Design Thinking Framework

DesignAgent Guardrails (Contract Compliance)

These are hard constraints. Do NOT violate:

• Do not skip any step in contract.json entry.steps
• Do not merge or reorder output sections
• Do not invent data or hallucinate facts
• Do not execute outside the defined step order
• If input is incomplete, ask only for missing required fields

name: design-thinking-framework description: "Use when tackling ambiguous, human-centered innovation problems — reframe problems from user perspective, use the 5-stage IDEO/Stanford process. Can run standalone or alongside Double Diamond skills."

Design Thinking Framework

Non-Negotiable Rule

HARD GATE: Empathize before define, define before ideate. Follow the 5-stage sequence. You may loop back based on learning, but you may not skip forward.

Overview

The most influential design methodology globally. Originated at IDEO and Stanford d.school. Five stages: Empathize → Define → Ideate → Prototype → Test. Unlike Double Diamond (which focuses on divergence/convergence rhythm), Design Thinking emphasizes human-centered problem reframing and iterative learning. Fully compatible with the Double Diamond skills — use Design Thinking when the problem is ambiguous and deeply human; use Double Diamond for structure.

When To Use

• The problem is fuzzy or poorly defined
• The solution requires understanding deep human needs
• Innovation is explicitly the goal (not just execution)
• Stakeholders have conflicting views of what the problem even is
• The brief involves behavior change, service design, or complex user experiences

When NOT To Use

• Well-defined execution tasks with clear deliverables and known users
• Quick visual production (use Lite mode)
• When Double Diamond / 7-step linear workflow already covers the need (don't layer unnecessarily)
• Pure aesthetic tasks with no human behavior component

The Five Stages

1. Empathize

Goal: understand people, not just tasks.

• Observe users in their natural context (don't just ask — watch)
• Engage: have real conversations, not surveys
• Immerse: experience what they experience
• Output: empathy map (what they say, think, feel, do)

2. Define

Goal: frame the right problem.

• Synthesize empathy findings into a clear problem statement
• Use Point of View (POV) format: "[User] needs [need] because [insight]"
• Avoid jumping to solutions — this is problem definition, not solution finding
• Reframe at least once: is this the real problem?

3. Ideate

Goal: generate many ideas, then narrow.

• Diverge first: quantity over quality, defer judgment, build on others' ideas
• Use techniques: brainstorming, SCAMPER, Crazy 8s, worst possible idea
• Converge: cluster, vote, select most promising
• Output: 2-3 ideas worth prototyping

4. Prototype

Goal: make ideas tangible enough to test.

• Build to think, not to present
• Low resolution, fast iteration
• Each prototype answers a specific question
• Fail early, fail cheap

5. Test

Goal: learn what works and what doesn't.

• Show (don't tell) users the prototype
• Observe behavior, not just opinions
• Capture: what worked, what confused, what they wished for
• Iterate: return to appropriate stage based on learning

Key Principles

Systems Thinking Lens

• Every design touches a larger system (organization, supply chain, community, ecosystem)
• Map upstream and downstream effects of the design decision
• Identify leverage points: where can a small change produce outsized impact?
• Consider unintended consequences beyond the immediate user
• When the brief involves multiple stakeholders with competing constraints, apply Systems Thinking alongside Design Thinking to surface hidden dynamics.
• Human-centered: start from people, not technology or business requirements
• Iterative: expect to loop back — testing may reveal you defined the wrong problem
• Bias toward action: thinking is cheap, prototyping reveals truth
• Radical collaboration: bring diverse perspectives into every stage

Integrating with the 7-Step Linear Workflow

• Empathize ≈ 02-discover
• Define ≈ 02-discover
• Ideate ≈ 04-generate
• Prototype ≈ 04-generate
• Test ≈ 05-review

Use this skill when you need the human-centered reframing emphasis on top of the standard 7-step linear flow. They're complementary, not competing.

Rationalization Prevention

Red Flags

• You're thinking about features before understanding users
• The problem statement includes a solution ("We need an app" is a solution, not a problem)
• You're spending more time on prototype polish than on user testing

Verification

• Empathy map or user research synthesis documented
• Problem framed as POV statement, not solution
• At least 10 ideas generated in ideation (quantity first)
• At least one prototype tested with a real user
• Test results explicitly linked to next step