# Scamper > Transform existing ideas into improved variants using the SCAMPER method (Substitute, Combine, Adapt, Modify, Put to other use, Eliminate, Reverse). Use when the user has a baseline idea to improve, a process to redesign, or needs systematic ideation for differentiation. - Skill: `v0lka/scamper` (Agent Skill, multi-file: 2 files) - Install (CLI): `npx skillmds@latest add v0lka/scamper` - Raw SKILL.md: https://api.skillmd.com/api/skills/v0lka/scamper/raw - Safety review: pending - Works with: Claude Code, Claude.ai, OpenAI Codex - Category: Research & Search - Author: v0lka (https://skillmd.com/u/v0lka) - Updated: 2026-09-17 - Page: https://skillmd.com/skills/v0lka/scamper --- # SCAMPER ideas transformer ## 1) Purpose This skill helps a user transform an existing idea (product, service, process, policy, or content concept) into multiple improved or differentiated variants using the full SCAMPER method: - S: Substitute - C: Combine - A: Adapt - M: Modify / Magnify / Minify - P: Put to another use - E: Eliminate - R: Reverse / Rearrange The skill is optimized for a continuous, back-and-forth dialogue. It maintains a running "idea model" and iteratively enriches that model with relevant information found on the public web (benchmarks, analogs, constraints, technical options, examples, user pain points, pricing patterns, regulations, etc.). ## 2) When to Use Use this skill when the user has: - A baseline idea they want to improve, differentiate, or repurpose. - A process they want to redesign (reduce steps, cost, time, defects). - A concept that is "stuck" and needs systematic variation. - A need for a structured ideation checklist that reliably produces options. Do NOT use this skill for: - Purely exploratory ideation without any starting subject. - Requests primarily about verifying facts (use a research skill instead). - Problems requiring rigorous root-cause analysis before ideation (do that first). ## 3) Inputs and Outputs Inputs (minimum): - Subject: what is being transformed (one sentence). - Objective: what "better" means (e.g., cheaper, faster, safer, more adoption). - Constraints: must-haves and must-not-do (budget, timeline, compliance, tech). Optional inputs: - Audience / user segment - Current design or workflow steps - Known blockers, risks, and assumptions - Success metrics and target ranges Outputs: - A set of labeled idea variants mapped to SCAMPER letters and prompts. - A short-list (typically 3-7) of best candidates with rationale. - A "next experiments" plan for validating key assumptions. - A consolidated context brief enriched with web findings (sources summarized, not copied). ## 4) Core Operating Principles 1) Coverage: Always run all SCAMPER letters (S, C, A, M, P, E, R) in order unless the user explicitly asks to stop early. 2) Diverge then converge: Generate options first; evaluate later. 3) Traceability: Every generated variant must be tagged with its SCAMPER letter and the exact prompt lens used. 4) Minimal assumptions: Ask for missing constraints early; explicitly mark any assumption. 5) Context enrichment: Use targeted web research to reduce guesswork, expand option space, and de-risk feasibility. 6) Iteration: After each SCAMPER pass, deepen the best candidates with another pass on selected components. ## 5) Conversation Contract (Continuous Dialogue) At any time, the assistant maintains and updates this structured "Idea State": - Subject / baseline - Objective(s) and success metrics - Audience / context of use - Constraints and non-goals - Decomposition: key components / steps / resources - Known pain points / failure modes - Candidate variants (with SCAMPER tags) - Open questions and assumptions - Web findings summary (bullets, with relevance notes) - Next actions / experiments The assistant explicitly confirms updates to the Idea State in compact form after each major step. ## 6) High-Level Flow ### Step 0: Intake and framing Goal: clarify what is being transformed and what "success" means. Ask (as needed): - What is the subject (one sentence) and current version (short description)? - Who uses it and in what context? - What is the primary objective (pick up to 2)? - What constraints are non-negotiable? - What is the time horizon (now, 3 months, 1 year)? Then produce: - A one-paragraph baseline restatement. - A component breakdown (3-10 elements) or process steps (3-15 steps). ### Step 1: First web enrichment (baseline context) Goal: avoid ideating in a vacuum. Run web research to gather: - What already exists (category benchmarks, typical features, standard practices). - Common user complaints / pain points in this domain. - Hard constraints (regulations, safety, standards, platform rules). - Typical cost/time ranges (if relevant). - Adjacent solutions and analogs in other industries. Summarize findings into the Idea State and explicitly note how they impact constraints or opportunity space. ### Step 2: SCAMPER pass (full coverage) Goal: generate transformations systematically. Process: - For each letter: 1) Ask 1-3 targeted questions to the user to ground the prompt in reality. 2) Generate 5-12 variants (or fewer if user requests). 3) Ask the user to pick 1-3 variants to carry forward (or the assistant selects based on stated metrics). 4) Add chosen variants to the short-list and record why. ### Step 3: Second web enrichment (for top variants) Goal: test plausibility and deepen options. For each short-listed variant, research: - Existing implementations / similar patterns - Materials/technologies/vendors/tools that enable it - Adoption drivers and blockers (including UX and behavior) - Risks and edge cases - Rough cost/time/complexity signals Update the Idea State with evidence-based refinements. ### Step 4: Converge + plan experiments Goal: produce actionable next steps. - Rank the short-listed variants by objective fit. - Identify critical assumptions per variant. - Propose 1-3 cheap experiments per variant (prototype, landing page, user test, A/B, pilot, etc.). - Define success criteria per experiment. ### Step 5: Iterate If the user wants more: - Run another SCAMPER cycle on a specific component (not the whole system). - Or combine top ideas into hybrid concepts and re-check constraints. ## 7) SCAMPER Prompt Library (Full Method) ### S - Substitute Intent: replace ingredients, components, people, tools, channels, rules, or resources. Ask the user: - What elements are fixed vs. swappable? - Where do we see high cost, high friction, or high risk? Generate variants by substituting: - Materials, inputs, or data sources - Roles (who does the work), automation vs. human - Channels (web/app/in-person/partner) - Mechanisms/technologies (method A -> method B) - Business model primitives (one-time -> subscription, etc.) - Constraints/rules (change a rule while staying compliant) Output format: - S1..Sn: "Substitute X with Y to achieve Z" ### C - Combine Intent: merge functions, steps, features, teams, or experiences. Ask the user: - What adjacent tasks occur before/after this? - What do users already bundle manually? Generate variants by combining: - Two features into one flow (reduce context switches) - Product + service wrapper - Data streams to unlock personalization - Partner capabilities into a bundle - Multiple steps into a single action Output: - C1..Cn: "Combine X and Y to create Z" ### A - Adapt Intent: borrow from analogs; fit the idea to new contexts or constraints. Ask the user: - What is a similar problem in a different industry? - What patterns do users already understand? Generate variants by adapting: - Proven patterns from other domains (subscriptions, loyalty, freemium, etc.) - Interfaces/metaphors users know - Mechanisms from a known solution (but adapted to constraints) - Workflows optimized for a new environment (mobile-first, low bandwidth, etc.) Output: - A1..An: "Adapt pattern X from domain Y to our subject to accomplish Z" ### M - Modify / Magnify / Minify Intent: change attributes (size, shape, frequency, intensity, scope), amplify or shrink. Ask the user: - Which attribute drives value? Which drives pain? - Where does "more" or "less" likely help? Generate variants by modifying: - Scale up (magnify): add capacity, speed, fidelity, personalization - Scale down (minify): simplify, strip to core, reduce steps/features - Change form factor, packaging, sequencing, UX - Change timing/cadence (batch vs. real-time) - Change quality attributes (durability, security, aesthetics) Output: - M1..Mn: "Modify attribute X from A to B to improve Z" ### P - Put to Another Use Intent: repurpose for new users, contexts, or secondary markets. Ask the user: - Who else has this problem in a different setting? - What byproducts or capabilities are currently unused? Generate variants by repurposing: - New segments (B2B vs B2C, education, healthcare, etc.) - New contexts (offline, emerging markets, enterprise compliance) - Byproduct monetization (data, waste, assets) - Component reuse as a standalone offering Output: - P1..Pn: "Put X to another use for user Y in context Z" ### E - Eliminate Intent: remove elements that are redundant, costly, risky, or confusing. Ask the user: - What do users tolerate but do not value? - Where are the biggest failure points? Generate variants by eliminating: - Steps in a process (remove approvals, reduce handoffs) - Features that add complexity without impact - Dependency on scarce resources - High-risk interactions - Optionality and configuration overload Output: - E1..En: "Eliminate X to reduce Y and improve Z" ### R - Reverse / Rearrange Intent: invert direction, roles, or order; restructure components. Ask the user: - What is the current sequence and why? - What if we start from the end state? Generate variants by reversing/rearranging: - Reverse the user journey (result first, details later) - Swap producer/consumer roles (users generate value for each other) - Reorder steps for earlier feedback loops - Invert constraints (treat constraint as a feature) - Flip push/pull (on-demand vs scheduled) Output: - R1..Rn: "Reverse/rearrange X -> Y to accomplish Z" ## 8) Web Research Protocol (Iterative Context Enrichment) ### When to research Trigger web enrichment when any of the following is true: - The domain is unfamiliar or regulated. - There is uncertainty about typical solutions, costs, or constraints. - A variant depends on specific technologies/materials/vendors. - Validation needs competitor benchmarks or user sentiment. - The user requests "best practices" or "examples". ### What to research (typical buckets) - Baseline market: common feature sets, pricing models, adoption patterns - User voice: reviews, complaints, forums, case studies - Constraints: standards, regulations, safety, platform policies - Enablers: tools, materials, algorithms, suppliers, APIs - Analogs: "how other industries solve this" - Evidence: performance claims, comparative studies (when available) - IP awareness: patents as inspiration (no legal interpretation) ### Query templates - "How do people solve in ?" - " common complaints" / "top features" - " regulation standard requirements" - " alternatives to " - "case study " - "pattern examples" / " best practice" - "patent " (for inspiration) ### How to incorporate findings - Summarize in 3-8 bullets per bucket. - Tie each finding to a decision: constraint, opportunity, or risk. - Explicitly mark confidence (high/medium/low) if evidence is weak. - Never copy large text; only paraphrase and cite/attribute in the conversation. ## 9) Convergence and Evaluation After the full SCAMPER pass: - Cluster duplicates. - Score candidates against the user's success metrics (simple 1-5 scoring). - Flag high-risk assumptions. - Recommend a short-list (3-7) with one-sentence rationale each. Then propose experiments: - Prototype: paper/mockups, clickable demo, service roleplay - User tests: 5-10 interviews with scripted questions - Market test: landing page + ads, waitlist, pricing survey - Operational test: limited pilot with one segment ## 10) Quality Checklist (Must Pass) - All SCAMPER letters used (S, C, A, M, P, E, R). - Each variant tagged with letter + prompt lens. - Clear baseline and constraints recorded. - Web enrichment performed at least twice (baseline + shortlist deepening), unless the user opts out. - Short-list produced with rationale. - Experiments defined with success criteria. - No unmarked assumptions; no invented facts. **Language.** Always respond in the language the user speaks. ## 11) Example (Mini Walkthrough) Subject: "A weekly newsletter for remote product managers" Objective: "Increase retention and referral" Constraints: "No paid acquisition, 1 person team" S: Substitute format (audio, short video), substitute source (curated community), substitute cadence (daily micro-brief). C: Combine with template library, combine with community Q&A, combine with job board. A: Adapt language-learning streak mechanics, adapt "morning briefing" format. M: Minify to 2-minute read, magnify personalization by role/seniority. P: Repurpose as onboarding kit for teams, repurpose as internal enablement. E: Eliminate long essays, eliminate multiple CTAs, eliminate topic breadth. R: Reverse flow (question first -> answers next issue), rearrange content (actionable first). Then research: - retention levers for newsletters, best onboarding sequences, common PM pain points, and comparable newsletters. ## 12) Safety, Ethics, and IP - Do not request or store sensitive personal data. - Do not present copied content as original; summarize and attribute. - Patents and competitor research are for inspiration only; do not provide legal advice. - Respect the user's confidentiality: avoid sharing proprietary details in public searches; use generic queries if needed.