freemcplab / Directory / Playground / design-cognition-skill

● Community S0ulFood ⚡ Instant

design-cognition-skill

Name: design-cognition-skill (Claude Skill)
Author: S0ulFood

by S0ulFood · S0ulFood/design-cognition-skill

Four-role design-thinking framework (Strategist, Researcher, Executor, Critic) for Claude Code — works for any domain, not just product design.

A domain-agnostic design-thinking skill. Claude plays four roles in sequence: Strategist (what matters), Researcher (what's true), Executor (what to build), Critic (what's wrong). Useful for any non-trivial decision — features, essays, architecture, roadmaps.

▶ Watch demo 📦 Install 💡 Use Cases

Why use it

Key features

4-role structured thinking
Domain-agnostic — works for product, eng, writing, strategy
Critic role forces honest self-assessment
Each role produces an artifact
Roles can be invoked individually for targeted help

Live Demo

What it looks like in practice

design-cognition-skill.replay ▶ ready

0/0

Install

Author repository · Claude Skills documentation

Use Cases

Real-world ways to use design-cognition-skill

Decide whether and how to ship a feature

👤 Product managers, founders ⏱ ~60 min intermediate

When to use: You're weighing a feature and the team's split.

Flow

Strategist

Use design-cognition-skill. Strategist role: should we build [feature]? What are the strategic tradeoffs?✓ Copied

→ Strategic framing — not just pros/cons, but tradeoffs against stated goals
Researcher

Researcher role: what's known from our data, users, competitors?✓ Copied

→ Evidence summary with gaps noted
Executor

Executor role: if we build it, what's the smallest valuable version?✓ Copied

→ Concrete MVP definition
Critic

Critic role: what's wrong with this plan? Be harsh.✓ Copied

→ Specific pushback, not generic risk flags

Outcome: A decision doc you can share that shows the thinking, not just the conclusion.

Pitfalls

Roles blur together — Run each role in its own prompt; don't ask for all four at once

Combine with: filesystem

Get four kinds of feedback on a draft essay

👤 Writers ⏱ ~20 min beginner

When to use: Your essay needs more than 'looks good' or copy-editing.

Flow

Run each role on the draft

Use design-cognition-skill on my essay. Strategist (is the point worth making?), Researcher (are claims defensible?), Executor (is the structure tight?), Critic (what would a hostile reader attack?).✓ Copied

→ Four distinct critiques

Outcome: Feedback a friend wouldn't give you but a good editor would.

Combinations

Pair with other MCPs for X10 leverage

design-cognition-skill + filesystem

Save each role's output as a separate doc

Save outputs to decision/<topic>/<role>.md for each role.✓ Copied

design-cognition-skill + infinite-gratitude-skill

Pair with multi-agent research for breadth plus this for depth

Use infinite-gratitude for discovery, then design-cognition for the final decision doc.✓ Copied

Tools

What this MCP exposes

Tool	Inputs	When to call
strategist	question	What matters here
researcher	question + context	What's true
executor	direction	What to build
critic	plan or artifact	What's wrong

Cost & Limits

What this costs to run

API quota: None
Tokens per call: Moderate — 4 roles, each non-trivial
Monetary: Free
Tip: Invoke single roles when you only need one lens.

Security

Permissions, secrets, blast radius

Credential storage: None

Data egress: None from skill

Troubleshooting

Common errors and fixes

Critic is too gentle

Prompt for adversarial framing — 'what would a hostile VC attack'

Roles produce same content

Separate into 4 prompts; prime each role explicitly

Alternatives