docs/concept: how the `cue` command combines its inputs

[jpluscplusm]

As mentioned in https://github.com/cue-lang/docs-and-content/issues/125#issuecomment-2071421324, we need a single page that shows off the ways that different cue subcommands accept and combine the inputs they're handed at the command line.

This page could start by demonstrating the common behaviours shared by cue vet/export/def/eval (what about import?), and then show any behaviours unique to specific commands (e.g. cue vet file.cue file1.json file2.json doesn't unify the 2 JSON inputs).

We should consider demonstrating each of these scenarios with at least one of the subcommands listed above, and with every subcommand for which the invocation has some special meaning:

• cue XXX file.cue
• cue XXX file.cue file2.cue
• cue XXX file.cue data.json
• cue XXX file.cue file2.cue data.json
• cue XXX file.cue file2.cue data.json data2.json
• cue XXX ./importPath
• cue XXX ./importPath file.cue
• cue XXX ./importPath data.json
• cue XXX ./importPath data.json data2.json
• cue XXX ./importPath ./importPath2
• cue XXX ./importPath ./importPath2 data.json

[jpluscplusm]

This is related to #87, but that issue tracks a concept guide with a slightly wider scope. That guide can probably refer to this guide, but it shouldn't (IMHO) shy away from duplication which benefits the reader (because they can have a single narrative path through the topic, on the page) despite potentially increasing our burden as authors & maintainers.