Exam Prep — Exam Preparation Full Pipeline

Workflow

Phase 1: Preparation (Orchestrator performs directly)

1. Extract from user input:
   • Exam name: Which exam is being prepared for (college entrance, civil service, certification, TOEIC, etc.)
   • Subject/area: All subjects or a specific subject
   • Current level (optional): Existing scores, mock exam results, self-assessment
   • Goal (optional): Target score, pass threshold
   • Remaining time (optional): Time until the exam
   • Existing data (optional): Score reports, error notebooks, study history
2. Create a _workspace/ directory at the project root
3. Organize the input and save it to _workspace/00_input.md
4. If existing data is provided, copy it to _workspace/ and adjust the relevant phase accordingly
5. Determine the execution mode based on the scope of the request (see "Task-Scale Modes" below)

Phase 2: Team Assembly and Execution

Inter-agent communication flow:

• trend-analyst completes -> sends frequently tested areas and trap types to diagnostician; sends predicted topics to learning-designer; sends frequency and difficulty data to examiner
• diagnostician completes -> sends weak areas, weakness types, and goal gap to learning-designer; sends weakness-focused item requests to examiner
• learning-designer completes -> sends study scope and difficulty guide to examiner
• examiner completes -> sends question sheet, answer key, and item design intents to error-analyst
• error-analyst completes -> feeds remediation items back to learning-designer; sends next-exam reflection items to examiner

Phase 3: Integration and Final Report

1. Verify all files in _workspace/
2. Confirm that the error analysis remediation suggestions are reflected in the study plan
3. Report the final summary to the user:
   • Trend analysis — 01_trend_analysis.md
   • Weakness diagnosis — 02_diagnosis_report.md
   • Customized study plan — 03_learning_plan.md
   • Mock exam question sheet — 04_mock_exam.md
   • Mock exam answer key — 04_mock_exam_answer.md
   • Error analysis — 05_error_analysis.md

Task-Scale Modes

Existing data usage: If the user provides score reports, existing error notebooks, etc., store the data in _workspace/ and augment the relevant phase's analysis.

Data Transfer Protocol

File naming convention: {order}_{agent}_{deliverable}.{extension}

Error Handling

Test Scenarios

Normal Flow

Prompt: "Help me prepare for the AWS Solutions Architect certification. I have one month left, and I passed the Cloud Practitioner with 75%." Expected Results:

• Trend analysis: Most recent 3-year exam trends, frequently tested domains
• Diagnosis: Weakness areas inferred from 75% Cloud Practitioner score + diagnostic items provided
• Study plan: 30-day schedule, weakness-area-focused curriculum
• Mock exam: Certification-format mock questions (scenario-based + multi-select)
• Error analysis: Pattern analysis + remediation plan

Existing File Flow

Prompt: "Analyze my mock exam results" + error data attached Expected Results:

• Switches to Error Analysis Mode (error-analyst only)
• Pattern analysis based on the attached data
• Remediation study plan proposal

Error Flow

Prompt: "Help me study for an exam" Expected Results:

• Exam not specified -> ask which exam
• Confirm target score / pass threshold
• After minimum information acquired, proceed with full pipeline

Agent Extension Skills

Extension skills that enhance each agent's domain expertise: