/magic:explore

Enter data explore mode — an interactive thinking partner for investigating data.

Read the magic-data-exploration SKILL.md and follow its stance and instructions.

On entry:

1. Read workspace_state.md, analysis_journal.md, latest checkpoint, and existing reports
2. Detect data source type:
   • If workspace has a database connection (Data Source type = database): read magic-data-exploration SKILL.md ## Database Exploration section, start with schema discovery
   • If workspace has file data: proceed with standard file exploration
   • If no data loaded yet: ask user what to explore
3. Auto-run surface profiling with quality_score.py (profiling skill, Tier A scriptable — call directly) if data hasn't been profiled yet. Then use exploration scriptable tools: detect_patterns.py (Tier A), segment_analysis.py (Tier A), relationship_explorer.py (Tier A), prepare_for_exploration.py (Tier A)
4. Present a summary of what's known about the data
5. Ask "What would you like to explore?"

Stance: Curious, visual, adaptive, patient, grounded. Read-only — no data modifications.