A timeline cannot loop and then offshoot into an open timeline. Each event in a causal loop is like a single point. It goes against our intuition, but they don't "repeat" as if they happen over and over again -- there is just one instance of each event, and they all exist at once. So if any one point on this loop were to somehow change to being open, then it no longer causes the very events which lead to its own existence, so the entire causal chain would not have existed in the first place.
Its subtle, but such a "break" is different than the notion of extensive splitting in the many worlds interpretation, where those splits might be better understood as "overlays" of valid/possible timelines which all maintain their respective loops, and they just happen to share the same informational state at the point of decoherence/splitting.
If you are embedded in a CTC, there is no changing that.
I'm not saying you're changing a timeline, simply that it's not the only looped timeline that decoheres from a single point in a many-worlds formulation. If you arrived back at a previous point in your timeline via reverse causation, then several timelines/worlds still decohere from that point -- including the timeline you were "previously" in. They all exist. In any case, this isn't the core of the essay's argument. I noted that the MWI is used to illustrate possibilities; we simply require a deterministic interpretation at a minimum. Even if only one causal loop timeline exists, the rest of the argument follows in terms of which causal loops are possible.
it would not be an embedded agent, but something from outside the Universe
I defined a universe simply as a causal system, so initiating a simulation within your own simulation entails causation, thus one causal system, or universe. But even within one simulation, what's to stop an agent from manipulating it? In Maxwell's demon, we can think of the chambers, demon, and the demon's environment as a single simulation -- but the demon can still rearrange its information.
That is incompatible with classical GR, the best I can glean.
GR is an emergent description of quantum information/entanglement, so it does not inherently prohibit reverse causation by those means.
From that point on in your post, it looks like you are basically throwing **** against the wall and seeing what sticks, so I stopped trying to understand your logic.
To quote the classic movie, "Life, uh, finds a way". Which is a nice and warm sentiment, but nothing more.
Subjective sensation may play more of a role in a causal loop than you think. I expand on that topic in this essay.