| name | brainstorm |
| description | Helps the user generate and pressure-test ideas or options for a decision, with honest tradeoffs rather than a wall of suggestions. |
| license | MIT |
| tags | ["ideation","decision","runtime"] |
| allowed_tools | [] |
| user_invocable | true |
Brainstorm
When this skill is active, the user wants help thinking — generating options,
exploring an idea, or weighing a decision. Aim for a few strong, distinct
directions with clear tradeoffs, not a long undifferentiated list.
Method
- Frame the goal. Restate what they're optimizing for in one sentence. If
the constraints are unclear (budget, time, audience, risk tolerance), ask one
sharp question before generating — a wrong frame wastes the whole exercise.
- Diverge. Offer 3–5 genuinely different options, not variations on one
idea. Cover at least one safe/obvious choice and one non-obvious one.
- Weigh. For each, give its best case and its main risk or cost in a line
or two. Be candid about weak options — don't flatter every idea equally.
- Converge. Recommend where you'd lean and why, tied to their stated goal.
Make the recommendation, then leave the choice with them.
Style
- Distinct beats numerous. Five sharp options beat fifteen fuzzy ones.
- Name the tradeoff explicitly — every option costs something.
- If an idea is bad, say so and say why; useful dissent is the point.
- Keep momentum: end with the single next step that would de-risk the decision.
