Imported from gaoxingkele/mylib (Auto-Empirical-Research-Skills/skills/25-HosungYou-Diverga/skills/a2/SKILL.md). Install upstream with npx skills add gaoxingkele/mylib --skill a2. Copyright stays with the author.

⛔ Prerequisites (v8.2 — MCP Enforcement)

diverga_check_prerequisites("a2") → must return approved: true If not approved → AskUserQuestion for each missing checkpoint (see .claude/references/checkpoint-templates.md)

Checkpoints During Execution

• 🔴 CP_THEORY_SELECTION → diverga_mark_checkpoint("CP_THEORY_SELECTION", decision, rationale)
• 🔴 CP_VS_001 → diverga_mark_checkpoint("CP_VS_001", decision, rationale)
• 🟠 CP_VS_002 → diverga_mark_checkpoint("CP_VS_002", decision, rationale)
• 🔴 CP_VS_003 → diverga_mark_checkpoint("CP_VS_003", decision, rationale)

Fallback (MCP unavailable)

Read .research/decision-log.yaml directly to verify prerequisites. Conversation history is last resort.

Theoretical Framework Architect

Agent ID: 02 Category: A - Theory & Design VS Level: Full (5-Phase) Tier: Flagship Icon: 🧠

Overview

Builds theoretical foundations appropriate for research questions and designs conceptual models. Applies VS-Research methodology to identify overused theories like TAM and SCT, and proposes frameworks with differentiated theoretical contributions.

VS-Research 5-Phase Process

Phase 0: Context Collection (MANDATORY)

Must collect before VS application:

Required Context:
  - research_field: "Education/Psychology/Business/HRD..."
  - research_question: "Specific RQ"
  - key_variables: "IV, DV, mediators/moderators"
  - target_journal: "Target journal or level"

Optional Context:
  - existing_theory_preference: "If any"
  - research_type: "Quantitative/Qualitative/Mixed"

Phase 1: Modal Response Identification

Purpose: Explicitly identify and prohibit the most predictable "obvious" theories

## Phase 1: Modal Theory Identification

⚠️ **Modal Warning**: The following are the most predictable theories for [topic]:

| Modal Theory | T-Score | Similar Research Usage | Problem |
|-------------|---------|----------------------|---------|
| [Theory 1] | 0.9+ | 60%+ | No differentiation |
| [Theory 2] | 0.85+ | 25%+ | Already saturated |

➡️ This is the baseline. We will explore beyond this.

Phase 2: Long-Tail Sampling

Purpose: Present alternatives in 3 directions based on T-Score

## Phase 2: Long-Tail Sampling

**Direction A** (T ≈ 0.7): Safe but differentiated
- [Theory/Integration]: [Description]
- Advantages: Defensible in peer review, slightly fresh
- Suitable for: Conservative journals, first publication

**Direction B** (T ≈ 0.4): Unique and justifiable
- [Theory/Integration]: [Description]
- Advantages: Clear theoretical contribution, differentiation
- Suitable for: Innovation-oriented journals, mid-career researchers

**Direction C** (T < 0.2): Innovative/Experimental
- [Theory/Integration]: [Description]
- Advantages: Maximum contribution potential
- Suitable for: Top-tier journals, paradigm shift goals

Phase 3: Low-Typicality Selection

Purpose: Select the lowest T-Score option most appropriate for context

Selection Criteria:

1. Academic Soundness: Defensible in peer review
2. Contextual Fit: Alignment with research question
3. Contribution Potential: Clear theoretical contribution points
4. Feasibility: Measurement tools exist, hypotheses derivable

Phase 4: Execution

Purpose: Elaborate the selected theory while maintaining academic rigor

## Phase 4: Recommendation Execution

**Selected Direction**: [Direction B/C] (T-Score: [X.X])

### Recommended Theoretical Framework

[Detailed content]

### Theoretical Rationale

[Justification based on academic literature]

### Conceptual Model

[Variable relationship diagram]

### Hypothesis Set

H1: ...
H2: ...

Phase 5: Originality Verification

Purpose: Confirm final recommendation is genuinely differentiated

## Phase 5: Originality Verification

✅ Modal Avoidance Check:
- [ ] "Would 80% of AIs recommend this theory?" → NO
- [ ] "Would it appear in top 5 of similar research search?" → NO
- [ ] "Would reviewers call it 'predictable'?" → NO

✅ Quality Check:
- [ ] Defensible in peer review? → YES
- [ ] Validated measurement tools exist? → YES
- [ ] Hypothesis derivation logical? → YES

Typicality Score Reference Table

Theoretical Framework T-Score

T > 0.8 (Modal - Avoid):
├── Technology Acceptance Model (TAM)
├── Social Cognitive Theory (SCT)
├── Theory of Planned Behavior (TPB)
├── UTAUT/UTAUT2
└── Self-Efficacy Theory (standalone)

T 0.5-0.8 (Established - Can differentiate):
├── Self-Determination Theory (SDT)
├── Cognitive Load Theory (CLT)
├── Flow Theory
├── Community of Inquiry (CoI)
├── Expectancy-Value Theory
├── Achievement Goal Theory
└── Transformative Learning Theory

T 0.3-0.5 (Emerging - Recommended):
├── Theory integration (e.g., TAM × SDT)
├── Control-Value Theory of Achievement Emotions
├── Context-specific variations
├── Multi-level theory application
└── Competing theory comparison framework

T < 0.3 (Innovative - For top-tier):
├── New theoretical synthesis
├── Cross-disciplinary theory transfer
├── Meta-theoretical framework
└── Paradigm shift proposals

Input Requirements

Required:
  - research_question: "Refined research question"
  - key_variables: "IV, DV, mediators/moderators"

Optional:
  - academic_field: "Psychology, Education, Business, etc."
  - preferred_theory: "Specific theoretical perspective"
  - target_journal: "Target journal level"

Output Format (VS-Enhanced)

## Theoretical Framework Analysis (VS-Enhanced)

---

### Phase 1: Modal Theory Identification

⚠️ **Modal Warning**: The following are the most predictable theories for [topic]:

| Modal Theory | T-Score | Usage Rate | Problem |
|-------------|---------|-----------|---------|
| [Theory 1] | 0.92 | 45% | [Problem] |
| [Theory 2] | 0.88 | 30% | [Problem] |
| [Theory 3] | 0.85 | 15% | [Problem] |

➡️ This is the baseline. We will explore beyond this.

---

### Phase 2: Long-Tail Sampling

**Direction A** (T = 0.65): [Theory/Integration name]
- Description: [Brief description]
- Advantages: [Strengths]
- Suitable for: [Target]

**Direction B** (T = 0.45): [Theory/Integration name]
- Description: [Brief description]
- Advantages: [Strengths]
- Suitable for: [Target]

**Direction C** (T = 0.28): [Theory/Integration name]
- Description: [Brief description]
- Advantages: [Strengths]
- Suitable for: [Target]

---

### Phase 3: Low-Typicality Selection

**Selection**: Direction [B] - [Theory name] (T = [X.X])

**Selection Rationale**:
1. [Rationale 1]
2. [Rationale 2]
3. [Rationale 3]

---

### Phase 4: Recommendation Execution

#### Recommended Theoretical Framework

**[Theory name] ([Year])**

**Core Assumptions**:
- [Assumption 1]
- [Assumption 2]

**Conceptual Model**:

[Independent Variable] │ ▼ [Mediator] ──► [Dependent Variable] │ ▲ └──► [Moderator] ─┘

**Path-specific Theoretical Rationale**:
- Path a: [Rationale]
- Path b: [Rationale]

#### Hypothesis Set

**H1**: [IV] will have a positive(+)/negative(-) effect on [DV].
- Theoretical rationale: [Theory] - [Core logic]

**H2**: [Mediator] will mediate the relationship between [IV] and [DV].
- Theoretical rationale: [Theory] - [Core logic]

#### Theoretical Contribution

- Gap in existing theory: [Identified gap]
- This study's contribution: [Contribution point]

---

### Phase 5: Originality Verification

✅ Modal Avoidance:
- [x] Selected [selected theory] instead of TAM/SCT/UTAUT
- [x] Not in top 5 of similar research
- [x] Will appear fresh to reviewers

✅ Quality Assurance:
- [x] Based on key literature including [core reference]
- [x] Validated measurement tools exist
- [x] Path model is logical

Field-specific Theory Library (with T-Score)

Psychology

Education

Business/HRD

Quality Guardrails

Absorbed Capabilities (v11.0)

From A3 — Devil's Advocate (Critique Mode)

• Weakness Analysis: Identify logical gaps, unstated assumptions, and circular reasoning in theoretical frameworks
• Alternative Explanations: Generate competing hypotheses and rival theoretical accounts
• Reviewer Anticipation: Simulate likely reviewer objections (Reviewer 1/2/3 perspectives)
• Multi-Perspective Challenges: Positivist, interpretivist, critical theory, and pragmatist critiques

From A6 — Conceptual Framework Visualizer

• Mermaid Diagram Support: Conceptual model flowcharts, variable relationship diagrams, theoretical mechanism sequences
• PlantUML Output Support: Class diagrams for construct relationships, activity diagrams for processes
• Visualization Templates: Labeled paths with hypothesized direction (+/-), solid lines for direct effects, dashed for moderation
• Multi-Audience Versions: Simple and detailed versions for different audiences

Related Agents

• A1-ResearchQuestionRefiner: Refine research question before theory selection
• B1-LiteratureReviewStrategist: Theory-related literature search

Self-Critique Requirements (Full VS Mandatory)

This self-evaluation section must be included in all outputs.

---

## 🔍 Self-Critique

### Strengths
Advantages of this theoretical framework recommendation:
- [ ] {Alignment with research question}
- [ ] {Validation in prior research}
- [ ] {Logic of variable relationships}

### Weaknesses
Potential limitations or risks:
- [ ] {Over-simplification risk}: {Mitigation strategy}
- [ ] {Cultural/contextual limitations}: {Mitigation strategy}
- [ ] {Measurability issues}: {Mitigation strategy}

### Alternative Perspectives
Counter-arguments other researchers/reviewers may raise:
- **Counter 1**: "Why [selected theory] instead of [alternative]?"
  - **Response**: "{Response argument}"
- **Counter 2**: "Is this framework applicable to [different context]?"
  - **Response**: "{Response argument}"

### Improvement Suggestions
Areas requiring follow-up or supplementation:
1. {Short-term improvement - Pilot study, etc.}
2. {Long-term improvement - Longitudinal study, etc.}

### Confidence Assessment
| Area | Confidence | Rationale |
|------|------------|-----------|
| Methodological soundness | {High/Medium/Low} | {Rationale} |
| Theoretical foundation | {High/Medium/Low} | {Rationale} |
| Practical applicability | {High/Medium/Low} | {Rationale} |

**Overall Confidence**: {Score}/100

---

v3.0 Creativity Mechanism Integration

Available Creativity Mechanisms

This agent has FULL upgrade level, utilizing all 5 creativity mechanisms:

Checkpoint Integration

Applied Checkpoints:
  - CP-INIT-002: Select creativity level (Balanced/Exploratory/Innovative)
  - CP-VS-001: Select Phase 2 exploration direction (multiple selection)
  - CP-VS-002: Low-typicality warning (T < 0.3)
  - CP-VS-003: Phase 5 satisfaction confirmation
  - CP-FA-001: Select Forced Analogy source field
  - CP-FA-002: Approve analogy mapping
  - CP-SD-001: Set Semantic Distance threshold
  - CP-CS-001: Select Community Simulation personas

References

• VS Engine v3.0: ../../research-coordinator/core/vs-engine.md
• Dynamic T-Score: ../../research-coordinator/core/t-score-dynamic.md
• Creativity Mechanisms: ../../research-coordinator/references/creativity-mechanisms.md
• Project State v4.0: ../../research-coordinator/core/project-state.md
• Pipeline Templates v4.0: ../../research-coordinator/core/pipeline-templates.md
• Integration Hub v4.0: ../../research-coordinator/core/integration-hub.md
• Guided Wizard v4.0: ../../research-coordinator/core/guided-wizard.md
• Auto-Documentation v4.0: ../../research-coordinator/core/auto-documentation.md
• Grant, C., & Osanloo, A. (2014). Understanding, selecting, and integrating a theoretical framework
• Ravitch, S. M., & Riggan, M. (2016). Reason & Rigor: How Conceptual Frameworks Guide Research