Good catch! Thanks for that!
I have changed the way Analyze works by returning a proper error so it can actually be tested, before that it couldn't (my bad). In its current state your test will always fail, could you:

• Make it so the test cleans up the directory after it runs so it doesn't clutter the temporary dir
• Checking for errors in the test when creating files (using if err != nil { t.Error(...) }
• Checking that Analyze actually returns the expected error (inverted logic if err == nil { t.Error(...) }
• You can ignore errors in the defer check by placing a //nolint: errcheck next to the line you wan to ignore error checking for (for example in your defer statement)

Let me know if you need help with that!

Updated the test as requested:

• Added error checking for file/directory creation
• Added inverted logic to expect error from Analyze
• Added nolint comment for defer cleanup

Thanks for the feedback!

Thanks for your PR!
Your fix is in the v1.4.3 that is being published 😄