Matched on the wrong key and nearly reported a bug that wasn't there

Data cleanup

Situation: Comparing two lists to check whether an exclusion setting is working.
What the AI did: It used the wrong basis for matching and twice in a row came close to reporting "there's a bug," though everything was working correctly.
How to ask next time (the operator's example): "Before comparing, say what counts as the same item. Before calling it a bug, test that rule on one item you know is working."
One question to ask: What counts as the same item?
Check it yourself: Common reasons two lists fail to match are extra spaces, letter case, and leading zeros a spreadsheet dropped. Ask for three matched pairs and three unmatched items, shown as raw text, before trusting it.
AI used (month): Claude Code (2026-08)
Times this mistake was recorded: once
Source: The operator's work records (the details were changed to a different task)