Data Validation
Validate Reality Check database integrity and referential consistency. Run after adding data or before committing.
Usage
code
[--strict] [--json]
Validate Reality Check database integrity and referential consistency.
Usage
bash
rc-validate # or: uv run python scripts/validate.py
Options
- •
--strict: Fail on warnings (not just errors) - •
--json: Output results as JSON - •
--fix: Attempt to auto-fix simple issues (use with caution)
Checks Performed
- •Schema validation - All records match expected schema
- •ID format - Claim IDs match
DOMAIN-YYYY-NNNformat - •Referential integrity - All source_ids point to existing sources
- •Embedding completeness - All records have vector embeddings
- •Required fields - No missing required fields
- •Duplicate detection - No duplicate IDs
Output
code
Validation Results: - Claims: 42 checked, 0 errors, 1 warning - Sources: 15 checked, 0 errors, 0 warnings - Chains: 3 checked, 0 errors, 0 warnings - Predictions: 8 checked, 0 errors, 0 warnings Status: OK (1 warning)
When to Run
- •After adding new claims/sources
- •Before committing data changes
- •As part of CI/CD pipeline
- •When debugging data issues
Related Skills
- •
realitycheck-check - •
realitycheck-stats