I'm not 100% sure what you mean by this. Generally the distributed stuff is done by each developer having a private repository, and a public-facing repository. If Joe wants changes from Alice, he pulls them from her public repository, into his private repository. When he has changes that he wants to make public to Alice, he pushes his changes from his private repository into his public repository. The project maintainer will generally maintain the "master" public repository by pulling changesets from the contributors' public repositories.

You can pull from multiple repositories, without ever pushing anything back, so in effect, everything is a dynamic read-only link in that regard.

git also has the notion of a submodules, which allows you to create a directory in your tree, and associate that with a particular revision of some other repository. This is what gives you the ability to track dependencies between projects. When you "check out" a git repo containing a submodule, the submodule isn't automatically checked out.