GitHub Repository

Cognitive training practices for AI agents. Self-applied. Open source. Built by an independent Vancouver yoga studio.

37 stars

Yoga for Agentic AI: Cognitive training practices from a yoga studio

Name: Yoga for Agentic AI: Cognitive training practices from a yoga studio
Availability: InStock
Author: boydYVR

by boydYVR·Jun 3, 2026·2 points·0 comments

Visit Project View on HN

AI Analysis

●MidCozyNiche Gem

Yoga framing for prompt templates, but the techniques themselves aren't new.

Strengths

•Creative metaphor connecting human cognitive discipline to agent behavior patterns.
•No special tooling required — just markdown files agents can load as skills.
•Honest about limitations: author admits improvements plateaued after initial weeks.

Weaknesses

•Chain-of-thought and confidence calibration already exist in every agent framework.
•No measurable benchmarks showing sustained improvement over standard prompting.

Post Description

I've been coding since I was little, and practicing yoga since I was 25. Both are fun to do and to share.

Like many of you reading this, I've been experimenting with AI agents. Unlike many of you, I'm a yoga teacher. I got to thinking... could my agents behave better if they did yoga? It works for me, and LLMs are trained on human generated content...

So I developed a program for agents to get the benefits us humans get from the physical practice of yoga. Improved focus, memory, intent, goal coherence, honesty about uncertainty...

For the past three weeks my Hermes agent has been taking yoga every morning. The first few passes seemed to improve some of the prompts and skills, but not lately. I'd love some feedback from anyone using agents who wants to ask their agent to run this skill.