In CVSS v3, the "scope" indicates whether a vulnerability in an application impacts resources beyond its means. It can have the values "changed" or "unchanged".
I don't fully understand when the scope is changed. For example, in the CVSS examples, an XSS has scope changed because a vulnerability in the application impacts the user's browser. But with XSS you can still only run Javascript within the application context, something which the application could do in the first place. It doesn't seem to me that this impacts resources beyond its means.
And the example CSRF vulnerability has scope unchanged, even though just like the XSS it triggers behavior in the browser.
What does it mean when the scope is changed or unchanged?
It seems from the documentation that the scope is changed when the vulnerability traverses authorization authorities. What is an "authorization authority"?