The share of your code that ran during the tests. It says what ran. It says nothing about whether anything checked the result.
Low coverage tells you something real: that code has never been run by a test. High coverage tells you much less, and it is trivially easy for an agent to hit any number you give it.
Say it to your agent
“Run bun test --coverage and list the lines nothing runs. Don't add tests just to raise the number. For each uncovered line, say what would go wrong if it were broken, then write an assertion that would catch it.”