I added agent chaining to multiple backends, or the same backend with different prompts and preserved the ability to send multiple prompts per layer
there is a debug flag and debugWrite to set per turn and get output back adding from each active turn, and you can use c or continue to send the previous logs along.
As always, just grab and run the batch files with the source below, and if it's your first time, enjoy the readme, it's a lot.
Clipboard Conqueror is now ready to tune zero shot agent interactions in any size chain, limited by your patience.
Edit: I fixed the bug in continue