Free will and determinism

What did you ask exactly?

Here is my message:

Helen,

Greetings. A few of us have had a years-long discussion about libertarian free will. I keep asking a question about it but do not get a satisfactory answer. Perhaps you can point me to something you have written about this question.

Let's accept that we have libertarian free will and that an agent can somehow influence the nexus of a decision with some sort of top-down causation. In other words, the decision is influenced by more than just deterministic events and true randomness, by some sort of indeterministic factor.

My question is: How does the agent make the decision that influences the nexus in a way that is not itself just deterministic and random? We agree that the agent is producing a top-down effect on the decision, but how? I cannot imagine, nor have I read any description of, how the agent itself is free from the confines of determinism and coin-flipping.

I'll accept that the top-down effect is free from prior event determinism, at least in part. But what does that leave to complete the decision-making "process" other than randomness? (I scare-quote "process" because many people tell me that by using such a word, I am assuming determinism.)

What is the process/procedure/method/algorithm/technique/trick that a free agent uses to make a free decision?

~~ Paul