sstklen

sstklen / yes.md

Public

PUA says NO. YES says YES. — AI governance skill for Claude Code. Safety gates, evidence rules, anti-slack detection. 🌐 EN | 中文 | 日本語

36
3
100% credibility
Found Mar 12, 2026 at 28 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Shell
AI Summary

YES.md is a skill for AI coding assistants that enforces multiple layers of safety checks, evidence-based reasoning, and thorough debugging practices through automated reminders and guards.

How It Works

1
🔍 Discover YES.md

You hear about YES.md, a friendly guide that makes your AI coding helper safer, smarter, and more thorough without cutting corners.

2
📥 Add to your AI assistant

You quickly add YES.md to your AI coding tool using a marketplace or simple download, and pick your preferred language like English or Chinese.

3
💬 Start chatting with AI

You describe your coding problem to the AI, and YES.md activates automatically or with a simple /yes command.

4
🛡️ Safety kicks in

YES.md watches over the AI, blocking dangerous moves, reminding it to backup and check everything carefully before changes.

5
🔍 AI thinks deeply

The AI now gathers real evidence, checks for ripple effects, and avoids guesses, giving you solid fixes with proof.

6
✅ Enjoy reliable results

Your project gets fixed completely and safely, with no surprises, ready for real use.

🎉 Coding made easy

You confidently use your improved AI helper for all tasks, knowing it's thorough and protected.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 28 to 36 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is yes.md?

yes.md is a Claude Code skill that enforces rigorous AI governance during coding sessions, using safety gates, evidence rules, and anti-slack detection to prevent lazy habits like guessing or deflecting. It solves the unreliability of AI assistants by layering machine-enforced shell hooks with self-correction prompts, ensuring backups before edits, ripple checks after fixes, and verified deploys. Developers get a more thorough, production-safe Claude that pairs well with PUA for persistence.

Why is it gaining traction?

Unlike single-layer prompts or PUA's focus on persistence, yes.md stacks six enforcement layers—including unbreakable shell hooks that block destructive commands and mandate health checks post-deploy—making AI behavior stick without user nagging. The anti-slack detection catches seven common dodges like empty questions or surface fixes, forcing evidence-based action. Developers try it for the "YES says YES" hook: automatic triggers on failures or guesses, plus multilingual skills in English, Chinese, and Japanese.

Who should use this?

Backend engineers debugging multi-service production issues with Claude Code, where unverified claims or missed ripples cause outages. DevOps teams handling deploys, needing gates against rm -rf disasters or untested pushes. AI power users frustrated by Claude's advice-without-action in complex repos, especially those pairing it with PUA or systematic-debugging.

Verdict

Grab it if you're deep in Claude Code workflows—installs easily via marketplace or git clone, with solid docs and battle-tested rules. Low 19 stars and 1.0% credibility score flag early maturity and sparse tests, so test in non-prod first.

(178 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.