Self-improvement skill for OpenClaw. It captures learnings, errors, and feature requests to support continuous improvement across sessions.

self-improving-agent

Self-improvement skill for OpenClaw. It captures learnings, errors, and feature requests to support continuous improvement across sessions.

This version is OpenClaw-only. To use the skill with other agents (Claude Code, Codex, GitHub Copilot, …), use the original multi-agent version instead: https://github.com/pskoett/pskoett-ai-skills/tree/main/skills/self-improvement

Repository Layout

The publishable skill package is the self-improving-agent/ subfolder — that is what ClawHub installs and what you copy into ~/.openclaw/skills/. Repo-level files (this README, .github/ CI) stay outside the package on purpose.

Skill entry point: self-improving-agent/SKILL.md
OpenClaw hook: self-improving-agent/hooks/openclaw/

Development and Amp orbs

Use the Node.js version in .nvmrc, then run:

npm ci
npm test
npm run typecheck
npm run check:shell

The root npm manifest and lockfile are development-only; the publishable skill subfolder remains dependency-free. Tests use an OpenClaw type stub, so no OpenClaw installation, credentials, or backing services are required.

Amp runs .agents/setup to install the pinned Node.js toolchain and locked development dependencies before taking an orb snapshot. Warm setup runs reuse installed dependencies. A repository-scoped login-shell hook selects this Node version without replacing the orb's system Node. .agents/resume intentionally does no installation because there is no per-wake repair or authentication work.

Installation

With OpenClaw's built-in skill installer — this installs into the active OpenClaw workspace:

openclaw skills install @pskoett/self-improving-agent

Or with the ClawHub CLI (npm i -g clawhub). Note this installs into ./skills under the current working directory, not the workspace:

clawhub install @pskoett/self-improving-agent

Or manually — copy the skill subfolder (not the repo root):

git clone https://github.com/pskoett/self-improving-agent.git /tmp/self-improving-agent-repo
cp -r /tmp/self-improving-agent-repo/self-improving-agent ~/.openclaw/skills/self-improving-agent

Attribution

Remade for OpenClaw from the original repo:

Upgrading

See self-improving-agent/CHANGELOG.md for version history and per-version upgrade notes. After upgrading, re-copy the OpenClaw hook and restart the gateway:

cp -r ~/.openclaw/skills/self-improving-agent/hooks/openclaw ~/.openclaw/hooks/self-improvement

Uninstalling

See self-improving-agent/references/uninstall.md for disable vs. full-removal steps. Review .learnings/ before deleting — it contains your captured learnings, not skill code.