GitChain SQL Query

Execute SQL queries directly on the GitChain Supabase database.

Process

1. Parse $ARGUMENTS for the query intent
2. Translate to safe SQL (read-only by default)
3. Execute via Supabase MCP
4. Format results

Common Queries

Count by namespace

SELECT namespace, COUNT(*) as containers, 
       pg_size_pretty(SUM(pg_column_size(data))) as data_size
FROM containers 
GROUP BY namespace ORDER BY containers DESC;

Search products

SELECT id, namespace, identifier, data->>'name' as name, 
       data->'pricing'->>'listPrice' as price
FROM containers 
WHERE to_tsvector('english', data::text) @@ to_tsquery($1)
LIMIT 20;

Data quality check

SELECT namespace,
  COUNT(*) as total,
  COUNT(data->'pricing') as has_pricing,
  COUNT(data->'dimensions') as has_dimensions,
  ROUND(COUNT(data->'pricing')::numeric / COUNT(*)::numeric * 100, 1) as pricing_pct
FROM containers WHERE type = 'product'
GROUP BY namespace;

Output Format

Present results as clean markdown tables. Include:

• Query executed (sanitized)
• Row count
• Execution time
• Formatted results

Rules

• READ-ONLY queries by default
• Never expose connection strings
• Always LIMIT results (max 100)
• Sanitize any user input in queries
• Show query plan for complex queries