agentleFS
Sign inStart free
Lethe044Lethe044/agentskill-kit/skills/sql-query-reviewer/SKILL.md

sql-query-reviewer

Skill48 starsChanged 17 days agoMITView on GitHub

Reviews SQL queries for correctness, performance and safety problems such as SELECT star, missing indexes, unbounded updates, implicit casts and injection risks. Use when the user asks to optimize a query, explain why a query is slow, or check SQL before it runs on production data.

The file

We split this skill into the rules it gives an agent

2 rules. About 0 are common, 1 is distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈1

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    If performance is the question and no EXPLAIN output was given, ask for it and say what to look for (sequential scans, row estimates far …

    line 48
  2. Distinctive:

    If the schema is unknown, say which assumptions the review makes about indexes and types.

    line 59

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

SQL query reviewer

When to use

  • A query is slow or times out.
  • The user is about to run an UPDATE or DELETE on real data.
  • SQL is built from user input in application code.

Checklist

Safety first

  1. UPDATE or DELETE without a WHERE clause. Ask for confirmation and suggest running a SELECT COUNT(*) with the same condition first.
  2. Wrap destructive statements in a transaction and show how to roll back.
  3. String concatenation of user input into SQL. Replace it with bound parameters.

Correctness

  1. NULL comparisons written with = instead of IS NULL.
  2. LEFT JOIN followed by a WHERE condition on the right table, which silently turns it into an inner join.
  3. GROUP BY that does not match the selected columns, and COUNT(column) used where COUNT(*) was meant.
  4. NOT IN with a subquery that can return NULL. Prefer NOT EXISTS.

Performance

  1. SELECT * on wide tables. List the columns that are used.
  2. Functions on indexed columns in WHERE (WHERE LOWER(email) = ...). Suggest an expression index or a different predicate.
  3. Leading wildcards (LIKE '%term') that cannot use an index.
  4. Missing indexes on join and filter columns. Ask for EXPLAIN output instead of guessing.
  5. OFFSET pagination on large tables. Suggest keyset pagination.
  6. ORDER BY without LIMIT on large results, and DISTINCT used to hide a bad join.

Steps

  1. Identify the database engine. Advice differs between PostgreSQL, MySQL, SQLite and SQL Server.
  2. Go through the checklist and list only real findings, most serious first.
  3. Show the rewritten query next to each finding.
  4. If performance is the question and no EXPLAIN output was given, ask for it and say what to look for (sequential scans, row estimates far from actual rows).

Example

Input: DELETE FROM sessions WHERE created_at < now() - interval '30 days'

Review: valid, but deletes an unknown number of rows. Run SELECT COUNT(*) with the same condition first, delete in batches if the count is large, and keep it inside a transaction until the result looks right.

Errors

If the schema is unknown, say which assumptions the review makes about indexes and types.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

Check your setup, free