An AI-driven framework for capturing errors, corrections, and discoveries into markdown-based project memory for continuous agent evolution.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install self-improving-agent-1-0-0
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-improving-agent-1-0-0 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 Skill?
The Self-Improvement Skill is a specialized protocol designed to help AI agents learn from their environment and user interactions. By systematically logging unexpected command failures, user-provided corrections, and newly discovered best practices into a local .learnings directory, this skill ensures that Openclaw Skills become more efficient over time. It provides a structured way to transform temporary conversation context into persistent technical debt tracking and architectural knowledge.
Self-Improvement Skill Use Cases
- Automatically logging terminal command failures or API timeouts for later debugging.
- Documenting user corrections when an agent makes an assumption that contradicts project conventions.
- Tracking missing features or capability requests identified during natural language interactions.
- Building a long-term knowledge base that can be promoted to project-level instruction files.
How Self-Improvement Skill Works
- The agent identifies a trigger such as an error, a user correction, or a knowledge gap during a task.
- A structured markdown entry is created in the .learnings directory using a unique ID format like LRN-YYYYMMDD-XXX.
- The agent populates the entry with critical metadata, including priority levels, the affected area of the codebase, and suggested fixes.
- During periodic reviews, pending items are resolved or promoted to permanent memory files like CLAUDE.md to guide future Openclaw Skills behavior.
Self-Improvement Skill Setup
To start using this skill in your repository, create the required directory structure from your project root:
mkdir -p .learnings
Once the directory exists, you can populate it with template files for LEARNINGS.md, ERRORS.md, and FEATURE_REQUESTS.md to allow Openclaw Skills to begin logging improvements immediately.
Self-Improvement Skill Data Schema & Taxonomy
The skill maintains a structured taxonomy using markdown files to categorize different types of feedback. Each file uses a specific header format for easy parsing and filtering.
| File Name | Logged Data Type | Key Metadata Fields |
|---|---|---|
| LEARNINGS.md | Best practices, corrections, and knowledge gaps | Priority, Status, Area, Source |
| ERRORS.md | Command failures, stack traces, and environment logs | Reproducible, Context, Suggested Fix |
| FEATURE_REQUESTS.md | Capability requests and implementation ideas | Complexity, Frequency, User Context |
All entries utilize a standardized status lifecycle: pending, in_progress, resolved, or promoted.
Self-Improvement Skill Advanced Features
- Pattern detection through grep-based searches to identify recurring issues across the development lifecycle.
- Promotion workflow that distills verbose learnings into concise, actionable rules for global project memory.
- Metadata-rich logging that supports filtering by codebase area such as frontend, backend, or infra.
- Collaborative learning support when the .learnings directory is tracked via version control for team-wide Openclaw Skills enhancement.
SKILL.md
Loading
Related Openclaw Skills
Frontend Design Skill
A specialized design-focused skill for creating high-end, production-ready frontend interfaces that escape the generic look of standard AI-generated code.
clscli
A powerful command-line interface for querying, filtering, and analyzing Tencent Cloud Cloud Log Service (CLS) data.
Cognary Tasks
A terminal-based task management skill that enables AI agents to perform full CRUD operations on Cognary task lists using a structured CLI.
Amap (Gaode Maps)
A powerful geospatial skill providing AI agents with real-time weather data, POI searching, and driving route planning via the Amap Web Service API.
Free Mobile SMS
Enables your AI agent to send SMS notifications directly to your mobile phone via the Free Mobile API.
SAP Journal Auditor
An automated auditing tool for SAP FI/CO journal exports that identifies financial anomalies, duplicates, and compliance risks.
Featured*
