Yes, I can see the usecase for legacy desktop apps etc, but the web? it's a DOM. WebMCP coming now too, no need for screenshotting or DOM querying then either..
no, you don't 'have to do so manually'. all agents can run 'git commit' for you. if you end up with too many commits for your taste; squash on merge, or before push; `git reset --soft HEAD~3; git commit -m "Squashed 3 commits"`
I just don't get why would you would want an agent to use the browser to do these mundane things (check email, work with calendar etc), when you can simply give it a few tools, and save maybe six gazillion tokens per task?
nice that you saw the business opportunity here! I've had LLMs write commits for me with custom scripts, but this is way more convenient. Privacy concerns arise for enterprise use, but for open source, I'll use it!
https://mrorigo.github.io/envkey/