Shmayro

Shmayro / singularity-claude

Public

Self-evolving skill engine for Claude Code — skills that create, score, repair, and harden themselves through recursive improvement loops

ai-agents automation claude-ai claude-code developer-tools
10
1
100% credibility
Found Mar 23, 2026 at 10 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Shell
AI Summary

A plugin for Claude Code that enables AI skills to autonomously create, score, repair, and mature through usage feedback loops with local data storage.

How It Works

1
🚀 Discover self-improving skills

You learn about a helpful add-on for your AI assistant that makes custom skills get smarter and better over time on their own.

2
📥 Add it to your AI chats

You easily install the tool into your AI assistant so it starts working in your everyday conversations.

3
💡 Notice a repeated task

While using your AI, it spots a task you do often and suggests making a custom skill to handle it perfectly.

4
⚙️ Build your first skill

You guide the creation of a new skill step by step, tailored exactly to what you need it for.

5
🔧 Use the skill on a job

You try the skill on a real task, and it helps complete your work smoothly.

6
📊 Score how well it did

Right after, you rate the skill's performance in simple categories like accuracy and thoroughness.

7
🔄 See it fix itself

If the score is low, the tool automatically analyzes what went wrong and improves the skill for next time.

✅ Celebrate perfect skills

Over repeated uses, your skills mature into reliable helpers you can trust completely, saving you time forever.

Sign up to see the full architecture

6 more

Sign Up Free

Star Growth

See how this repo grew from 10 to 10 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 singularity-claude?

Singularity-claude is a self-evolving skill engine for Claude Code, where skills create, score, repair, and harden themselves through recursive improvement loops. It fixes static skills that break on edge cases or degrade over time by automating gap detection, 5-dimension scoring (correctness, completeness, etc.), auto-repairs, and crystallization into immutable versions. Delivered as a Shell-based Claude Code plugin with commands like `/singularity-create`, `/singularity-score`, and `/singularity-dashboard`, it stores everything locally in JSON.

Why is it gaining traction?

This self-evolving AI agent stands out by running zero-dependency loops natively in Claude Code—no Docker or external services—while tracking maturity from draft to hardened via telemetry and git tags. Developers hook into auto-repair for low scores and gap detection for new patterns, turning one-off skills into production-grade ones. As an awesome self-evolving agent on GitHub, it delivers Claude code singularity without the overhead of tools like SmythOS.

Who should use this?

Claude Code users building repetitive tasks like API proxies or form generators, tired of manual skill fixes. AI experimenters tweaking self-evolving LLMs for local coding workflows. Teams enforcing skill quality in Anthropic CLI sessions without constant reviews.

Verdict

Experimental with 10 stars and 1.0% credibility score, but solid docs and quick install make it worth a spin for Claude Code enthusiasts. Adopt for personal tinkering; hold for team use until more runs prove the loops.

(178 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.