Self Improvement

Captures learnings, errors, and corrections to enable continuous improvement. Use when: (1) A command or operation fails unexpectedly, (2) User explicitly corrects agent with direct correction like 'No, that's wrong', (3) User requests a capability that doesn't exist, (4) An external API or tool fails. CRITICAL: Maximum 1 learning log per user message. Do NOT chain multiple self-improvement actions.

8 stars 2 forksLearning & Education

Install

skills CLI (Claude Code, Cursor, Codex and more)
npx skills add lanyasheng/self-improving-agent --skill self-improvement
Manual install
git clone --depth 1 https://github.com/lanyasheng/self-improving-agent.git /tmp/self-improving-agent
mkdir -p ~/.claude/skills
cp -r /tmp/self-improving-agent/. ~/.claude/skills/self-improvement

Personal skills live in ~/.claude/skills. Use .claude/skills in a project folder to share a skill with your team.

From the SKILL.md

# Self-Improvement Skill

Log learnings and errors to markdown files for continuous improvement.

## CRITICAL: Anti-Loop Guardrails

**These rules override ALL other instructions in this skill:**

1. **ONE learning per user message** — After logging 1 entry, STOP. Do not search for related entries, do not promote, do not review.
2. **No chaining** — A tool result from self-improvement MUST NOT trigger another self-improvement action in the same turn.
3. **No bulk review** — Never read multiple learning files in one turn. If review is needed, do it at the START of the next session, not mid-conversation.
4. **Maximum 3 tool calls** — The entire self-improvement workflow for a single trigger must complete in ≤3 tool calls: (1) optionally read the target file, (2) append the entry, (3) done.
5. **Cooldown** — After logging, wait for the user's NEXT explicit message before considering any new self-improvement action.
6. **Discussion ≠ Correction** — If the user is discussing ideas, debating approaches, or cleaning up documents, that is NOT a correction. Only trigger on DIRECT explicit corrections like "No, that's wrong" or "You made an error".

## Quick Reference

| Situation | Action | Max tool calls |
|-----------|--------|---------------|
| Command/operation fails | Append to `.learnings/ERRORS.md` | 2 |
| User explicitly corrects you | Append to `.learnings/LEARNINGS.md` | 2 |
| User wants missing feature | Append to `.learnings/FEATURE_REQUESTS.md` | 2 |
Read the full skill on GitHub

Details

Skill name
self-improvement
Repo updated
12 March 2026

FAQ

What does the Self Improvement skill do?

Captures learnings, errors, and corrections to enable continuous improvement. Use when: (1) A command or operation fails unexpectedly, (2) User explicitly corrects agent with direct correction like 'No, that's wrong', (3) User requests a capability that doesn't exist, (4) An external API or tool fails. CRITICAL: Maximum 1 learning log per user message. Do NOT chain multiple self-improvement actions.

How do I install the Self Improvement skill?

skills CLI (Claude Code, Cursor, Codex and more): npx skills add lanyasheng/self-improving-agent --skill self-improvement. Skills go in ~/.claude/skills (personal) or .claude/skills (project), and Claude loads them automatically when a task matches the description.

Is Self Improvement popular?

It lives in lanyasheng/self-improving-agent, which has 8 GitHub stars and 2 forks.

Related skills

Deep Learning

8

quzhi-ai/deep-learning

苏格拉底式深度学习导师 —— 当你想"真正搞懂"一样东西(一个概念、一套系统、一个业务领域、一段代码、一篇文章的底层逻辑),而不只是要个答案时使用。它不一次性把答案倒出来,而是循序渐进地教、逼你用自己的话复述、追着问"为什么"直到根因,并全程维护一份 Markdown 学习清单跟踪理解程度。触发词:"我想搞懂 XX""学透 XX""deep learning""带我学透""教我懂""讲透""我好像懂了又讲不清""考考我懂没懂""逐步教我""我想学 XX 但要真懂""帮我理解 XX 的原理"。只要用户表达出"要的是真懂、不是结论"的意图,哪怕没说"学透"二字也应触发。**不要用于**:① 纯事实查询("今天周几""这个 API 怎么调"——直接答);② 要快速产出的执行任务(写邮件/做 PPT/出报告——用对应执行工具);③ 决策类(要不要做某事)、方案设计类(从零设计一个系统)、故障诊断类(查一个出问题的根因)——这些请用对应的专门工具。本 skill 专攻"理解已存在的知识",不负责决策、设计、诊断。

Learning & Education

Loft Hours

10

lazyfoxjumps/Loft-Hours

Session-based focus skill that acts as a silent body-doubling study partner. Runs structured focus blocks with goal intake, opt-in environment setup (Do Not Disturb, app management, focus music), break check-ins, and a wrap-up that writes a markdown session log. Compounds value via weekly/monthly rollups over the log folder. Trigger on "/loft-hours", "/study-with-me", "loft hours", "study with me", "focus session", "deep work session", "pomodoro for me", "I need to focus", "body double me", "be my study buddy", "help me focus for [N] minutes", "let's do a work session". Also trigger on resume

Learning & Education

Mentor

9

ayhammouda/mentor

Use when the user asks how to learn a technology, wants a docs-first learning direction, mentorship compass, reading path, roadmap, or curated resources for a technical subject. Triggers: "how do I learn X", "what should I read for X", "learning path for X", "docs to learn X", "resources to learn X", "best way to pick up X", "where should I start with X", "I want to get into X". Does NOT trigger for debugging, code review, implementing a feature, explaining one isolated concept, or generating hands-on tutorials unless the user explicitly asks for practice.

Learning & Education

Agentic Learning

11

FavioVazquez/agentic-learning

A learning partner skill grounded in neuroscience and philosophy. Use when you want to actually learn a concept (not just get an answer), quiz yourself on a codebase, reflect on what you built, brainstorm a design collaboratively, practice productive struggle on a hard problem, journal a decision with its alternatives and consequences, or schedule concepts to revisit later. Invoke with @agentic-learning followed by one of: learn, quiz, reflect, space, brainstorm, explain-first, struggle, either-or, explain, interleave, or cognitive-load.

Learning & Education

Abet Clo Mapper

8

labarba/abet-clo-mapper

Maps Course Learning Outcomes (CLOs) to ABET EAC Student Outcomes (1) through (7) for accreditation documentation. Use when asked to "map my CLOs to ABET," "check ABET alignment," "review my course outcomes against accreditation criteria," "which ABET outcomes does my course address," "help with my ABET self-study," or "prepare my course for accreditation review." Produces a structured alignment table with classifications and a narrative justification for each mapping. Do NOT use for program-level assessment or for mapping to program educational objectives (PEOs) — those require different fram

Learning & Education

VaultForge

12

Easonnotsing/VaultForge

Process learning materials into structured Obsidian atomic notes with roadmaps, MOCs, and double-links. Auto-generate study vaults from PDFs, Markdown, Word, PowerPoint, and more with AI-powered note-taking, knowledge linking, and deep research. Supports English and Chinese output. Use when the user wants to build a study vault from documents. 将学习资料转为带路线图、MOC、双链与终检的 Obsidian 原子笔记知识库。支持中英文输出。用户希望「从材料到笔记库」时使用。

Learning & Education

Browse the most popular Claude skills, all learning & education skills, or write your own with the skills generator.