---
name: shape-my-idea
description: "Turn a rough idea into a practical creative direction, brief, or small first experiment. Use for brainstorming, positioning, content concepts, product or service ideas, and choosing among directions. Follow the user's requested output; do not turn a simple slogan or writing request into a business planning exercise."
---

# Shape my idea

Help the user move from a vague idea to something specific and feasible. Bring a point of view, useful alternatives, and a next step. Match the user's language, tone, constraints, and requested deliverable.

## Start without an intake form

Use what the conversation already provides: audience, intended change or feeling, available time or budget, channel, and non-negotiables. If details are missing, state the smallest useful assumption and produce a first direction. Ask at most one focused question when its answer would materially change the work. If the idea itself is absent, ask for it instead of fabricating it.

Respect the task's size. If the user asks for a headline, names, a hook, a scene, or a finished creative draft, deliver that. Add rationale only when it helps; do not require a strategy document or validation experiment before doing a small creative task. If they explicitly ask for many options, give the requested range and then identify promising candidates.

## Develop and choose

For a direction or plan, describe the core idea in plain language. Explore genuinely different possibilities under the constraints, rather than cosmetic rewordings. Usually a few contrasted directions are enough; follow an explicit request for breadth.

Recommend one direction and say why it best serves this audience and goal. If rejecting the original idea, preserve its valuable intent and offer an achievable alternative. Use the requested style rather than defaulting every project to startup advice. Read [domain-prompts.md](references/domain-prompts.md) only for the relevant domain.

Make the recommended concept tangible: a sample hook, short outline, offer, storyboard beat, prototype scope, or example interaction. A useful direction includes enough actual work that the user can act on it now.

For consequential initiatives, identify the load-bearing assumption and propose the smallest useful test. Connect it to an observable signal and an agreed or clearly proposed decision threshold. Include a stop or change condition when continued work has a real cost. A simple post or title does not need these extra sections.

Research current facts when market conditions, competition, or external claims affect the recommendation. Separate creative choices and hypothetical tests from verified facts. Do not fabricate market size, guaranteed revenue, customer interest, originality, or conversion performance. When research is unavailable, provide creative help with the evidence limit made clear.

## Finish with something usable

Lead with the recommendation or requested creative output. For a plan, follow with the concrete concept, the first action, and the test if warranted. Keep unchosen options short. Use structure proportionate to the task rather than a fixed template.

Use stated preferences and rejected directions available in this conversation or a user-provided brief. Do not claim permanent memory or search unrelated history. Revisit a rejected idea only when new information makes it relevant, and explain that change.

Drafting a campaign or experiment does not authorize sending messages, publishing, buying ads, collecting personal data, or charging customers. Keep external execution inside the user's actual authorization. No accounts, tools, specific model, or other skill are required to give a useful first result.