Humans should not be involved with generating changelogs any more.
Meat based processing is far to inefficient and our silicon overlords can take on this task with ease.

It will take some time for people to get their own automation implemented, but still, this project and standard should start to swing in that direction.

An immediate example of the changes that could occur due to LLM automation, is the removal of the## [Unreleased] heading.

I've created this script for local use to generate changelogs for me and it is working well. I'll move it to CI/CD at some point, early days for it.

• Script:https://github.com/grantcarthew/scripts/blob/main/lib/git/ucl
• Script prompt:https://github.com/grantcarthew/scripts/blob/main/lib/git/ucl-prompt.md

I added a line in the prompt to specifically not include the unreleased header. It is for my use-case though and I'm not keeping a changelog of "Releases". It's a changelog for changes in branches that get merged to main.

It would be easy enough to change this to create or update the changelog from release tag to release tag though.

I'm just opening this discussion. I love that this standard exists and appreciate your work. Thanks.