skills.rest
zai-org
zai-orgOFFICIAL
Explore all skills by this author →
Website𝕏@Zai_orgGitHub

self-improvement

Capture AI session learnings, errors, and corrections into a .learnings repository.

588|41|Updated May 13, 2026
Search Tags:#automation#logging#learning#memory#errors#promotion#openclaw
One-click install
npx skills add https://github.com/zai-org/Synapse --skill self-improvement-zai-org
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: self-improvement
Source: https://github.com/zai-org/Synapse/tree/main/.setup/skills/clawhub/skills/self-improving-agent
Command: npx skills add https://github.com/zai-org/Synapse --skill self-improvement-zai-org

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) components.

What problem does it solve?

Automates capturing learnings, errors, and corrections during AI sessions to build a reusable knowledge base and improve future performance.

Core Features & Use Cases

  • Automatic logging targets: Logs to .learnings/LEARNINGS.md, ERRORS.md, and FEATURE_REQUESTS.md as appropriate.
  • Promotions to project memory: Facilitate promotion to CLAUDE.md, AGENTS.md, or .github/copilot-instructions.md when recurring or broadly applicable.
  • OpenClaw and hook integration: Hooks to trigger after prompts or command failures; usable with OpenClaw to automatically remind.
  • Extraction workflow: Extract learnings into reusable skills using scripts like extract-skill.sh.

Quick Start

Enable the self-improvement hook to capture learnings after each task and log them to the .learnings directory.

Frequently Asked Questions about self-improvement

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I automatically log AI session errors and corrections to a knowledge base?▼

You can automatically log errors and corrections by enabling the self-improvement hook, which captures tool failures and user feedback directly into centralized .learnings/ERRORS.md and LEARNINGS.md files.

What is the best way to capture recurring AI agent mistakes into project memory?▼

Capturing recurring mistakes into project memory involves logging corrections to .learnings/LEARNINGS.md, then promoting broadly applicable patterns into CLAUDE.md or AGENTS.md when they emerge.

Can I use OpenClaw hooks to track tool failures during coding sessions?▼

Yes, OpenClaw hooks can trigger after command failures and in-session prompts to automatically remind users to log issues and track errors into the .learnings directory.

How do I extract captured learnings into reusable skills?▼

You extract learnings into reusable skills by running the helper script extract-skill.sh, which processes the centralized .learnings repository and converts documented patterns into functional skill components.

Does this approach work with GitHub Copilot instruction files?▼

Yes, it works with GitHub Copilot by promoting recurring corrections and learnings directly into the .github/copilot-instructions.md file to guide future AI sessions.

Related Skills

Explore More Software Engineering Skills→
diriendirien
19

agent-rules

Generate and verify AGENTS.md files with bash, jq, and git.

Community
Advanced
marcoamumarcoamu
1

self-improvement

Log learnings and errors to markdown files under .learnings and promote them to project-wide memory files.

Community
Advanced
thewoolleymanthewoolleyman

Hooks Automation

Coordinate Claude Flow hooks across swarm agents for automated development workflows.

Community
Advanced
ut0p1anut0p1an

learn

Capture knowledge topics into a centralized knowledge base with duplicate checking and link suggestions.

Community
Intermediate