Managing corrections and open questions outside tasks.md/plan.md

[Kulagin-G]

Hi all,

Based on suggested videos on YouTube from Den Delimarsky and other authors I still don’t quite understand the concept of clarifying implementation details - how is the model supposed to know that tasks.md or plan.md contain inaccuracies/errors or require clarification.

I can interact with the model directly in the chat, but it’s more convenient to do this through a file where I describe all the necessary clarifications for the new plan. It seems that a separate files can make it more simple:

• clarifications.md would be useful - the single source of corrections/clarifications for the plan and tasks
• (optional) decisions.md as the decision log
• (optional) questions.md for open questions (if you don’t want to keep them in clarifications.md).

Then, using commands like /update scope=plan or /update scope=tasks, we can ask the model to make clarifications or apply corrections based on clarifications.md / decisions.md / questions.md, which contain the user input.

Additionally, the command /validate will be useful to validate integrity (check for contradictions, dangling task references, etc).