Skip to content

// MUST-use-first creative skill — explores intent, requirements, constraints, divergent options before you build.

Brainstorming Pro

skill --stat
  • last update

    Jun 2, 2026

  • license

    MITv1.0.0

  • format

    SKILL.md

quickstart.sh
3 steps
  1. Install

    // Downloads SKILL.md into ~/.claude/skills/brainstorming-pro/

    $ curl -fsSL --create-dirs -o ~/.claude/skills/brainstorming-pro/SKILL.md https://claudeskil.com/skills/brainstorming-pro/SKILL.md
  2. Invoke

    // Inside Claude Code, type the skill name as a slash command

    > /brainstorming-pro help me ship this
  3. Or just ask

    // Claude loads the skill by itself when a request matches its description

    > help me ship this
brainstorming-pro/
references
SKILL.md
readonly
name:
Brainstorming Pro
slug:
brainstorming-pro
version:
v1.0.0
license:
MIT
author:
@chiragthummar16
categories:
tags:
#brainstorming#ideation#requirements#discovery#design-thinking
description:

MUST-use-first creative skill — explores intent, requirements, constraints, divergent options before you build.

README.md

brainstorming-pro/README.md
10 sections

Brainstorming Pro

Brainstorming Pro is the skill you invoke before writing a line of code, a spec, or a piece of copy. Instead of jumping straight to a solution, it interrogates the request first — what the user actually wants, what's required, what constraints apply — then runs a structured divergent-then-convergent ideation pass so you commit to the right idea rather than the first one. It's built for builders who tend to start fast and discover the misunderstanding three days later.

When to use it

  • You're handed a vague feature request ("add notifications") and need to pin down scope before designing.
  • You're starting a new product, page, schema, or content piece from a blank page.
  • A stakeholder gave you a solution ("build a dashboard") when you suspect the real problem is upstream.
  • You have several plausible approaches and want a defensible reason to pick one.

How it works

The skill runs a four-phase loop and writes its findings to a working artifact (brainstorm.md in your project by default) so decisions are traceable:

  1. Intent — clarifying questions that separate the stated ask from the underlying goal and success criteria.
  2. Requirements — surfaces functional, non-functional, and explicitly out-of-scope items.
  3. Constraints — maps technical, time, budget, and dependency limits that prune the option space.
  4. Diverge → Converge — generates many distinct options (deliberately including weird ones), then scores them against the requirements and constraints to recommend one with rationale.

It is conversational, not a fill-in form: it asks one cluster of questions at a time and adapts follow-ups based on your answers. The output is a ranked shortlist plus a written "here's why" you can paste into a PR description or design doc.

Usage

bash
curl -fsSL --create-dirs -o ~/.claude/skills/brainstorming-pro/SKILL.md https://claudeskil.com/skills/brainstorming-pro/SKILL.md

# then, in a session:
> /brainstorm "we want users to be able to share their reports"
# → asks: share with whom? read-only or editable? public links or
#         account-gated? expiry? then maps constraints, then proposes
#         3-4 ranked approaches with trade-offs written to brainstorm.md

Tips & gotchas

  • Answer the intent questions honestly, even the obvious-seeming ones. The skill's value comes from the gap between what you asked for and what you meant — skip the questions and you get generic output.
  • It will resist solutioning early. If you push for code in phase 1, it'll redirect. That friction is the point; don't fight it on genuinely trivial tasks.
  • Keep brainstorm.md — feeding it to your implementation skill or agent later gives downstream work the full requirement and constraint context for free.
  • Re-run only the diverge/converge phase when requirements change rather than restarting the whole loop.

FAQ

Is this overkill for small changes? Yes — for a one-line fix, skip it. Reach for it when the cost of building the wrong thing exceeds a few minutes of questioning.

Does it write any implementation code? No. It stops at a recommended direction and rationale. Hand the artifact to your build skill afterward.

Can I skip phases? You can jump straight to diverge/converge if requirements are already nailed down, but the convergence scoring is only as good as the constraints you fed it.

$ cat reviews/

Reviews

// No reviews yet. Be the first.
Loading review form…

$ ls related/

explore all →