Staged mode for safe outputs seems like something we underplay for cautious adopters - it's such a great story - "try it out and see what it would do"

• It applies to all safe outputs right?  And just displays them in the Actions run?
• It looks like it's not currently documented except a few little bits for custom safe outputs and glossary?
• Could we add intent links to the display of items (like "create-issue" ) that support them? And download-artifact-and-apply sequences of steps for other items like adding comments?