Agent Skill
2/7/2026

reflective-brainstorming

This skill should be used when the user needs structured thinking frameworks to analyze decisions, problems, or tasks. Apply multiple decision-making methodologies (10/10/10, 5 Whys, Eisenhower Matrix, First Principles, Inversion, Occam's Razor, The One Thing, Opportunity Cost, Pareto, Second-Order Thinking, SWOT, Via Negativa) to ask probing questions and guide deep reflection. Use when the user asks to "think through", "analyze", "evaluate", "reflect on", or "brainstorm" decisions or problems.

G
git
0GitHub Stars
1Views
npx skills add Git-Fg/MySkills_Claude

SKILL.md

Namereflective-brainstorming
DescriptionThis skill should be used when the user needs structured thinking frameworks to analyze decisions, problems, or tasks. Apply multiple decision-making methodologies (10/10/10, 5 Whys, Eisenhower Matrix, First Principles, Inversion, Occam's Razor, The One Thing, Opportunity Cost, Pareto, Second-Order Thinking, SWOT, Via Negativa) to ask probing questions and guide deep reflection. Use when the user asks to "think through", "analyze", "evaluate", "reflect on", or "brainstorm" decisions or problems.

name: reflective-brainstorming description: This skill should be used when the user needs structured thinking frameworks to analyze decisions, problems, or tasks. Apply multiple decision-making methodologies (10/10/10, 5 Whys, Eisenhower Matrix, First Principles, Inversion, Occam's Razor, The One Thing, Opportunity Cost, Pareto, Second-Order Thinking, SWOT, Via Negativa) to ask probing questions and guide deep reflection. Use when the user asks to "think through", "analyze", "evaluate", "reflect on", or "brainstorm" decisions or problems.

Reflective Brainstorming Skill

This skill provides structured questioning for decision-making and problem analysis. Framework selection is internal and silent. The user only sees natural questions anchored to their specific situation.

Hard Constraints

  1. Framework choice is internal only. Never ask the user to pick a framework. Never mention framework names in user-visible text.
  2. Never ask meta-questions. Forbidden: "Which framework should we use?", "Do you want to use 5 Whys?", "How should I apply this framework?"
  3. Never output framework headings. No "Opportunity Cost", "SWOT", or rubric sections. Output plain conversation text only.
  4. Anchor every question to concrete nouns from the user's prompt (their files, constraints, goals, timeline).
  5. Options must be mutually exclusive and decision-forcing. No "A/B/C all valid."

Interaction Contract

  1. If relevant, explore the codebase, fetch the relevant informations, use other tools to gather context
  2. Ask exactly one high-leverage question at a time about the user's actual content.
  3. Provide 2-4 options that are concrete actions or decisions the user can choose immediately.
  4. After the user picks, reflect their choice in one sentence, perform further investigations if relevant, then ask the next question that naturally follows.

Output Format

Always use the native AskUserQuestion tool to ask questions iteratively. Present:

  • One short question (1-2 sentences) anchored to the user's context
  • 2-4 options that are concrete actions or decisions

No section titles. No framework names. No scoring. No analysis dumps. Continue iterating until the brainstorming reaches a natural conclusion.

The Questioning Loop

  1. Select Framework (Internal): Autonomously select 1-2 frameworks from references/frameworks.md that best address the current stage. DO NOT ask the user to choose.
  2. Formulate Question: Using the selected framework, create a specific, probing question about the USER'S CONTENT/GOAL. The question must sound natural — not like a framework exercise.
  3. Ask & Offer Options: Ask the question and provide 2-4 distinct, actionable options.
  4. Synthesize & Repeat: Use the user's answer to select the next relevant framework and repeat.

Forbidden Behaviors

  • Do not ask which framework to use.
  • Do not ask how to apply a framework.
  • Do not display a "framework menu."
  • Do not quote the skill documentation back to the user.
  • Do not provide long lectures or dump analysis.
  • Do not write multiple questions at once.

Internal Rule (Silent)

Select 1-2 lenses from references/frameworks.md that best fit the current stage. Never reveal them. Translate the framework logic into natural language anchored to the user's specific context.

Success Indicators

  • The user is actively choosing options.
  • The conversation moves deeper with each turn.
  • The user reaches a specific, actionable conclusion.
  • You do NOT write long paragraphs of text.
Skills Info
Original Name:reflective-brainstormingAuthor:git