What delineates a mechanism and its contents, the circle in the CMO diagram, is not so much facts about the world but the things which I happen to need to call upon to make a causal explanation

I found the concept of a trigger in realist evaluation totally baffling because RE is supposed to be somehow scientific, yet most forms of scientific explanation don't involve actual triggers.

But then I realised, triggers (and mechanisms) are best understood from an epistemic perspective.

"This mechanism gets triggered here" can be parsed as: there is a need for an explanation here: "I am invoking this mechanism to explain something that needs explaining".

We don't invoke oxygen to explain the forest fire, although both oxygen and the spark are (let's say) necessary.