Regenerate and interpret the sweep database dashboard, surfacing actionable items.

## Steps

1. Regenerate the dashboard (optionally filtered to one asset/TF):
   ```
   python3 mine_sweep_db.py
   # or filtered:
   python3 mine_sweep_db.py --asset COINBASE_BTCUSD --timeframe 4H
   ```

2. Read `results/sweep_dashboard.md` and provide a structured summary:

   **Data inventory**: for each combo, show tier (SPARSE/EMERGING/USABLE/RICH/DEEP) and row count.
   Flag any combos that just crossed a tier threshold.

   **Actionable locks** (from Action Summary section):
   - List all HIGH CONFIDENCE locks with the param name, direction, and evidence strength
   - Suggest running `/lockParams` if there are ≥3 HIGH CONFIDENCE items

   **Range tightening** (from Range Tightening section):
   - List params flagged with ✂ Tighten, showing current range → suggested range
   - Note any step-size adjustments recommended
   - If significant tightening is suggested, note the approximate search-space reduction

   **Cross-asset universality** (if RICH/DEEP data available):
   - Note any params that show consistent sign across all assets → candidates for global locking

   **Red flags**:
   - Any combo with >30% degenerate solutions (entry ≈ exit threshold)
   - Any param where sign flips between bull and bear regimes

3. Recommend next actions in priority order:
   - If HIGH CONFIDENCE locks exist → `/lockParams`
   - If range tightening is substantial → update params JSONs manually or prompt to do it
   - If most combos are SPARSE/EMERGING → recommend longer optimization runs before acting on data
   - If RICH/DEEP data is available on BTC combos → consider running SHAP: `python3 tools/shap_feature_importance.py`