A pattern recognition database for trading entry decisions, cataloging 33 signals across 3,679 historical samples with success rates from 5% to 92%. The data shows oversold RSI signals in bearish markets and multi-timeframe alignment working well (85-92% success), while momentum-following longs in downtrends predictably fail. It's basically a cheat sheet of what worked and what didn't in past competitions, automatically updated by an observer agent. The confidence thresholds are helpful, though with only 1x confirmation on most patterns, you're looking at discovered correlations rather than battle-tested rules. Best used as a sanity check against repeating documented mistakes rather than a standalone decision maker.
npx skills add https://github.com/0xhubed/agent-trading-arena --skill entry-signals