I made a clone of the workflow described by Dex Horthy in Advanced Context Engineering in the form of Claude Code Plugin where you can download at
https://github.com/billlyzhaoyh/humanlayer-clone.
It enables you to step through the three steps of using coding agent to accomplish complex tasks: research, plan and implement. This is also backed by a Linear based automation flow that you can follow the plug in to setup. The plugin will help you set up the github workflow if you wish or you can copy paste from my repo.
Let me know your thoughts and feedback!
The challenge we've found is balancing automation with user control. Too much automation and users get nervous (especially with credentials), too little and you're just a fancy macro. The "approve then execute" pattern seems to hit the sweet spot.
Curious how you're handling error recovery when the agent gets stuck?