vasilievyakov

vasilievyakov / miracle-infrastructure

Public

17 skills for Claude Code: memory, decision-making, research, security, and business workflows. Zero dependencies.

11
0
100% credibility
Found Feb 26, 2026 at 10 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Python
AI Summary

A zero-dependency collection of Markdown prompts and rules that add memory, decision-making, research, business, and productivity capabilities to an AI coding assistant.

How It Works

1
🔍 Discover the toolkit

You hear about Miracle Infrastructure, a simple set of tools that make your AI coding helper smarter and more reliable.

2
📥 Get and set up

Download the starter files and run the quick setup to add helpful abilities like memory and thinking to your AI helper.

3
🧠 Prepare the memory

Open your AI helper and give it a simple command to organize its memory for your projects, so it never forgets important details.

4
💭 Start your work session

Chat with your AI about a project, and it automatically loads past notes, decisions, and lessons without you reminding it.

5
🤝 Get smart help

Ask for advice from virtual experts, verify facts, or extract tasks from notes, feeling like you have a team supporting you.

6
💾 Save your progress

At the end, tell your AI to save the session, capturing key insights and next steps for tomorrow.

🎉 Work flows smoothly

Your AI now remembers everything, makes better suggestions, and saves you hours of re-explaining, making projects fun and efficient.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 10 to 11 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 miracle-infrastructure?

Miracle infrastructure equips Claude Code users with 17 Markdown-based skills across seven packs—memory, thinking, research, business, content, productivity, and meta—that add persistent memory, decision boards, web research with verification, and workflow automation like turning call transcripts into proposals and prototypes. It solves Claude's session amnesia and solo decision-making pitfalls by letting you invoke commands like /session-save, /directors, or /research directly in chats, all with zero dependencies beyond a bash install to your Claude skills folder. Built in Python with pure Markdown prompts, it runs offline after a two-minute setup.

Why is it gaining traction?

Unlike MCP-heavy Claude skills github alternatives like memory-bank or claudemem, miracle infrastructure demands no databases, Docker, or servers—just readable Markdown files in ~/.claude/skills that version with git and work across Claude desktop, CLI, or web. Developers hook on token-efficient progressive disclosure (search 100 observations for 4k tokens, not 15k), auto-capture rules, and real usage stats from 1,169 sessions proving it prevents re-debating decisions like JWT vs sessions. It's the github skills ai creator's minimal viable nervous system for Claude code skills.

Who should use this?

Solo full-stack devs or indie hackers using Claude code cli for daily coding who hate re-explaining projects each session. AI workflow builders needing claude skills web research, security audits (/security review), or business tools like /proposal from transcripts. Teams experimenting with skills claude github agents without infra overhead.

Verdict

Grab it if you're deep in Claude Code—docs are thorough, MIT-licensed, and installer is idiot-proof, but with 10 stars and 1.0% credibility, treat as an early alpha for custom tweaks. Solid for personal use; scale cautiously until more battle tests surface.

(198 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.