Plan with one model Implement with another now live in Command Code Your planning model explores the codebase and writes the plan. Approve it, and your implementation model takes over mid-run. Set both in /config → Feature models

Read docs about choosing planning and implementation models.

Love to see it! Keep up the awesome work you guys do!

Finally, you guys have done it Thanks a lot!

Took you long enough.

Lessgoo, multi-agent orchestration in cc! Also i am not able to use cc inside herdr, sent @MrAhmadAwais details of the issues in dm. Kindly check.

omp already doing this; what's the point?

I do planning, but always "keep on this mode" because I already have pipeline to store "plans" elsewhere. Will this feature still work? The prompt always include "do not code yet, plan first yada yada...", normally commandcode will prompt either to enter plan mode or remains.

🐐

This is a good job, but is there any test to prove an effect similar to 1+1 > 2

plan model explores then mid-run handoff to implement… that /config feature-models split is neat

Interesting

Won’t that consume more tokens because of cache miss?

一个建议，可以优化一下 /loop 和 /goal 吗，它们真的非常非常容易停下来，甚至比不使用更容易停下来， 我不得不去使用 omp 来接入 commandcode 才能长时间驱动 Agent

Can I use command code with my openai and Claude subscription account and not an APIKey?

welcome to 2025.

watch the handoff: the implementer inherits the plan but none of the evidence behind it. carry what the planner ruled out and why, or it re-explores, re-litigates decisions you already approved, and you can't tell which model made which call.

which model gets the blame when it breaks

What context does the implementation model inherit when it takes over mid-run?

为什么现在Deepseek v4.1 Flash的TPS慢到发指，经常是20、30？那以前很快的呀。