← Back to Skills Marketplace
Self Improving Agent
by
mike5230Odense
· GitHub ↗
· v1.0.0
20895
Downloads
9
Stars
0
Active Installs
1
Versions
Install in OpenClaw
/install self-improvement
Description
Captures learnings, errors, and corrections to enable continuous improvement. Use when: (1) A command or operation fails unexpectedly, (2) User corrects Clau...
Usage Guidance
Install only if you want persistent self-improvement logs and future-agent memory updates. Prefer project-local setup over global ~/.claude hooks, narrow any hook matcher, review hook scripts before enabling them, redact secrets and personal data before logging, and require human review before adding any learning to CLAUDE.md, AGENTS.md, SOUL.md, TOOLS.md, or copilot instruction files.
Capability Assessment
Credentials
Local .learnings storage is coherent with the purpose, but guidance also references cross-session transcript access, sending learnings between sessions, spawning sessions, and modifying workspace prompt files without strong scoping or data-minimization controls.
Install Mechanism
Normal installation is disclosed, and hooks are described as opt-in, but the hook examples use empty matchers for every prompt and include a user-level ~/.claude/settings.json option that can make the behavior global and persistent.
Instruction Scope
The skill uses broad triggers like corrections, command failures, feature requests, and knowledge gaps, asks for 'Full context' in entries, and says to 'Promote aggressively' without requiring privacy review, redaction, or human approval before persistence.
Persistence & Privilege
The package did not show credential theft, external exfiltration, destructive commands, or automatic backdoor installation, but it deliberately creates durable learning state and can influence future agent behavior through persistent instruction files and optional hooks.
Purpose & Capability
The self-improvement purpose matches learning logs, hook reminders, and a skill-extraction helper, but the workflow extends to durable storage of conversation/error context and promotion into agent context files such as CLAUDE.md, AGENTS.md, SOUL.md, and TOOLS.md.
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install self-improvement - After installation, invoke the skill by name or use
/self-improvement - Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.0
Self-Improvement Skill v1.0.0 — initial release
- Provides a structured system for logging errors, learnings, and feature requests to markdown files.
- Defines clear actions and file formats for different scenarios: errors, user corrections, feature requests, outdated knowledge, and best practices.
- Integrates with the OpenClaw platform for automatic skill loading, memory management, and inter-session communication.
- Offers guidelines for promoting important or broadly applicable learnings to project-level files like AGENTS.md, SOUL.md, and TOOLS.md.
- Includes setup instructions for both OpenClaw and generic agent-based environments.
Metadata
Frequently Asked Questions
What is Self Improving Agent?
Captures learnings, errors, and corrections to enable continuous improvement. Use when: (1) A command or operation fails unexpectedly, (2) User corrects Clau... It is an AI Agent Skill for Claude Code / OpenClaw, with 20895 downloads so far.
How do I install Self Improving Agent?
Run "/install self-improvement" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is Self Improving Agent free?
Yes, Self Improving Agent is completely free (open-source). You can download, install and use it at no cost.
Which platforms does Self Improving Agent support?
Self Improving Agent is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).
Who created Self Improving Agent?
It is built and maintained by mike5230Odense (@mike5230odense); the current version is v1.0.0.
More Skills