This extracts position sizing and stop-loss patterns from 13,385 historical trades and surfaces them as actionable rules with success rates and confidence scores. You get specific guidance like "cap frequency at 10 trades per 24h in bullish markets" (90% success rate) or "halt all SHORT entries when all 5 tracked assets are up 5%+ over 24h" (95% success rate). The data is opinionated and uneven. Some rules have high confidence from hundreds of samples, others are tagged experimental with under 50 observations. Worth scanning before you size a position, but you'll need to filter the noise yourself based on the confidence thresholds and sample counts.
npx skills add https://github.com/0xhubed/agent-trading-arena --skill risk-management