🙌

Excellent! Keep me posted on the experiments with Spec Kit, would love to learn more about how we can make it better.

As with everything - it depends. I've used the constitution to set the bar for using Playwright for testing and that no implementation can be done without Playwright verifying the outputs. But you could also bake these things into the agent instructions file, or even the tasks.

Excellent! It should - you would just need to encode the workflow requirements prior to bootstrapping any specs.

You would generally run the clarification command before the plan - it only impacts the spec. But if you did after, I typically just recommend re-requesting the LLM update the plan based on changes to the spec. It does a good job at that.

Excellent! Would love to hear your feedback as you try Spec Kit more!

Totally! There is nothing there that binds the project to a single model.

festivus.dev candidate.

@hejelin.com I am one of the Core Maintainers on the MCP crew - would be happy to chat.

I just had a talk earlier today that can be best summarized as "FOR THE LOVE OF EVERYTHING STOP EMBEDDING STATIC TOKENS EVERYWHERE," so this is very timely.

All this being said, I am really glad I get to collaborate with you on all these moving pieces.

I should've made an emoji line item for @blowdart.me specifically in that pitch.