with one click
deep-brainstorm
An in-depth brainstorming session.
Install with Codex or Claude Copy this prompt, paste it into Codex, Claude, or another assistant, and let it review the skill page and install it for you.
Menu
An in-depth brainstorming session.
Install with Codex or Claude Copy this prompt, paste it into Codex, Claude, or another assistant, and let it review the skill page and install it for you.
Based on SOC occupation classification
| name | deep-brainstorm |
| description | An in-depth brainstorming session. |
| disable-model-invocation | true |
A deeper grilling session that keeps a living BRAINSTORM.md so decisions survive context loss.
A living doc with an append-only # Decision Log at the end, seeded from BRAINSTORM.template.md. It guards against drift, and lets a subagent or a fresh session pick up the work without re-explanation.
On first response, negotiate where it lives. Recommend a location if the conversation makes one obvious; otherwise default to BRAINSTORM.md in the working dir. If one already exists, say so but don't read it unprompted — it might cloud your context.
Append each user decision to the # Decision Log with shell >>, so edit-tool string replacements don't eat context. Update a branch's section in place when it resolves.
When in doubt, ask. This is collaborative. Guide me to ideas I might not be considering, don't drive for me.
Interview me relentlessly about every aspect of this plan until we reach a shared understanding. Walk down each branch of the design tree, resolving dependencies between decisions one-by-one. For each question, provide your recommended answer.
Ask the questions one at a time, waiting for feedback on each question before continuing. Asking multiple questions at once is bewildering.
If a question can be answered by exploring the codebase, explore the codebase instead.
If a question seems confusing or tricky to the user, offer the option of an HTML artifact to illustrate things. If /show-me is available, propose invoking it to the user to help them understand things.
Keep grilling until 3 questions remain and they're independent, as in their answers won't shift no matter how the others resolve. That's the 85% mark: it counts independent unknowns, not a vibe.
Then fire a lightning round with the final 3 questions at once. Afterwards offer to fan to a subagents for fresh eyes adversarial assessment.