A specialized logging and feedback loop system that enables AI coding agents to learn from errors, user corrections, and discovered best practices.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install self-improvement-ai
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install self-improvement-ai using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is Self-Improvement Agent Skill?
The Self-Improvement skill is a framework designed to give AI agents a long-term memory of their own performance and mistakes. By systematically capturing learnings, errors, and feature requests into structured markdown files, agents can avoid repeating past failures and refine their behavior over time. It is a core component of the Openclaw Skills ecosystem, enabling agents to operate more autonomously by building a persistent knowledge base within any project workspace.
This skill bridges the gap between transient chat sessions and permanent project documentation. It provides a standardized way to promote local learnings into project-wide conventions, ensuring that every interaction contributes to the overall intelligence of the development environment. Whether used for error tracking or best practice discovery, it transforms the agent from a stateless tool into a self-optimizing collaborator.
Self-Improvement Agent Skill Use Cases
- Logging unexpected command or API failures to prevent recurrence in future tasks.
- Capturing specific user corrections when an agent provides outdated or incorrect information.
- Documenting non-obvious project conventions that are discovered during the development lifecycle.
- Tracking requested capabilities that require future implementation or skill development.
- Promoting recurring patterns into permanent project memory files like CLAUDE.md or AGENTS.md.
How Self-Improvement Agent Skill Works
- The agent identifies a trigger such as a failed bash command, a direct user correction, or a discovered optimization.
- A structured entry is generated using a standardized ID format (e.g., LRN-YYYYMMDD-XXX) and appended to the appropriate file in the .learnings directory.
- The agent utilizes metadata such as priority, area, and pattern-keys to categorize the entry for easy retrieval.
- During periodic reviews or session starts, the agent scans existing learnings to apply previously discovered context to the current task.
- High-value or recurring items are promoted to global project memory or extracted into entirely new Openclaw Skills.
Self-Improvement Agent Skill Setup
To get started with this skill using the recommended platform for Openclaw Skills, run the following command:
clawdhub install self-improving-agent
For manual installation into your workspace:
git clone https://github.com/peterskoett/self-improving-agent.git ~/.openclaw/skills/self-improving-agent
mkdir -p ~/.openclaw/workspace/.learnings
Optional: Enable automatic reminders and error detection by activating the built-in hooks:
openclaw hooks enable self-improvement
Self-Improvement Agent Skill Data Schema & Taxonomy
The skill organizes data within a .learnings directory using specific markdown files:
| File | Purpose | Key Metadata |
|---|---|---|
| LEARNINGS.md | Corrections and best practices | Priority, Area, Pattern-Key |
| ERRORS.md | Command and tool failures | Exit Codes, Stack Traces, Context |
| FEATURE_REQUESTS.md | Desired agent capabilities | Complexity, Frequency, Status |
Each entry includes a unique ID, timestamps, and status tags (pending, resolved, promoted) to track the lifecycle of the improvement.
Self-Improvement Agent Skill Advanced Features
- Skill Extraction: Automatically generate new standalone Openclaw Skills from verified, high-value learnings using helper scripts.
- Recurring Pattern Detection: Tracks the frequency of specific issues across different tasks to prioritize systemic fixes.
- Inter-Session Communication: Tools to send learnings between active sessions or read the history of other sub-agents.
- Promotion Workflow: A formal process for moving distilled knowledge into project-level configuration files for permanent agent context.
- Hook Integration: Support for UserPromptSubmit and PostToolUse hooks to automate the evaluation of learnings without manual intervention.
SKILL.md
Loading
Related Openclaw Skills
Web Fetcher
A sophisticated web scraping skill that converts complex web pages, blogs, and documentation into AI-ready Markdown format.
Reddit Competitor Analysis API Skill
A specialized skill for extracting structured Reddit posts and comments to perform competitor research and brand sentiment analysis without hallucinations or CAPTCHA issues.
Openclaw Memorize
A lightweight memory management tool for OpenClaw agents to store and retrieve persistent information across multiple sessions.
Art Blocks Token Metadata Retriever
A high-performance tool for retrieving comprehensive on-chain metadata, media assets, and project context for specific Art Blocks tokens.
LLM Chat Scraper
A comprehensive scraping utility to extract structured conversation data and citations from leading AI chat platforms via the Scrapeless API.
9527 GitHub Trending
A zero-cost automation tool that fetches daily GitHub Trending projects and sends customized notifications to Telegram, DingTalk, or WeChat.
Featured*
