← Back to Skills Marketplace
78
Downloads
0
Stars
0
Active Installs
1
Versions
Install in OpenClaw
/install self-improvement-v2
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
This skill appears to do what it says: surface reminders and log learnings to .learnings/, optionally install a small hook, and provide a helper to scaffold new skills. Before installing or enabling hooks: 1) Inspect the included scripts (activator.sh, error-detector.sh, extract-skill.sh) yourself — extract-skill.sh creates files under the current directory; 2) Only enable OpenClaw/Claude hooks if you trust the source, since hooks run with the agent's user permissions and can inject content into sessions; 3) Be cautious about enabling PostToolUse hooks globally (error-detector.sh parses tool output) if you don't want that reminder in every Bash tool run; and 4) If you want stricter visibility, run these scripts manually rather than wiring them into global hook settings. If you want additional assurance, provide the upstream repository URL you plan to clone so you (or I) can verify the remote source fingerprints and commit history.
Capability Analysis
Type: OpenClaw Skill
Name: self-improvement-v2
Version: 1.0.0
The 'self-improvement-v2' skill bundle is designed to create a feedback loop for AI agents by logging errors, corrections, and new insights into markdown files (e.g., `.learnings/LEARNINGS.md`). The bundle includes shell scripts (`activator.sh`, `error-detector.sh`, `extract-skill.sh`) and OpenClaw hooks that provide reminders and automate the creation of new skill scaffolds. The code includes appropriate safety checks, such as path validation and input sanitization in the extraction script to prevent directory traversal. All behaviors, including the promotion of learnings to workspace configuration files like `SOUL.md` or `AGENTS.md`, are transparently documented and align strictly with the stated purpose of continuous agent improvement.
Capability Tags
Capability Assessment
Purpose & Capability
Name, description, and included files (reminder hook, activator, error detector, extract helper, templates) align with a self-improvement/logging skill. Files placed under ~/.openclaw and .learnings, hook injection of a reminder, and a script to scaffold extracted skills are proportionate to the stated purpose.
Instruction Scope
Runtime instructions and hooks are narrowly scoped to emitting reminders, detecting error text in a tool output env var, and scaffolding markdown-based skills. They reference the CLAUDE_TOOL_OUTPUT environment variable (a platform-provided variable) and instruct copying hooks into ~/.openclaw/hooks and adding workspace files; those are expected for integration but are actions that grant file-system and session-context influence, so verify you want that behavior before enabling.
Install Mechanism
No install spec; repository/clone instructions point to a GitHub repo and included scripts are local. No downloads from untrusted URLs or archive extraction are present. The included scripts are readable and their behavior is straightforward (print reminders, scan an env var for error patterns, create a skill scaffold).
Credentials
The skill declares no required environment variables or credentials, which matches its purpose. Scripts do rely on a platform variable (CLAUDE_TOOL_OUTPUT) to detect errors; this is normal for hook-based integrations but is not listed in requires.env. The scripts can write files when run (extract-skill.sh scaffolds files under the current directory) — this is expected but worth awareness.
Persistence & Privilege
always is false and the skill is user-invocable; hooks are opt-in. The hook code injects a virtual bootstrap reminder (no permanent changes unless you copy files yourself). Enabling hooks gives the skill the ability to influence session bootstrap output and to run the provided scripts with the agent's user permissions — expected for this integration but should be enabled intentionally.
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install self-improvement-v2 - After installation, invoke the skill by name or use
/self-improvement-v2 - Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.0
Initial release of the self-improvement skill for OpenClaw and generic agents:
- Provides structured workflow for logging errors, learnings, and feature requests in markdown files for continuous improvement.
- Includes step-by-step setup guides for both OpenClaw and other agent environments.
- Defines clear templates and logging formats for learnings, errors, and feature requests.
- Outlines promotion process for broadly useful insights to project memory files (e.g., AGENTS.md, SOUL.md, TOOLS.md).
- Supports inter-session communication of learnings within the OpenClaw workspace.
- Offers optional startup hook for automatic reminders to review and log learnings.
Metadata
Frequently Asked Questions
What is Self Improvement?
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 78 downloads so far.
How do I install Self Improvement?
Run "/install self-improvement-v2" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is Self Improvement free?
Yes, Self Improvement is completely free, licensed under MIT-0. You can download, install and use it at no cost.
Which platforms does Self Improvement support?
Self Improvement is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).
Who created Self Improvement?
It is built and maintained by janeaaaa (@janeaaaa); the current version is v1.0.0.
More Skills