brainstorm_skill

Overview

This skill transforms vague ideas into fully formed designs and specs through collaborative, incremental dialogue. It guides a structured process—from context gathering and focused questioning to approach comparison and stepwise design sections—so teams reach clear, implementable decisions. Use it to reduce ambiguity, evaluate trade-offs, and produce a committed design document ready for implementation.

How this skill works

The skill inspects project context (directory structure, recent commits, and relevant code/docs) and then asks one clear question at a time to refine purpose, constraints, and success metrics. It presents 2–3 approaches with trade-offs and leads the user to select one. After selection, it delivers the design in 200–300 word sections (Overview, Architecture, Data Model, Error Handling, Testing) and waits for confirmation before proceeding. Once validated, it compiles the final design and offers optional implementation setup steps like worktrees and TDD plans.

When to use it

• You have a vague idea that needs concrete scope and success criteria.
• Designing a feature that has multiple viable implementation approaches.
• Clarifying constraints, integrations, or unknown dependencies before coding.
• Preparing a design document for team alignment or PRs.
• Exploring architecture or data model changes before implementation.

Best practices

• Gather project context first: check key directories, recent commits, and existing patterns.
• Ask one question at a time; prefer 2–4 mutually exclusive choices with a recommended option.
• Always show 2–3 approaches and lead with a recommendation and trade-offs.
• Present design in 200–300 word sections and stop for confirmation after each section.
• Apply YAGNI: ruthlessly remove nonessential features for the MVP.

Example use cases

• Convert a product request like “add offline sync” into a scoped design with success metrics and tests.
• Decide between microservice vs. monolith approaches with clear pros/cons and recommended option.
• Design a new database schema and migration plan with relationships and rollback strategies.
• Draft error-handling and retry strategies for flaky third-party APIs with trade-off analysis.
• Produce a final design doc and create an implementation worktree plus a TDD-backed plan.