Somewhat recently I've started to use LLM-based coding tools, notably Claude Code. My workflow here involves creating a dedicated Git worktree to do the work, creating a (Hatch) Python virtual environment for that worktree, and letting the tool do its thing within that worktree.

A critical part of this workflow is reviewing what the LLM actually did, and having it fix some bad coding practices (overuse of global variables, overuse of mocks in tests, public docstrings that describe removed internal behavior, ...).

Meanwhile, I have a fairly robust VSCode setup for local Python development. It has some extensions, but not a lot: GitHub and Python and Jupyter extensions, but not more than a dozen or so in total. So the general review process is

1. Use the VSCode File > Open menu item to locate and open the worktree
2. Wait for VSCode to reload its UI, reload the extensions, and discover source control
3. (Optionally) hatch run which python in a terminal tab, then CmdShiftP, "Python: Select Interpreter" to activate it
4. In the source-control sidebar, right-click the topmost commit and "Compare with Merge Base"

This works, and it's the UI I generally want, beyond the overhead of this process. Once I've reviewed the code, fixed it up, and pushed it, I'll never see this worktree again. I can wind up doing this on multiple changes a day, and it's enough time for me to context-switch away from the code review I was all ready to do.

Are there any workflow changes, alternate tools, or specific VSCode tweaks that can make this code-review cycle faster?