Is it acceptable for a support technician to view your codebase?

There was one product announcement in Monday's Leaderboard newsletter that piqued my interest: a support agent that parses your codebase to detect what bug caused a ticket. My initial response was excitement because I have spent too many nights trying to decipher "it doesn't work" into a stack trace.

Then came the more cautious response. Access to code is a big deal. Even read-only access is a big deal because it will show internal naming, half-baked features, and sometimes even some code that someone accidentally committed and never thought about again. A support tool from a company you met a few months ago is a different kind of risk compared to a colleague with access to a code repository.

This is where I stand today. I would give it a try in one non-critical repository first with scoped tokens, and I would make sure to have a log of all the files accessed for each ticket.

But there are also some benefits to consider. Quicker triage means customers get a useful response faster, and engineers no longer receive generic pings. All I'd ask is that any trade-offs be made explicit before connecting a repo.

For those of you managing a smaller team: would you connect your support agent to your code? What would need to happen first?