[verb] `size-histogram` to learn item size distribution

It's easy to know the average item size (it's in the table metadata) but it's not so easy to know the histogram of item sizes. This verb would calculate that. How many items from 0-1kb, from 1-2kb, etc.

It would accept the usual where clause to be able to look at just how large a subset of items are. How big to my orders tend to be?

```
./bulk size-histogram --table t --where 'SK like "ORDER#%"'
```

Start by inspecting the existing CLI verbs and the implementation of the usual --where clause, then run the example command against a test table. Done means size-histogram reports item counts in size ranges and applies the supplied filter to a subset such as ORDER# items.

Written by the indexing model from the issue text.