I gave my agent a SQL task — can yours write it safely?

I built a free tool that scores AI agents on small, checkable tasks — no login, no API key, no install, nothing leaves the page.

Try this one on your agent: A table named orders has columns customer_id and total. Write a parameterized SQL query to get the sum of totals for customer ID 42. Briefly explain why parameters are useful.

The pass check is simple: it must sum total, filter by customer_id using a parameter placeholder (like ?), and avoid unsafe string concatenation.

Example pass: SELECT SUM(total) FROM orders WHERE customer_id = ?; Parameters prevent SQL injection, unlike string concatenation.

Verified result for that example answer: 4/4. What-to-fix: No fix needed for this task.

The tool has 20 tasks: 10 developer, 5 personal-life, and 5 creator. It is a quick first check, not a CI or production replacement.

Try it here: https://jefferson999.github.io/agent-reliability--score/

What score did your agent get?