≈1.3k tokens on demand. Your agent reads this file only when SKILL.md points to it.
SCAMPER Creativity Technique Reference
Purpose: Structured idea-modification checklist developed by Bob Eberle (1971), adapting Alex Osborn's earlier creative-thinking questions into seven prompt lenses — Substitute, Combine, Adapt, Modify (magnify/minify), Put to other use, Eliminate, Reverse. SCAMPER fires best when an existing object, process, or service exists and the goal is incremental transformation rather than first-principles reinvention.
Scope Boundary
- flux
scamper: lens-by-lens transformation of an existing artifact. Produces a 7-cell idea matrix with prompt-bank-driven ideation per lens.
- flux
reframe (default): full DEEP pipeline — assumptions, cross-domain combine, perspective shift, crystallize. Use when the frame, not the artifact, is stuck.
- flux
cross: cross-domain Bisociation only. SCAMPER is intra-artifact modification; cross is inter-domain knowledge transfer.
- flux
challenge: First Principles and Assumption Reversal. Use when the question is "should this exist at all?" not "how do we modify it?".
- flux
analogy: structural mapping from a source domain. SCAMPER lens A (Adapt) is analogy-adjacent but shallower — analogy goes deeper into structural alignment.
- flux (elsewhere): iterative dialogue brainstorming. SCAMPER is a single-pass checklist; Flux is multi-turn co-creation.
- spark (elsewhere): feature proposal from data. SCAMPER feeds Spark with raw idea candidates.
Workflow
ENTER → receive target artifact (product, process, feature, service)
→ verify it is concrete enough for lens-by-lens probing
PROBE → apply each of 7 lenses with its prompt bank
→ generate ≥3 ideas per lens (minimum 21 raw ideas)
FILTER → ASN test per idea (Actionability / Specificity / Novelty)
→ drop synonym-substitutions and trivially-true modifications
MATRIX → arrange surviving ideas into 7-lens × N-ideas grid
→ flag lens-pair tensions (e.g., Eliminate vs Combine)
DELIVER → hand to Spark (feature proposals) or Magi (decision among
modified variants) with lens citations
The 7 Lenses (Eberle 1971)
| Lens |
Core question |
Prompt bank |
Fires best when |
| S Substitute |
What can be swapped? |
Swap material, person, rule, ingredient, place, approach |
Component is taken for granted |
| C Combine |
What can be merged? |
Combine purposes, audiences, products, ideas, services |
Two adjacent things share users |
| A Adapt |
What can be borrowed? |
Adapt from nature, other industry, history, child's view |
Stuck on novel problem with old precedent |
| M Modify |
What can be magnified or minified? |
Larger, smaller, faster, slower, stronger, lighter, more frequent |
Scale or intensity is wrong |
| P Put to other use |
What else could this serve? |
New audience, new context, new market, new lifecycle stage |
Asset is underutilized |
| E Eliminate |
What can be removed? |
Drop step, feature, role, constraint, decoration, dependency |
Bloat or YAGNI suspected |
| R Reverse |
What can be inverted? |
Reverse order, roles, polarity, direction, cause-effect |
Flow feels arbitrary |
Lens Selection Heuristics
| Symptom |
Strongest lenses |
| "It's too complicated" |
E (Eliminate), M (minify) |
| "It's too generic" |
S (Substitute), P (other use) |
| "It's too slow" |
M (modify timing), E (eliminate steps), R (reverse order) |
| "Users don't understand it" |
A (adapt familiar metaphor), S (substitute language) |
| "Competitors copy us" |
C (combine unique pairs), R (reverse industry default) |
| "We've optimized everything" |
R (reverse), P (other use) |
Anti-Patterns
- Running SCAMPER alone as the entire reframing — produces incremental ideas, never paradigm shifts. Always pair with
challenge or shift upstream.
- Stopping at the first idea per lens — the first idea is usually the obvious one. Push to ≥3 per lens to escape the local maximum.
- Treating "Combine" as "list features side by side" — Combine means fusion (shared mechanism, shared user, shared moment), not concatenation.
- Skipping Reverse because it feels silly — Reverse is the highest-novelty lens; silliness is the signal it is working.
- Substitute-as-synonym — "use blue instead of red" passes Substitute syntactically but fails ASN. Substitute the function, not the label.
- Eliminate without asking what depends on this — produces fragile suggestions Builder cannot ship.
- Applying SCAMPER to a vague target ("our company") — lenses need a concrete artifact. Narrow scope first.
- Reporting all 21+ raw ideas without filtering — overwhelms downstream agents. CRYSTALLIZE to top 5-7 cross-lens candidates.
Handoff
- To Spark: surviving ideas with lens citations as feature proposal candidates.
- To Magi: when 3+ modified variants compete and a decision is needed.
- To Flux: when one cell looks promising but underdeveloped — open iterative dialogue on that cell.
- To Omen: Eliminate and Reverse cells that may carry hidden failure modes — pre-mortem before committing.
- To
analogy: Adapt-lens hits that deserve deeper structural mapping, not just surface borrowing.
