Self-Improvement Skill for Coding Agents for Openclaw

Workspace Initialization

Ensure the .learnings/ directory and baseline templates exist in your workspace root:

mkdir -p .learnings
[ -f .learnings/LEARNINGS.md ] || printf "# Learnings\n\nCorrections, insights, and knowledge gaps captured during development.\n\n**Categories**: correction | insight | knowledge_gap | best_practice\n\n---\n" > .learnings/LEARNINGS.md
[ -f .learnings/ERRORS.md ] || printf "# Errors\n\nCommand failures and integration errors.\n\n---\n" > .learnings/ERRORS.md
[ -f .learnings/FEATURE_REQUESTS.md ] || printf "# Feature Requests\n\nCapabilities requested by the user.\n\n---\n" > .learnings/FEATURE_REQUESTS.md

OpenClaw Installation

Configure this skill directly within your Openclaw Skills ecosystem:

Via ClawdHub:

clawdhub install self-improving-agent

Manual Method:

git clone https://github.com/peterskoett/self-improving-agent.git ~/.openclaw/skills/self-improving-agent

Enabling Reminder Hooks (Optional)

To enable automatic prompts and post-tool execution checks, register the opt-in hook scripts in your project configuration (e.g., .claude/settings.json):

{
  "hooks": {
    "UserPromptSubmit": [{
      "matcher": "",
      "hooks": [{
        "type": "command",
        "command": "./skills/self-improvement/scripts/activator.sh"
      }]
    }]
  }
}

The self-improvement skill organizes tracking logs into structured, machine-readable Markdown lists located in the .learnings/ workspace directory.

Directory and Files

Taxonomy Fields

All entries must support standard identification and filtering fields:

• ID Schema: TYPE-YYYYMMDD-XXX where TYPE is LRN (learning), ERR (error), or FEAT (feature).
• Priority: Must use critical, high, medium, or low.
• Status: Transitioned through pending, in_progress, resolved, wont_fix, and promoted.
• Area: Tracked by codebase layers: frontend, backend, infra, tests, docs, and config.