self-improving-agent

Organize learnings, errors, and feature requests into persistent markdown logs.

1|Updated Feb 13, 2018
One-click install
npx skills add https://github.com/g-lok/gs-dotfiles --skill self-improving-agent-g-lok
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: self-improving-agent
Source: https://github.com/g-lok/gs-dotfiles/tree/main/install.d/dotfiles/ai_skills/.agents/skills/self-improving-agent
Command: npx skills add https://github.com/g-lok/gs-dotfiles --skill self-improving-agent-g-lok

SYSTEM DOCUMENTATION & REQUIREMENTS

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

What problem does it solve?

Self-improvement skill for OpenClaw captures learnings, errors, and feature requests to enable continuous improvement across sessions and projects.

Core Features & Use Cases

  • Logs corrections, errors, and knowledge gaps to centralized markdown files for later analysis.
  • Supports promotion of recurring learnings to CLAUDE.md, AGENTS.md, or Copilot hooks to reuse across projects.
  • Provides a guided workflow to extract practical skills from learnings for reuse.

Quick Start

Create the .learnings directory at the project root and start logging with the prescribed format.

Frequently Asked Questions about self-improving-agent

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

FAQPage Schema
How do I log errors and learnings across OpenClaw workspace sessions?

Promote recurring learnings into project memory by running scripts that elevate persistent markdown log entries into CLAUDE.md, AGENTS.md, or SOUL/TOOLS.md, making knowledge gaps and corrections reusable across future sessions.

Can I automatically log errors and feature requests without manual entry?

Yes, you can automatically log errors and feature requests by configuring optional hook integrations that capture corrections and knowledge gaps into centralized markdown files without manual input.

What is the best way to extract practical skills from error tracking logs?

The best way to extract practical skills from error tracking logs is using the guided workflow to analyze persistent markdown entries, identify recurring feature requests, and promote consistent learnings into project memory files.

Does the self-improving-agent skill work with sub-agents in OpenClaw?

Yes, the self-improving-agent skill works with sub-agents in OpenClaw by applying persistent markdown logs and frontmatter metadata across sub-agent sessions, ensuring corrections and learnings are captured consistently throughout the workspace.

When should I promote learning logs into CLAUDE.md or AGENTS.md files?

Promote learning logs into CLAUDE.md or AGENTS.md files when corrections or knowledge gaps become recurring, ensuring that validated learnings transition from session-specific records into permanent project memory for continuous improvement.