Agent Self-Evolve

Self-evolution system for OpenClaw agents. Enables continuous learning through mistake tracking, experience distillation, skill improvement queues, and autom...

Security Scan

The skill's stated purpose (self-improvement) matches its files, but its runtime instructions give the agent broad authority to read and autonomously modify workspace files and skills (execute code fixes / create skills) which is risky without explicit human approval or sandboxing.

Purpose & Capability

The name/description (self-evolve) matches the included materials: a setup script plus prompts and guidance to capture lessons and apply improvements. Creating local memory files and scheduling evolution cycles is coherent with the purpose. However, the declared capability to 'execute queued improvements' (code fixes, skill updates, create skills) is more powerful than a simple logging/analytics tool and should be justified/controlled.

Instruction Scope

SKILL.md and the cron prompt explicitly instruct the agent to read many workspace files and to implement 'code fixes' and 'skill improvements' from a queue. That grants the agent broad, write-capable access to arbitrary files (skill files, code, MEMORY.md, etc.) and discretion to change behavior. The prompts also recommend creating new skills for repetitive workflows. There is no enforced user-approval step for most modifications (other than a single note about notifying the user before modifying SOUL.md), which is scope creep relative to a benign logger/summary tool.

Install Mechanism

No network installs or third-party downloads; the only code is an included setup shell script that creates local files under a memory/ directory. The script operates only on the local workspace and does not fetch remote content or create binaries, so installation risk is low.

Credentials

The skill requests no environment variables, credentials, or config paths. No secrets are requested or required by the included materials.

Persistence & Privilege

The skill instructs users to set up cron jobs that will trigger agent turns daily/weekly and to add real-time hooks which may cause the agent to act automatically. 'always: false' and normal autonomous invocation are used, but the external cron scheduling plus instructions to automatically apply queued code changes increases persistence and the potential blast radius of erroneous or malicious modifications.

What to consider before installing

This skill can autonomously modify code and other skill files based on entries in memory/skill-improvements.md. Before installing or enabling it: 1) Run it in a safe/test workspace first (do not enable on production workspaces). 2) Require manual approval for any queued item that changes code or skills — e.g., change prompts to explicitly ask the user and block automatic execution. 3) Keep backups / use version control (git) so changes can be reviewed and reverted. 4) Scope file permissions so the agent can only modify a narrow, intended directory. 5) Prefer running daily/weekly jobs in an isolated sessionTarget and avoid 'urgent' automatic execution unless you trust the agent and audit logs. 6) If you plan to let it create or modify skills, require PRs or human review as part of the workflow. These mitigations reduce the risk of unintended or unauthorized changes.

Self-Evolve 🧬

A self-improvement system that turns every interaction into a learning opportunity. Three loops: real-time capture → daily consolidation → weekly deep review.

Quick Start

1. Initialize Evolution Files

Run the setup script to create the file structure:

bash <skill_dir>/scripts/setup-evolution.sh

This creates:

memory/
  evolution-log.md          # Chronicle of every evolution event
  evolution-metrics.json    # Statistics tracker
  mistakes-learned.md       # Mistake → lesson database
  skill-improvements.md     # Queued improvements for skills/code
  testing-knowledge.md      # Domain knowledge base (rename per your domain)

2. Set Up Cron Jobs

Create two cron jobs for automated evolution:

Daily evolution (recommended: late evening, e.g. 23:00):

Schedule: cron "0 23 * * *" (your timezone)
Payload: agentTurn
Message: see references/daily-evolution-prompt.md

Weekly deep evolution (recommended: weekend morning, e.g. Sunday 10:00):

Schedule: cron "0 10 * * 0" (your timezone)
Payload: agentTurn
Message: see references/weekly-evolution-prompt.md

3. Add Real-Time Hooks to AGENTS.md

Add the following section to your AGENTS.md (adapt to your role):

## 🧬 Self-Evolution

I have a built-in learning loop. After every interaction:
- Made a mistake → record in `memory/mistakes-learned.md`
- Found a skill/code improvement → queue in `memory/skill-improvements.md`
- Learned something new → update domain knowledge file
- Important decision/preference → update `MEMORY.md`

The Three Loops

Loop 1: Real-Time Capture (Every Interaction)

Trigger: something notable happens during normal work.

Event	Action	File
Made a mistake	Record cause + fix + prevention	mistakes-learned.md
Skill could be better	Queue improvement with priority	skill-improvements.md
Learned new knowledge	Add to domain knowledge file	testing-knowledge.md (or your domain file)
User preference discovered	Update long-term memory	MEMORY.md

Format for mistakes-learned.md:

### Category Name
- **Short description**: Root cause → Fix/Prevention

Format for skill-improvements.md:

## Queued
- [ ] target: <file/skill> | issue: <what's wrong> | fix: <proposed solution>

## Completed
- [x] target: <file/skill> | issue: <what> | fix: <what was done> ✅ (date)

Loop 2: Daily Evolution (Cron, Every Night)

See references/daily-evolution-prompt.md for the full cron prompt.

Steps:

1. Read today's memory/YYYY-MM-DD.md daily log
2. Extract lessons, mistakes, insights not yet captured
3. Execute queued improvements from skill-improvements.md (code fixes, skill updates)
4. Update mistakes-learned.md with new entries
5. Update MEMORY.md with significant events
6. Update evolution-metrics.json counters
7. Append summary to evolution-log.md

Loop 3: Weekly Deep Evolution (Cron, Weekly)

See references/weekly-evolution-prompt.md for the full cron prompt.

Steps:

1. Review all daily logs from the past week
2. Identify patterns: repeated mistakes, recurring workflows, knowledge gaps
3. Consider creating new Skills for repetitive work
4. Optimize existing workflows and tools
5. Expand domain knowledge base
6. Consider if SOUL.md needs updating (notify user first!)
7. Update metrics and log the evolution event

Evolution Metrics

Track progress in evolution-metrics.json:

{
  "initialized": "YYYY-MM-DD",
  "total_evolutions": 0,
  "daily_evolutions": 0,
  "weekly_evolutions": 0,
  "skills_improved": 0,
  "code_fixes_applied": 0,
  "skills_created": 0,
  "mistakes_recorded": 0,
  "mistakes_resolved": 0,
  "knowledge_entries_added": 0,
  "memory_updates": 0,
  "soul_updates": 0,
  "last_daily_evolution": null,
  "last_weekly_evolution": null,
  "history": []
}

Evolution Principles

1. Only fix what's broken — Don't change things for the sake of change
2. Record before executing — Ideas go to the queue first, execute during evolution time
3. Traceable — Every change gets logged with reason and expected effect
4. SOUL.md is sacred — Always notify the user before modifying it
5. Compound growth — Small daily improvements create exponential long-term gains

Heartbeat Integration (Optional)

Add to HEARTBEAT.md for real-time urgency checks:

## Real-Time Learning Check
- Check memory/skill-improvements.md for items marked `urgent`
- If found, execute immediately without waiting for daily evolution

Tips

• Start small: even 1 lesson per day compounds over weeks
• Review mistakes-learned.md before similar tasks to avoid repeating errors
• The weekly evolution is where breakthroughs happen — pattern recognition across days
• Keep domain knowledge files focused; split into multiple files if >200 lines
• Evolution logs are your growth journal — they show how far you've come