We believe that the following happened (though we haven't tried to repro so we're not 100% sure):
* A commit was pushed to a remote branch
* This requested an audit
* Some time later, someone went through and deleted many remote branches as part of a clean-up effort
* The auditor was working through his queue and saw an old (incomplete) audit corresponding to a commit on one of the deleted branches
* Clicking on the audit shows:
```
Command failed with error #128!
COMMAND
git log -n 1 --format='%P' '08d5e6c2a164f10679fdf412dc187c50f2acb289'
STDOUT
(empty)
STDERR
fatal: bad object 08d5e6c2a164f10679fdf412dc187c50f2acb289
```
The problem we encountered is: it wasn't obvious (since the clean-up happened a while ago) that the reason for this error was that the commit was on a deleted branch. We wasted quite some time figuring out what was going on.
Potential solution ideas:
* Improve the error message
* Have a daemon that processes deleted commits/branches and deletes associated (incomplete?) audits