shibing624

shibing624 / skills

Public

Learn from Experience skill 是让 Agent 从经验中学习 -- 纠正过的错误不再重犯，好的方法自动沉淀，越用越懂你。原理是基于你跟Agent交互中的经验(踩坑、纠正、好方法)记下来、整理好、用起来，不再随会话结束而消失。

learn-from-experience memory self-improvement skills skills-sh
18
2
89% credibility
Found Apr 15, 2026 at 19 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
AI Summary

A repository of self-contained skill folders that AI agents load from markdown files to gain abilities like learning from past corrections and building persistent knowledge bases.

How It Works

1
🔍 Discover smarter AI helpers

You hear about a collection of simple add-ons that make your AI assistant learn and remember better over time.

2
📂 Pick a skill

Choose a helpful ability like 'learn from experience' to stop repeating mistakes or build a personal knowledge notebook.

3
📱 Add the skill to your AI

Copy the skill folder into your AI helper's special skills area, like a drawer for new tricks.

4
🗣️ Tell your AI to wake it up

Chat with your AI and say 'start using the learn-from-experience skill' – it sets up its own memory space right away.

5
✅ Use and teach it

As you work together, point out fixes or favorites, and your AI saves them to get smarter next time.

6
🔄 Watch it improve

Over chats and projects, your AI recalls lessons, avoids old errors, and matches your style perfectly.

🎉 Smarter companion forever

Your AI now grows wiser with every interaction, saving you time and frustration across all your sessions.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 19 to 18 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 skills?

Skills is a collection of drop-in Markdown-based modules that supercharge AI agents like Claude Code, OpenClaw, or Cursor by adding persistent capabilities. Drop a skill folder into your agent's skills directory, say "initialize learn-from-experience," and it starts logging corrections, preferences, and reflections to avoid repeating mistakes across sessions—think learn from experience cloud synced to your global config. It also includes an LLM wiki for building linked markdown knowledge bases, solving the "amnesiac agent" problem where good habits and fixes vanish per chat.

Why is it gaining traction?

Unlike one-off prompts or custom scripts, skills auto-accumulate user-confirmed patterns into hot, warm, and cold memory layers, with commands like "memory stats" or "show my patterns" for easy auditing. Multi-agent compatibility (Gemini CLI to Windsurf) and cross-session sync to files like CLAUDE.md make it a low-effort upgrade for learn github workflows or ci cd pipelines. Devs dig the "set it and forget it" hook: agents get smarter without you scripting learn github actions every time.

Who should use this?

Coding teams on Claude Code or Codex CLI tired of reteaching agents your coding style or project rules. DevOps folks building learn github copilot setups for persistent ci cd devops pipelines. Anyone iterating on agentic tools who wants learn experience design without vanishing context, like frontend devs enforcing component patterns or backend users banning repeated API pitfalls.

Verdict

Grab it if you're deep in agent tooling—solid docs and Apache license make it easy to try, especially for learn from experience meaning in daily grinds. With 18 stars and 0.9% credibility, it's early-stage raw; test on non-critical workflows first as adoption lags behind the clever persistence premise.

(198 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.