[CLI] search option

[jbolda]

Each time we run a unique hash is created so we have no way to tell if a particular tag has been published yet. Being able to consistently publish every tag for a release means that we need to be able to recover from a failure in the process.

If we have a search option (that can have a time element to it to limit the scope of search), we can search and check for a transaction. If it already exists, then we know we can skip publishing that particular tag. This would let us retry as needed without creating unnecessary tag+hash while attempting.

[jbolda]

Thinking about the time element more, it seems like an arbitrary number would still fall down (say for cases where a package is published once a quarter, what time range is "enough"?). If we specify a GitHub Release / tag, I imagine we could (optionally) use the date from that publish as the lower bounds from which to search.