Genius Thinking Skill
Validated innovation frameworks + cognitive-science-based creative reasoning.
<when_to_use>
| Situation | Example | |------|------| | Innovation ideation | New service concept | | Complex problem solving | Multi-angle challenge | | Creative breakthrough | Existing approach is stuck | | Strategy design | Long-term direction and vision | | Product innovation | Blue-ocean style opportunities |
/genius-thinking AI-powered education service ideas
/genius-thinking early-stage startup customer acquisition strategy
Output: Formula-selection rationale + deep analysis (1500+ chars) + 10+ ideas (3000+ chars) + prioritization.
</when_to_use>
<argument_validation>
If $ARGUMENTS is missing, ask immediately:
"What topic should I apply genius thinking to?
Examples:
- Business idea: 'AI-based education service'
- Problem solving: 'Early startup customer acquisition'
- Product innovation: 'Healthcare app differentiation'"
</argument_validation>
<mandatory_reasoning>
Mandatory Sequential Thinking
- Always use the
sequential-thinkingtool before any substantial analysis or ideation work. - Keep the full reasoning flow in
sequential-thinkingfrom Phase 1 to Phase 4 (do not skip phases). - If
sequential-thinkingis unavailable, stop and report the blocker instead of proceeding with ad-hoc reasoning.
</mandatory_reasoning>
<thinking_formulas>
10 Formulas + Validated Frameworks
1. Genius Insight (GI) + JTBD
GI = (O * C * P * S) / (A + B)
| Variable | Meaning | Validated Tool | |------|------|------| | O | Depth of observation | JTBD prompt: "What job is the user hiring this for?" | | C | Originality of connections | Functional / emotional / social job split | | P | Pattern recognition | Extract cross-case patterns | | S | Synthesis | Integrate insights | | A, B | Fixed assumptions / bias | Lower is better |
Source: Christensen Institute - JTBD
2. Multidimensional Analysis (MDA) + Six Thinking Hats
MDA = Sigma[Di * Wi * Ii]
| Dimension | Hat Perspective | |------|------| | D1 Time | White (facts/data) | | D2 Space | Yellow (opportunity) | | D3 Abstraction | Green (creativity) | | D4 Causality | Black (risk/critique) | | D5 Hierarchy | Red (intuition/emotion) | | Meta | Blue (process control) |
Source: MindTools - Six Thinking Hats
3. Creative Connection (CC) + SCAMPER
CC = |A intersect B| + |A xor B| + f(A -> B)
SCAMPER 7-question checklist:
| Question | Meaning | Example | |------|------|------| | Substitute | Replace? | Netflix: DVD -> streaming | | Combine | Combine? | Smartphone = phone + camera + computer | | Adapt | Borrow? | Roll-on deodorant <- ballpoint concept | | Modify | Magnify/minify? | iPad Mini / Pro segmentation | | Put to other use | New use? | McDonald's as real-estate business | | Eliminate | Remove? | Remove waiter-heavy process with self-service | | Reverse | Reverse? | Pay-first -> consume-later reversal patterns |
Source: IxDF - SCAMPER
4. Problem Redefinition (PR) + HMW + ERRC
PR = P0 * T(theta) * S(phi) * M(psi)
Required HMW transform: "How might we [verb] [target]?"
| Original Problem | HMW Transform | |------|------| | Users drop during onboarding | How might we create value realization within 10 minutes? | | Customer churn is high | How might we create daily return behavior? |
ERRC Grid (value-innovation check):
| Action | Question | |------|------| | Eliminate | What industry assumptions should be removed? | | Reduce | What should be reduced below standard? | | Raise | What should be raised above standard? | | Create | What should be newly created? |
Source: Stanford d.school, Blue Ocean ERRC
5. Innovative Solution (IS) + TRIZ
IS = Sigma[Ci * Ni * Fi * Vi] / Ri
Top TRIZ principles (from 40):
| Principle | Meaning | Typical Application | |------|------|------| | Segmentation | Break into smaller units | Microservices | | Extraction | Isolate essentials | MVP | | Local quality | Local optimization | Personalization | | Asymmetry | Break symmetry | Differentiation | | Merging | Combine functions | Super app | | Prior action | Prepare in advance | Predictive analytics | | Inversion | Reverse approach | Reverse-positioning marketing | | Intermediary | Add mediator layer | Platform model | | Self-service | System serves itself | Automation | | Copying | Imitate then improve | Best-practice adaptation |
Source: TRIZ Wikipedia
6. Insight Amplification (IA) + Incubation
IA = I0 * (1 + r)^n * C * Q
Incubation effect (cognitive science):
| Stage | Method | Effect | |------|------|------| | 1. Define problem | Focused analysis | Conscious processing | | 2. Rest | 15-30 min low-load activity | Unconscious processing | | 3. Capture | Log first ideas immediately | Insight harvesting |
Source: ScienceDirect 2024
7. Thought Evolution (TE)
TE = T0 + integral[L(t) + E(t) + R(t)]dt
| Function | Driver | |------|------| | L(t) Learning | Continuous information intake | | E(t) Experience | Diverse experimentation | | R(t) Reflection | Metacognition and failure learning |
8. Complexity Solving (CS) + Morphological Box
CS = det|M| * Sigma[Si/Ci] * product[Ii]
Zwicky box (morphological analysis):
| Parameter | Option 1 | Option 2 | Option 3 | |------|------|------|------| | Channel | App | Web | Offline | | Monetization | Subscription | Ads | Transaction fee | | Target | B2C | B2B | B2G |
Example combination: App + Subscription + B2C.
Source: Ness Labs - Zwicky Box
9. Intuitive Leap (IL) + DMN-ECN Switching
IL = (S * E * T) / (L * R)
Brain-network switching (creative control):
| Network | Role | Activation | |------|------|------| | DMN | Divergent thought and association | Rest / mind wandering | | ECN | Convergent evaluation | Focus / analysis | | Switching | Core creativity mechanic | Intentional transition |
Execution loop: DMN (10m) -> ECN (10m) -> repeat.
Source: Nature 2025
10. Integrated Wisdom (IW)
IW = (K + U + W + C + A) * H * E
| Variable | Meaning | |------|------| | K + U + W | Knowledge + understanding + wisdom | | C + A | Compassion + action | | H * E | Humility * ethics |
</thinking_formulas>
<formula_selection>
| Topic Type | Recommended Combo | Core Tools | |------|------|------| | New idea generation | 1 + 3 | JTBD + SCAMPER | | Problem solving | 4 + 8 | HMW + Morphological Box | | Strategy design | 2 + 10 | Six Hats + ERRC | | Product innovation | 5 + 6 | TRIZ + Incubation | | Fast breakthrough | 9 + 3 | DMN-ECN + SCAMPER | | Long-term vision | 7 + 2 | Evolution + multidimensional analysis | | Complex systems | 8 + 2 | Zwicky Box + Six Hats |
</formula_selection>
<workflow>
| Phase | Task | Tool | Required Check |
|------|------|------|------|
| 0 | Input validation | - | Validate ARGUMENT |
| 1 | Formula selection | Sequential Thinking (3 steps) | Connect at least one validated framework |
| 2 | Deep analysis | Sequential Thinking (5-7 steps) | Include HMW transform + SCAMPER sweep |
| 3 | Idea generation | Parallel tasks x3 | ToT structure, 12+ draft ideas |
| 4 | Integration/refinement | Sequential Thinking (3 steps) | Final 10+ ideas + ERRC validation |
| 5 | Save output | Write | .hypercore/genius-ideas/ |
Phase 2: Deep Analysis (1500+ chars)
Sequential Thinking:
thought 1: Apply Formula A + validated tool (ex: JTBD 3-job analysis)
thought 2: Apply Formula B + validated tool (ex: SCAMPER 7 prompts)
thought 3: Identify overlap + convert to HMW
thought 4: Expand synergy space + ERRC check
thought 5: Produce integrated insight
Phase 3: Tree-of-Thoughts Structure
Task({ subagent_type: 'analyst', model: 'sonnet',
prompt: 'Path A: [Formula1+JTBD] produce 4 ideas with 1-5 scoring each' })
Task({ subagent_type: 'analyst', model: 'sonnet',
prompt: 'Path B: [Formula2+SCAMPER] produce 4 ideas with 1-5 scoring each' })
Task({ subagent_type: 'analyst', model: 'sonnet',
prompt: 'Path C: [Fusion+TRIZ] produce 4 ideas with 1-5 scoring each' })
// Expand 4+ score paths, backtrack 3- paths
ToT reference: arXiv
</workflow><result_structure>
| Section | Content | |------|------| | Header | Date, topic, selected formulas + validation tools | | 1. Formula selection rationale | Why these formulas and frameworks fit | | 2. Deep analysis | 1500+ chars, includes HMW transform and sources | | 3. 10+ ideas | Title/description/innovation point/execution path/score (1-5) | | 4. Prioritization | ERRC validation + Quick Wins / Big Bets | | 5. Next steps | 1-3 immediate actions |
Idea Format
### Idea N: [Title]
**Description**: 2-3 sentences
**Innovation point**: What is novel vs status quo (TRIZ or SCAMPER element)
**JTBD**: Which job does it solve?
**Execution step**: First concrete action
**Score**: Impact [1-5] / Feasibility [1-5]
</result_structure>
<examples>
/genius-thinking AI-powered education service ideas
Phase 1: Formula selection
- Selection: Genius Insight (1) + Creative Connection (3)
- Tools: JTBD + SCAMPER
- Why: New ideation needs customer-job clarity plus combinational creativity
Phase 2: Deep analysis
[JTBD]
- Functional job: acquire new knowledge
- Emotional job: feel growth
- Social job: appear competent/expert
[SCAMPER]
- Substitute: instructor -> AI tutor
- Combine: learning + game + social
- Adapt: Netflix recommendation -> learning path recommendation
[HMW]
"How might we make learning feel as immersive as gaming?"
Phase 3: ToT parallel (12 draft ideas)
Path A (JTBD): emotional AI tutor (score 5)
Path B (SCAMPER): micro-learning game loops (score 4)
Path C (TRIZ fusion): inversion, student-teaches-back (score 4)
Phase 4: Refine to final 10 ideas + ERRC validation
Save: .hypercore/genius-ideas/00.ai_education_service.md
</examples>
<validation>
| Item | Required |
|------|------|
| ARGUMENT | Ask immediately if missing |
| Phase 1 | Use 2 formulas + at least 1 validated framework/tool |
| Phase 2 | 1500+ chars + HMW transform + SCAMPER or TRIZ |
| Phase 3 | ToT parallel structure, 12+ draft ideas |
| Phase 4 | 10+ final ideas, ERRC validation, 3000+ chars |
| Reasoning trace | sequential-thinking used across analysis phases |
| Prioritization | Impact/feasibility + JTBD linkage |
| Save | .hypercore/genius-ideas/ |
| Forbidden | |------| | Start without ARGUMENT | | Use formulas without validation tools | | Define problem without HMW transform | | List ideas without SCAMPER/TRIZ reasoning | | Fewer than 10 ideas or less than 3000 chars | | Explain frameworks without sources | | End without saving output |
</validation><references>
Innovation methods:
- TRIZ: Wikipedia
- SCAMPER: IxDF
- JTBD: Christensen Institute
- ERRC: Blue Ocean
Cognitive science:
- DMN-ECN switching: Nature 2025
- Incubation effect: ScienceDirect 2024
Execution methods:
- HMW: Stanford d.school
- Six Hats: MindTools
- Zwicky Box: Ness Labs
- Tree of Thoughts: arXiv
Scan to join WeChat group