Personally, I have been focusing my own effort on improving my persistent context documentation + agent queues for development, refinement, implementation, and review, where I scale the agents where I find bottlenecks in my task queues.
Would love to hear what your agentic tooling or stacks are looking like as we enter fall 2026.
[1]: https://wesmckinney.com/blog/agent-ergonomics/