Skill Creator

Pipeline

Phase 1: Content Extraction

Analyze all source materials and identify 3-8 knowledge modules (logical topic groupings). For each module, extract:

• Module title — clear, concise name
• Key concepts — 3-7 core ideas the MR must understand
• Key facts — specific clinical data, product specs, definitions
• Procedures — step-by-step processes (e.g., how to present a study)
• Common objections — HCP pushback the MR should anticipate
• Assessment criteria — how to measure MR mastery of this module

Phase 2: Learning Design

For each module, define:

1. Learning objectives using Bloom's Taxonomy (Remember -> Analyze)
2. Assessment items:
   • Multiple Choice (3-5 per module): 4 options, one correct, include explanation
   • Scenario-based (1-2 per module): realistic HCP interaction, scoring rubric 1-5
3. Scoring model: per-module scores, overall weighted average, pass threshold 70%

Phase 3: SOP Assembly

Generate a structured Standard Operating Procedure following the SOP Structure Guide provided in the reference file sop-structure-guide.md. The SOP must cover all 5 required stages (Opening, Needs Assessment, Product Discussion, Objection Handling, Closing) with the detail level specified in the guide.

Each SOP step must include:

• title: Step name
• description: What the MR should do
• key_points: Critical messages to deliver
• objections: Possible HCP pushback and suggested responses
• assessment_criteria: How to evaluate MR performance on this step
• knowledge_points: Product/clinical facts needed
• suggested_duration: Recommended time allocation

Phase 4: Skill Assembly

Combine everything into a complete coaching skill with:

• Skill metadata (name, description, product, therapeutic area)
• Complete SOP with all steps from Phase 3
• Knowledge base organized by modules from Phase 1
• Assessment questions and rubrics from Phase 2
• Scoring weights for the 6 standard evaluation dimensions
• Coaching tone and style guidelines

Output Format

Return a JSON object following the schema defined in output-schema.json. The key structure is:

{
  "name": "product-name-training",
  "description": "Comprehensive MR training skill for [Product]",
  "product": "Product Brand Name",
  "therapeutic_area": "Oncology",
  "sop_steps": [
    {
      "title": "Opening",
      "description": "Greet the HCP, establish rapport, set agenda",
      "key_points": ["Professional greeting", "Confirm available time"],
      "objections": [
        {"objection": "I only have 2 minutes", "response": "Focus on key message"}
      ],
      "assessment_criteria": ["Greeting professionalism", "Agenda clarity"],
      "knowledge_points": ["Product indication overview"],
      "suggested_duration": "1-2 minutes"
    }
  ],
  "modules": [
    {
      "title": "Product Fundamentals",
      "objectives": ["Explain mechanism of action", "Cite key efficacy data"],
      "content": "Detailed module content...",
      "questions": [
        {
          "type": "multiple_choice",
          "question": "What is the primary indication?",
          "options": ["A", "B", "C", "D"],
          "correct": 0,
          "explanation": "Approved for..."
        }
      ]
    }
  ],
  "scoring": {
    "pass_threshold": 70,
    "weights": {
      "sop_completeness": 0.20,
      "knowledge_accuracy": 0.25,
      "conversation_logic": 0.20,
      "assessment_coverage": 0.15,
      "difficulty_calibration": 0.10,
      "executability": 0.10
    }
  },
  "summary": "A 2-3 sentence overview of the skill's scope and purpose."
}

Rules

1. Content fidelity — NEVER invent facts not present in the source documents. Quote directly from source when key terminology matters. Mark ambiguous content with [NEEDS_CLARIFICATION].
2. SOP completeness — Include all 5 required SOP stages as defined in sop-structure-guide.md. Each stage must have actionable, specific steps with all required sub-fields.
3. Dimension optimization — Design content that scores well on all 6 evaluation dimensions from scoring-rubric.md: sop_completeness, knowledge_accuracy, conversation_logic, assessment_coverage, difficulty_calibration, and executability.
4. Language matching — Generate output in the SAME LANGUAGE as the source documents. If the source is Chinese, output Chinese text (JSON keys remain English). If multi-language, default to the predominant language.
5. Difficulty balance — Distribute assessment difficulty across Bloom's Taxonomy levels: approximately 30% Remember, 30% Understand, 25% Apply, 15% Analyze and above. Include progressive difficulty within each module.
6. Executability — Ensure all SOP instructions are precise enough for an AI coaching agent to execute automatically. Include clear decision criteria at conversation branch points and explicit guidance for handling unexpected HCP responses.