← ClaudeAtlas

sql-query-reviewlisted

Static review of a SQL query — finds anti-patterns, structural problems, and performance issues from the query text alone, then proposes an optimized rewrite with a verification plan. Use when the user says "review this query", "check my SQL", "is this query okay", "clean up this query", or "optimize this" without runtime profile data.
vermapragya/analytics-skill · ★ 0 · API & Backend · score 72
Install: claude install-skill vermapragya/analytics-skill
# SQL Query Review ## When to use this skill Use when reviewing or improving a SQL query **from its text alone** — no query profile or runtime stats required. Triggers: - "Review this query" - "Check my SQL before I ship it" - "Can this query be written better?" - "Optimize this query" (when no query_history / profile is available) - PR review of a `.sql` file Routing to related skills: - Runtime stats available (query_history, query profile, spilling)? → `warehouse-query-optimization` - Suspected **wrong results** (dupes, fanout, missing rows)? → `sql-correctness-review` - Query is correct and fast but unreadable? → `modular-sql-ctes` A full review often chains all three: correctness first, then this skill, then runtime profiling if still slow. ## Required inputs | Input | Why it matters | |---|---| | Query text | The thing being reviewed | | Intent (one sentence) | "What question does this answer?" — catches queries that are fast but wrong | | Expected output grain | One row per what? Needed to judge joins and aggregates | | Approx table sizes | A `SELECT *` on 1K rows is fine; on 1B rows it's a finding | | Run frequency | One-off ad-hoc vs hourly pipeline changes the bar | If table sizes / frequency are unknown, review anyway and mark size-dependent findings as conditional. ## Workflow 1. **Restate intent and grain.** One sentence each. If you can't infer them from the query, ask — every later judgment depends on this. 2. **Correctness scan first.** Optimizing