/speckit.implement - stay in same chat session or open a new one?

[wangyuqiang9509]

[aurexav]

This totally depends on your model’s context length.

"Remember" everything is always good for a machine.

[wangyuqiang9509]

This totally depends on your model’s context length.这完全取决于您的模型的上下文长度。

"Remember" everything is always good for a machine."记住" 对机器来说，一切总是好的。

Thank you for your question. Since the context length will be exceeded after a few rounds of conversation, I'd like to know whether it's better to start a new chat window or continue in the current one when the context is exceeded.

[aurexav]

You’ll have to open a new one sometimes. You can’t be sure the next one won’t exceed the context window.

Essentially, you don’t need to worry about this. Progress is recorded in task.md and each task is independent.

The downside is that the new command takes some time or tokens to recover the latest state.

Alternatively, you could tell your agent to summarise progress to a file when it finishes and then check the progress file in the next round.

[wangyuqiang9509]

You’ll have to open a new one sometimes. You can’t be sure the next one won’t exceed the context window.

Essentially, you don’t need to worry about this. Progress is recorded in task.md and each task is independent.

The downside is that the new command takes some time or tokens to recover the latest state.

Alternatively, you could tell your agent to summarise progress to a file when it finishes and then check the progress file in the next round.

非常感谢