| name | Data Intelligence & Fabric |
| description | Institutional-grade data cleaning, tick physics auditing, and gap reconstruction. |
Data Intelligence Skill
This skill ensures the "Fuel" (Data) for all other skills is pure. Follow these protocols for any OHLCV or Tick data processing:
1. Tick Physics Protocol
Before any strategy validation or feature engineering:
- Call
scripts/data_auditor.py on the target OHLCV dataset.
- Identify Spikes: Flag candles where volatility exceeds the 20-period standard deviation by more than 4x (potential "fake" broker spikes).
- Audit Wicks: Check if High/Low spreads are realistic compared to the peer-group (other symbols in the same group).
2. Time-Series Integrity (Gap Recon)
Institutional models cannot handle missing time steps.
- Use
scripts/gap_reconstructor.py to audit for missing bars in MT5 history.
- Interpolate: Fill gaps using forward-fill or linear interpolation to maintain a continuous time index.
- Alert: If a gap exceeds 4 bars (M15+) or 1 bar (H1+), log a "Data Quality Warning" in the Audit Trail.
3. Macro Filter Protocol
Institutional execution ignores technicals during extreme macro uncertainty.
- Call
scripts/macro_context.py before opening any new position.
- BLOCK: If an "ULTRA" or "CRITICAL" event is detected within the current session, HALT all execution for that currency pair.
- CAUTION: Increase stop-loss buffers by 50% during "HIGH" impact event windows.
Core Tools:
scripts/data_auditor.py: Spike and anomaly detector.scripts/gap_reconstructor.py: Gap detector and time-series filler.scripts/feature_factory.py: High-dimensional feature generator.scripts/macro_context.py: Fundamental impact scout.
