self-improvement
Captures learnings, errors, and corrections to enable continuous improvement. Use when: (1) A command or operation fails unexpectedly, (2) User corrects Claude ('No, that's wrong...', 'Actually...'), (3) User requests a capability that doesn't exist, (4) An external API or tool fails, (5) Claude realizes its knowledge is outdated or incorrect, (6) A better approach is discovered for a recurring task. Also review learnings before major tasks.
Install via CLI (Recommended)
clawhub install openclaw/skills/skills/biocrfhkust-cloud/self-improving-agent-2What This Skill Does
The self-improvement skill for OpenClaw is a framework designed to transform every failure, user correction, or knowledge gap into a permanent asset. By maintaining a structured .learnings/ directory, the agent builds a cumulative intelligence layer that prevents recurring errors and solidifies best practices. It serves as an automated feedback loop that captures transient interaction data and promotes critical insights into workspace-level documentation like SOUL.md, AGENTS.md, and TOOLS.md.
Installation
Installation is managed via the OpenClaw ecosystem. For the most seamless experience, utilize the package manager: clawdhub install self-improving-agent. Alternatively, you can perform a manual installation by cloning the source repository directly into your ~/.openclaw/skills/ directory. Once installed, ensure the workspace structure is initialized by running mkdir -p ~/.openclaw/workspace/.learnings to allow the agent to begin logging activities immediately.
Use Cases
This skill is essential for long-term project stability. Use it when an external API call fails, providing the stack trace to ERRORS.md. Trigger it when a user corrects a specific logic flow, allowing the agent to update LEARNINGS.md with a correction category to avoid future mistakes. Furthermore, use it to track missing features by logging requests to FEATURE_REQUESTS.md, or to document structural workflow improvements that should be promoted to the AGENTS.md file for team-wide adoption.
Example Prompts
- "The last command failed with a timeout error; please log this in the ERRORS file along with the current network context."
- "Actually, you should use the legacy API endpoint instead of the new one for this specific task; please update your learnings with this correction."
- "I have discovered a more efficient way to handle environment variable injection; please document this as a best_practice in the learnings file and update the TOOLS.md guide."
Tips & Limitations
To maximize the utility of this skill, conduct regular reviews of your .learnings/ directory before initiating major refactors or new features. Note that while this skill automates documentation, the actual promotion of content to high-level files like SOUL.md should be periodically audited by the human user to ensure alignment with project goals. Avoid bloating the logs with trivial information; focus on patterns that repeat across multiple sessions to ensure the agent remains performant and focused on high-impact optimizations.
Metadata
Not sure this is the right skill?
Describe what you want to build — we'll match you to the best skill from 16,000+ options.
Find the right skillPaste this into your clawhub.json to enable this plugin.
{
"plugins": {
"official-biocrfhkust-cloud-self-improving-agent-2": {
"enabled": true,
"auto_update": true
}
}
}Tags(AI)
Flags: file-write, file-read