposes punitive anger carrying a desire to destroy into protective boundary confirmation;

Fourth, jackal-language anti-pattern cleansing: built in with a checklist for intercepting communication toxins, spotting moral judgment, coercion, and shifting of responsibility at a glance;

Fifth, fine-tuning of the author's original-voice style: using an extremely sincere and restrained giraffe context to output high-quality feedback and sincere gratitude.

Many people think communication depends on emotional intelligence, but in fact, in contemporary high-intensity workplaces and intimate relationships, communication is purely a matter of algorithms.

You don't need to force yourself to be a saint in a fit of rage; when writing work reports, making requests to the team, passing the buck and tearing into each other across departments, arguing late at night with a partner...

before pressing Enter to send out the gunpowder, let it run through the compiler for you. When psychology is installed into the IDE and the terminal, restraint is no longer self-repression that violates human nature, but the natural result that follows once the code runs through.

The open-source address has been fully made public, supports one-click git clone installation, and the link is organized on the first floor for you to get yourself.