What problem does it solves?
When data is missing in the final database, it's difficult and time-consuming to pinpoint exactly where it got lost in a multi-stage pipeline. This Skill provides a clear, step-by-step trace, helping you quickly diagnose and resolve data flow issues.
Core Features & Use Cases
- Multi-Stage Trace: Traces a specific game or date through all pipeline stages: raw capture, derived data, packaged artifacts, and database ingestion.
- Loss Point Identification: Clearly identifies at which stage data was lost or failed to progress (e.g., found in raw but not derived).
- Actionable Diagnosis: Provides plain English explanations and recommendations for resolving data flow issues, guiding you to the next steps.
- Use Case: If a user reports "game X is not in the database," use this Skill to trace
game X and determine if it was captured, derived, packaged, or if ingestion failed.
Quick Start
Trace a specific NFL game through the pipeline
python .claude/skills/data-flow-tracker/scripts/trace_data.py --sport nfl --date 2025-10-23 --game-id 12345