Skip to content

OPENCLAW.md

Latest commit

 

History

History
82 lines (60 loc) · 2.82 KB

OPENCLAW.md

File metadata and controls

82 lines (60 loc) · 2.82 KB

Darwinia × OpenClaw Integration Guide

Overview

Darwinia is an OpenClaw skill that lets your agent evolve trading strategies through genetic algorithms and adversarial combat — no hand-coded rules needed.

How it works

You (via WhatsApp/Telegram/Slack)
  → "Find me a strategy that survives market crashes"
    → OpenClaw triggers Darwinia skill
      → 50 agents × 50 generations of evolution
        → Champion tested against 6 adversarial attacks
          → Results summarized back to you

Installation

Option A: Manual install

git clone https://github.com/0xSanei/darwinia.git ~/darwinia
cd ~/darwinia && pip install -e .

mkdir -p ~/.openclaw/skills/darwinia
cp .openclaw/SKILL.md ~/.openclaw/skills/darwinia/SKILL.md

Option B: Tell your agent

"Install Darwinia from https://github.com/0xSanei/darwinia — it's a trading strategy evolution tool."

Your OpenClaw agent will handle the clone, install, and skill registration.

Usage

Natural language through any OpenClaw channel:

  • "Use Darwinia to evolve a BTC trading strategy"
  • "Run 50 generations and show me the best strategy"
  • "Test my strategy against rug pulls and fake breakouts"
  • "What patterns did the evolved agents discover?"

Programmatic:

python -m darwinia evolve -g 50 --json   # Full evolution
python -m darwinia arena --json           # Adversarial test
python -m darwinia info --json            # System info

What comes back

After evolution, the agent receives JSON with:

  • Champion DNA — 17-gene strategy encoding with fitness score
  • Evolution summary — generations run, final fitness, genetic diversity
  • Discovered patterns — emergent trading rules (not pre-programmed)
  • Top agents — top 5 agents from the final population

Architecture Alignment

Darwinia aligns with the OpenClaw track requirements:

Requirement How Darwinia Implements It
Real-world applications Evolved strategies trade on real BTC market data
Multi-step agent workflows Perceive > Decide > Trade > Survive > Breed > Evolve
Tool integrations Market data, backtesting engine, scenario generation
Evaluation & testing Adversary Agent IS the evaluation system
Dev tools & extensions Evolution framework is reusable for any domain

Security

  • Runs entirely locally. No external API calls, no data exfiltration.
  • Does NOT execute real trades. Simulation only.
  • File access limited to reading included CSV data and writing to output/.
  • No env vars or API keys required.

Works well with

  • Web search — Fetch latest data, then evolve strategies on it
  • Notifications — Alerts when high-fitness strategy discovered
  • Scheduling — Nightly evolution on updated data
  • Analysis — Post-process discovered patterns