ideation

name	ideation
description	Creative brainstorming — SCAMPER, 5 Whys, How Might We, angle variations, idea stress-testing, concept refinement.
allowed-tools	Read, Write, AskUserQuestion, WebSearch, WebFetch

Ideation — Domain Knowledge

This skill is loaded into ideation agents via the skills: ideation frontmatter field. It provides domain expertise about brainstorming frameworks and creative techniques — not execution instructions. Agents read workflow files in workflows/ for step-by-step process via progressive disclosure.

Overview

Helps users who have a vague concept, a problem without a solution, multiple competing ideas, or an idea they want to stress-test. Produces a refined, validated pitch ready for structured intake or implementation planning.

Brainstorming Frameworks

SCAMPER (Track A — Flesh out a vague idea)

Best for: User has something but needs to expand, twist, or reframe it.

Letter	Technique	Prompt Pattern
Substitute	Replace a core element	"What if you replaced {X} with {Y}?"
Combine	Merge with adjacent domain	"What if you merged this with {Z}?"
Adapt	Borrow from another field	"What existing solution in {field} could you adapt?"
Modify	Change scale, audience, delivery	"What if this served {different audience}?"
Put to another use	Adjacent markets	"What other market could use exactly this?"
Eliminate	Remove complexity	"What if you removed the hardest part — what's left?"
Reverse	Flip relationships	"What if the user/provider roles were swapped?"

Pick 3-4 lenses most relevant to the specific idea. Don't run all 7 mechanically.

5 Whys + How Might We (Track B — Problem without a solution)

Best for: User has a pain point but no clear product concept.

5 Whys: Drill from symptom to root cause. Ask "Why?" iteratively (3-5 rounds). Stop when you reach a cause that's actionable — not every chain needs all 5.

How Might We (HMW): Convert each root cause into a design challenge:

• "How might we {solve root cause} for {target user}?"
• Good HMW statements are specific enough to inspire solutions but open enough to allow multiple approaches.
• Avoid "How might we build {specific solution}" — that's too narrow.

Comparison Matrix (Track C — Multiple competing ideas)

Best for: User has 2-4 ideas and can't choose.

Score each idea on 5 dimensions (1-3 stars):

1. Problem Clarity — How well-defined is the problem?
2. Target User Clarity — Can you describe exactly who uses this?
3. Feasibility — Can this be built with available skills/budget/time?
4. Uniqueness — How differentiated from existing solutions?
5. Personal Excitement — How motivated is the user to build this?

Present as a matrix. Commentary matters more than scores — explain why each score.

Regroup & Interpret (Track E — Scattered ideas need organizing)

Best for: User has multiple disconnected thoughts, notes, or fragments they've collected and needs help making sense of them as a coherent project direction.

Process:

1. Collect: Ask the user to dump all their ideas, notes, fragments — no filtering
2. Cluster: Group related ideas into themes (3-5 clusters typically emerge)
3. Name: Give each cluster a descriptive label
4. Connect: Identify relationships and tensions between clusters
5. Interpret: Synthesize what the clusters are "saying" as a whole — what's the underlying vision the user is circling around?
6. Propose: Present 2-3 coherent project narratives that weave the clusters together

Key principle: The user's scattered ideas often contain a hidden coherent vision. Your job is to surface it, not impose your own. Ask "Is this what you've been thinking?" not "Here's what you should think."

Inspiration Discovery (Track D — No idea at all)

Best for: User wants to build something but has no starting point.

Seed generation strategy:

1. Understand the user (background, skills, daily frustrations)
2. Identify friction points in their domain
3. Cross-pollinate: combine their skills with underserved markets
4. Generate 3-5 seeds, each with a one-line pitch
5. Iterate on whichever sparks interest

Angle Variations

• A brainstorming session is complete when ≥3 distinct angles have been stress-tested — if you have fewer, run another framework pass before presenting
• If you cannot find a clear Safe Bet / Ambitious Play / Lean Experiment split, ask the user which direction they lean before generating angles — don't invent a split that doesn't exist

After a direction is chosen, always generate three angles to prevent premature commitment:

Angle	Philosophy	Risk Profile
Safe Bet	Most straightforward interpretation, lowest risk, fastest to build	Low risk, potentially lower differentiation
Ambitious Play	Full vision, maximum differentiation, bigger bet	Higher risk, higher potential upside
Lean Experiment	Smallest testable version, fastest to validate with real users	Lowest investment, fastest learning

The user's choice of angle signals their risk appetite and informs downstream decisions (MVP scope, tech stack complexity, launch strategy).

Stress Testing

Every idea should survive basic scrutiny before entering intake. Four key probes:

1. Failure risk: "What's the #1 reason this could fail?" — surfaces assumptions
2. Willingness to pay: "Who pays? How much?" — validates there's a business
3. Existing alternatives: "What's closest today? Why isn't it enough?" — validates differentiation
4. Minimum viable scope: "What's the smallest valuable version?" — prevents scope creep from day one

Adapt probes based on context — skip obvious ones, dig deeper on weak spots.

Business Model Handoff

After ideation produces a Refined Pitch, the natural next step is validating the business model before entering Phase 1 Discovery.

Early-stage / first idea: use Lean Canvas (skills/strategic-frameworks/workflows/lean-canvas.md) — 9 sections, 15 minutes, surfaces whether the idea has a viable model before spending sprint cycles.

Established product / scaling model: use Business Model Canvas (skills/strategic-frameworks/workflows/business-model-canvas.md) — covers partnerships, key activities, and cost structure that Lean Canvas omits.

The ideation-to-model handoff is indeterministic: the agent selects based on pitch maturity signals — no AskUserQuestion required unless both frameworks are equally applicable.

Output Format

Ideation output always includes:

• Refined Pitch — one-liner, problem, solution, target user, differentiator, MVP hint
• Brainstorming Journey — which frameworks were used and why this direction won
• Angles Considered — the three variations and which was chosen
• Stress Test — answers to the four probes

Anti-Patterns

Anti-Pattern	Why It's Wrong	Do This Instead
Picking a direction for the user	Kills ownership and may miss their real intent	Always present options, let them choose
Running all frameworks mechanically	Feels like a checklist, not a conversation	Pick the right framework for the situation
Skipping stress-testing	Produces ideas that fall apart at intake	Always probe, even if the idea seems strong
Over-refining in ideation	Ideation produces a pitch, not a PRD	Stop at the pitch — intake handles the detail
Ignoring "none of the above"	The best idea might not be in your suggestions	Always leave room for user's own direction

Reference Files

File	When to Read
workflows/ideate.md	Step-by-step ideation execution process

Common Rationalizations

Rationalization	Reality
"The idea is obvious, let's just build it"	Obvious ideas have non-obvious failure modes. 20 minutes of stress-testing saves weeks of pivoting.
"We don't need multiple angles"	Single-angle thinking produces single-point-of-failure ideas. Always explore at least 3 angles.
"Brainstorming frameworks are overkill"	Frameworks prevent anchoring bias. Without structure, you optimize the first idea instead of exploring.

Customization

• Framework selection: Adapt which SCAMPER lenses to use based on domain
• Stress test depth: Add domain-specific probes (e.g., regulatory for healthtech)
• Angle definitions: Adjust the three angles for specific project types