Add requirements or recommendations about allow/deny lists for reference target retrieval

As multi-document OADs become more important due to both new use cases and an increasing number of OADs that are far too large to work with as a single document, we should be more clear about the resource location and retrieval process.

In particular, there is a security risk to fetching resources from arbitrary locations. We should at least RECOMMEND some sort of allow/deny functionality, and require (MUST) a sensible default behavior.

For example, it's conceivable that a poorly designed API runtime tool running at improperly elevated privilege could be induced to load a sensitive file, get confused, and display it in an error message over the network. Or, a tool that does something bad like eval() fetched JSON could be sent a malicious bit of JavaScript instead.

No file or test is named in the issue. Start by reviewing the OpenAPI requirements for multi-document resource location and reference target retrieval, then determine how allow/deny behavior and a sensible default should be specified. Done means the specification clearly states the security requirements and recommendations for implementations.