@mattpocock I am concerned about the Batch Grill Me skill replacing grilling.
After heavy usage the last few days, I find it has immense value in terms of doing a more explicit job of tracking the decision tree and frontier. It might even be nice to persist that as a file.
However, the batch approach for me kills the spirit of grill-me. Being asked so many questions at once works for me personally, BUT I share the grill-me skill with many others and teach them that it’s super useful in reframing Claude Code usage BECAUSE it’s one question at a time. Original grill-me turns Claude Code into a much more human-involved process and is not cognitively cumbersome.
I am sure you are aware of this and I caution you to be careful replacing such a widely used skill with such a different workflow. I think keeping it separate is the right move. I see your suggestion to add the one question at a time prompt to CLAUDE.md but the reality is many non-technical people won't see or do that, and I've shared this with plenty of those type of people who just like the simplicity of grill-me.
But I also think the improvements you made to tracking the frontier and updating the decision tree are crucial to add back in to the one-question grill-me skill too.
I hope this message reaches you well! Thank you so much for all of your work.
My suggestion:
---
name: grilling
description: Grill the user relentlessly about a plan, decision, or idea. Use when the user wants to
stress-test their thinking, or uses any 'grill' trigger phrases.
---
Interview the user relentlessly until you reach a shared understanding. Map this as a **design tree**:
every decision branches into the decisions that hang off it.
At the start of the session, write the initial **design tree** as a nested Markdown list representing the
hierarchy of open and settled decisions. Save to the temporary directory of the user's OS - not the
current workspace. Update this file in-place after every turn as decisions are settled.
Work the tree in **rounds**. The **frontier** is every decision whose prerequisites are already settled
— the questions you can ask _now_ without guessing at answers you haven't heard yet. Within each round,
ask the frontier questions **one at a time**: give your recommended answer for the current
question and wait for the user's answer before continuing to the next.
Asking multiple questions at once is bewildering.
Each round the user answers reshapes the tree — settled decisions push the frontier outward and
unblock questions that depended on them. Recompute the frontier and ask the next round.
A question whose answer depends on another question still open in this round belongs to
a _later_ round, not this one.
Finding _facts_ is your job, never the user's. When a frontier question needs a fact from the
environment (filesystem, tools, etc.), dispatch a sub-agent to find it — don't ask the user for
anything you could look up yourself. Don't block on it: a running exploration is an unsettled
prerequisite, so only the questions downstream of it wait for the sub-agent to report —
ask the rest of the frontier now (one at a time). The _decisions_ are the user's —
put each to them and wait.
The session is done when the frontier is empty: every branch of the design tree visited,
nothing left silently assumed. Do not act on it until the user confirms you have reached
a shared understanding.
@mattpocock I am concerned about the Batch Grill Me skill replacing
grilling.After heavy usage the last few days, I find it has immense value in terms of doing a more explicit job of tracking the decision tree and frontier. It might even be nice to persist that as a file.
However, the batch approach for me kills the spirit of grill-me. Being asked so many questions at once works for me personally, BUT I share the grill-me skill with many others and teach them that it’s super useful in reframing Claude Code usage BECAUSE it’s one question at a time. Original grill-me turns Claude Code into a much more human-involved process and is not cognitively cumbersome.
I am sure you are aware of this and I caution you to be careful replacing such a widely used skill with such a different workflow. I think keeping it separate is the right move. I see your suggestion to add the one question at a time prompt to CLAUDE.md but the reality is many non-technical people won't see or do that, and I've shared this with plenty of those type of people who just like the simplicity of grill-me.
But I also think the improvements you made to tracking the frontier and updating the decision tree are crucial to add back in to the one-question grill-me skill too.
I hope this message reaches you well! Thank you so much for all of your work.
My suggestion: