brainstorm

name	brainstorm
description	Facilitate a structured feature brainstorming session for the personal finance app. Use this skill when the user wants to explore a new feature idea, plan what to build next, or think through a product direction. Triggers on phrases like "brainstorm", "let's think about", "what should we build", "plan a feature", or "I have an idea".

Brainstorm

Help the user explore and shape a feature idea through structured thinking — from raw idea to a clear, actionable direction.

Step 1 — understand the seed idea

If the user has not provided a feature idea yet, ask one open question:

"What area or problem do you want to explore?"

If they already described something, confirm your understanding in one sentence before proceeding. Do not ask multiple questions at once.

Step 2 — explore the problem space

Before jumping to solutions, frame the why:

• Who benefits? Which user or persona does this serve?
• What's the pain? What frustration or gap does this address in the current app?
• How does this fit? Is it core (income/budgeting/savings/investments), or adjacent?

Present this as a concise framing paragraph, not a questionnaire. Infer what you can from context; ask only if genuinely ambiguous.

Step 3 — generate ideas

Produce 3–5 distinct approaches to solving the problem. Make them meaningfully different — not just variations in naming. For each:

• Give it a short name (2–4 words)
• One sentence on what it does
• One sentence on the key trade-off (complexity, scope, user friction, etc.)

Format as a compact list. Do not go deep on any single idea yet.

Step 4 — go deeper on the top picks

Ask the user which idea(s) resonate, or if they have no preference, pick the 2 most promising and say why.

For each selected idea, explore:

• Core interaction: What does the user actually do? (describe the UX in 1–2 sentences)
• Data model hint: What new data or state does this need?
• Dependencies: Does this require another feature to exist first?
• Scope gut-check: Is this a 30-minute spike, a half-day feature, or a larger initiative?

Keep each section tight — bullet points preferred.

Step 5 — identify risks and open questions

Surface the top 2–3 things that could go wrong or that need a decision before building:

• Technical unknowns
• UX assumptions that need validating
• Scope creep traps
• Anything that could block the implementation

Step 6 — produce a decision summary

End with a clear, copy-paste-friendly summary block:

---

Feature: [name] Problem it solves: [one sentence] Chosen approach: [one sentence] Key open questions: [bulleted list, max 3] Suggested first step: [one concrete action — e.g. "sketch the data model", "build a static mockup of X", "spike the API integration"]

Step 7 — offer to act

Ask: "Ready to start building, queue it for later, or explore further?"

• Build now: pass the decision summary to the feature-builder agent. Do not implement it yourself — feature-builder handles the full implementation, build check, and test run.
• Queue for later: call asana_create_task directly. Use asana_list_workspaces and asana_get_projects_for_workspace to resolve the "Personal Finance App" project GID, then create a task named [Feature] <feature name> with notes structured as:

  ## Problem
  <problem it solves>
  ## Approach
  <chosen approach>
  ## First Step
  <suggested first step>
  ## Open Questions
  <bulleted list, or "None">
  ## Context
  Source: brainstorm session
  App: Personal Finance App (teams/tom_artem/exercise_one)
  Stack: React 18 · Vite · plain CSS · JavaScript
  Implementation agent: feature-builder