Data Flow Tracker Skill

# Data Flow Tracker Skill

**Purpose**: Trace data flow from capture → database to identify where data gets lost.

## When Claude Should Use This Skill

Use this skill when:
- User says "where did the data go?" or "trace the data"
- User asks "why isn't game X in the database?"
- Debugging missing data issues
- User says "track game {game_id}"
- Data exists in artifacts but not in database

## What This Skill Does

1. **Traces** data for a specific game or date through all pipeline stages
2. **Checks** if data exists in: raw/ → derived/ → packages/ → database
3. **Identifies** where data gets lost (e.g., in raw/ but not derived/)
4. **Shows** timestamps at each stage
5. **Displays** data transformation at each step
6. **Explains** what happened in plain English

## Critical Constraints

- NO emoji (Windows environment)
- Read-only operations
- Handle missing data gracefully
- Clear explanation of where data was lost

## Example Output

```
DATA FLOW TRACE (Game: nfl_12345, Date: 2025-10-23)

STAGE 1: RAW CAPTURE
[OK] Found in artifacts/nfl/raw/2025-10-23/games/12345/game.json
      Captured at: 2025-10-23 15:20:34
      Game: MIN @ LAC
      Markets: 3 (moneyline, spread, total)

STAGE 2: DERIVED
[OK] Found in artifacts/nfl/derived/2025-10-23/unified_derived.json
      Processed at: 2025-10-23 15:25:12
      Metrics: L5 averages, usage data

STAGE 3: PACKAGES
[FAIL] NOT FOUND in artifacts/nfl/packages/2025-10-23/
       Packaging stage was not run

STAGE 4: DATABASE
[FAIL] NOT FOUND in games table
       Cannot ingest without packages

DIAGNOSIS
[WARN] Data was captured and derived but never packaged
[WARN] Run packaging scripts (09-13) to complete pipeline
```