B
brainstorming

brainstorming

obra
2026-04-06

You MUST use this before any creative work - creating features, building components, adding functionality, or modifying behavior. Explores user intent, requirements and design before implementation.


Introduction and Core Concept

The brainstorming skill is designed to act as a mandatory precursor to any creative implementation work. It solves the common pitfall of “unexamined assumptions causing wasted work,” especially in seemingly simple projects. The core goal is to force a structured, collaborative dialogue that transforms vague ideas into fully formed designs and specifications before a single line of code is written.

Core Concept

Always design before implementing. By exploring context, asking clarifying questions, and writing a formal specification, this skill prevents wasted effort and ensures alignment between user intent and the final architecture.

Installation and Usage Guide

https://github.com/obra/superpowers

Workflow Analysis (SOP)

Read original SKILL.md

  • Explore project context: The AI is forced to check existing files, docs, and commits before doing anything else, ensuring the design fits the current architecture.
  • Offer visual companion: For UI/UX tasks, it explicitly asks if visual mockups are needed, separating visual and conceptual thinking.
  • Ask clarifying questions: Limits the AI to asking one question at a time to prevent overwhelming the user and to deeply understand constraints.
  • Propose multiple approaches: Requires presenting 2-3 different solutions with trade-offs, preventing tunnel vision on the first available idea.
  • Present design and get approval: Breaks the design into manageable sections and mandates explicit user approval before proceeding.
  • Write and review spec: Forces the creation of a formal markdown specification document and an internal “self-review” for placeholders or ambiguities.
  • Transition to implementation: Strictly dictates that the terminal state is invoking a planning skill (writing-plans), creating a hard gate against premature coding.

Skill Design Evaluation

  • Strengths of SOP Integration: Enforces a strict order of operations before implementation. The inclusion of a formal “Spec self-review” step ensures the AI validates its own output for placeholders and ambiguities before requesting user feedback.
  • Potential Limitations: The strict “one question at a time” constraint may increase interaction time. Mandating a full markdown specification document can be unnecessary for minor tasks.
  • What Makes a Good Skill: It clearly defines anti-patterns (what not to do) and establishes a deterministic state machine for the interaction, effectively restricting premature code generation.