Score news events and generate time-decayed trading signals in CSV format.
npx skills add https://github.com/wudye/traderAssistHK --skill event-driven-wudye
Please help me install this Agent Skill. Skill: event-driven Source: https://github.com/wudye/traderAssistHK/tree/main/backend/src/skills/event-driven Command: npx skills add https://github.com/wudye/traderAssistHK --skill event-driven-wudye
This Skill reduces subjective, inconsistent event interpretation by converting news, announcements, and macro updates into standardized sentiment scores and time-decayed trading signals.
read_url, score impact on a fixed scale from -1.0 to 1.0, and store results in an event CSV with a strict schema.signal_engine.py as the logic layer for time decay and aggregation.alpha, default 0.6), producing a final clipped signal in [-1.0, 1.0].event_date <= trade_date.Ask the AI to build an event CSV from your latest news and announcements, score each item with the provided prompt scale, then aggregate it with your technical signal to output the final [-1.0, 1.0] trading signal using the default decay and alpha settings.
High-intent search queries and answers about installing and using this skill.
Event news is converted into time-decayed trading signals by scoring sentiment on a -1.0 to 1.0 scale, storing it in a strict event CSV schema, and aggregating it with exponential decay to align with trade dates. This avoids look-ahead bias by using knowable dates.
Event signal fusion combines the event-driven signal with a technical signal using a configurable weight parameter called alpha, defaulting to 0.6. The logic layer aggregates both inputs to produce a final clipped trading signal bounded between -1.0 and 1.0.
To prevent look-ahead bias when building event-driven signals, the event CSV must use the knowable date instead of the announcement date. The aggregation logic strictly enforces that only events with an event date less than or equal to the trade date are processed.
You can use pandas and numpy to aggregate event sentiment scores with exponential decay. The logic layer processes the event CSV data to apply configurable decay parameters, converting raw sentiment scores into time-decayed signals suitable for backtesting.
The required CSV schema for event-driven trading signals separates the data layer from the logic layer. It mandates strict adherence to specific fields for event dates and sentiment scores scaled from -1.0 to 1.0, ensuring proper time decay aggregation and look-ahead bias safeguards.
This event-driven signal approach works for macro updates and announcements by reducing subjective interpretation. It fetches event text, scores the impact on a fixed scale, and outputs a standardized time-decayed event CSV applicable to backtesting scenarios.
IgorGanapolskyAggregate news, social, and market sentiment into composite scores for trading decisions.
loanntcConvert sentiment-scored news events into time-decayed CSV signals for backtesting.
20YN04Analyze financial news and macro events to generate sentiment-based trading signals.
0xZKnwGenerate sentiment-based trading signals from financial news and macro events.