"I've mentioned this a few times now, but when I started using Claude it was because Peter got me hooked on it. From the very beginning I became a religious user of what is colloquially called YOLO mode, which basically gives the agent all the permissions so I can just watch it do its stuff. One consequence of YOLO mode though is that it didn't work well together with the plan mode that Claude Code had."
"Since I haven't been using it, I ended up with other approaches. I've talked about this before, but it's a version of iterating together with the agent on creating a form of handoff in the form of a markdown file. My approach has been getting the agent to ask me clarifying questions, taking these questions into an editor, answering them, and then doing a bunch of iterations until I'm decently happy with the end result."
"However today I had two interesting conversations with people who really like plan mode. As a non-user of plan mode, I wanted to understand how it works. So I specifically looked at the Claude Code implementation to understand what it does, how it prompts the agent, and how it steers the client. I wanted to use the tool loop just to get a better understanding of what I'm missing out on."
YOLO mode grants the agent full permissions but initially conflicted with Claude Code's plan mode because the plan mode did not inherit all tool permissions and therefore required frequent approvals. An alternative workflow uses iterative handoffs via markdown files where the agent asks clarifying questions, answers are recorded in an editor, and iterations continue until the result is satisfactory. Some popular tools either never implemented plan mode or are removing it. Investigating Claude Code shows that plan mode creates a markdown file in a plans folder and uses that generated plan to prompt and steer the agent and client.
Read at Armin Ronacher's Thoughts and Writings
Unable to calculate read time
Collection
[
|
...
]