MindForgeAgent

1. Overview

MindForge is a multi-agent collaborative system specifically designed for emotion analysis and journal processing. The system is managed by a central coordinator (Supervisor Agent) that orchestrates multiple specialized agents to perform tasks such as:

• Emotional diary analysis
• Mood tracking
• Theme extraction
• Report generation

By decomposing complex tasks into smaller, specialized components, MindForge efficiently processes user requests and provides deep emotional insights and personalized analysis.

---

2. System Architecture

MindForge employs a centralized coordination with distributed execution architecture:

             +-------------------+
             | supervisor_agent  |
             | (Task Coordinator)|
             +-------------------+
                      |
      +---------------+---------------+
      |               |               |

+-----------------+ +------------+ +---------------+ | summarize_agent | | emotion_ | | topic_agent | | (diary | | agent | | (Theme | | Summary) | | (Emotion | | Analysis) | +-----------------+ | Analysis) | +---------------+ +------------+ | | | +------------------+ | report_agent | | (Report | | Generation) | +------------------+

---

3. Agents Introduction

3.1 Supervisor Agent

The Supervisor Agent serves as the central coordinator of the system and is responsible for:

1. Analyzing the user's original request
2. Determining which specialized agents are needed
3. Coordinating workflows between agents
4. Integrating outputs and returning final results

3.2 Emotion Agent

The Emotion Agent conducts emotional analysis based on Plutchik's Wheel of Emotions. It is capable of:

1. Providing quantitative scores for 8 basic emotion dimensions
2. Detecting complex emotion combinations and intensity levels
3. Tracking emotional trends over time
4. Offering detailed emotional explanations

3.3 Topic Agent

The Topic Agent specializes in theme and keyword extraction. It performs:

1. Theme analysis across documents
2. Keyword and concept extraction
3. Topic clustering and classification
4. Foundation building for word clouds or knowledge graphs

3.4 Report Agent

The Report Agent synthesizes all analysis results and generates comprehensive reports by:

1. Transforming raw data into structured formats
2. Generating data visualizations (charts, graphs, etc.)
3. Providing insights and recommendations
4. Supporting various report formats

4. Workflow Example

User Request:

"I need to write an emotional journal about how meditation helps reduce anxiety, and analyze my emotional state from the past week."

Step-by-step Process:

1. Supervisor Agent analyzes the request and delegates tasks to:

   • summarize_agent → process the journal
   • emotion_agent → analyze emotional states
   • topic_agent → extract themes
   • report_agent → generate the report

2. User writes the journal

3. Agents execute:

   • summarize_agent → creates journal summary
   • emotion_agent → evaluates emotions in the entry
   • topic_agent → extracts terms like “meditation”, “anxiety”

4. For past emotional records:

   • emotion_agent → analyzes logs from the past week
   • topic_agent → identifies weekly recurring themes

5. Final stage:

   • report_agent integrates all findings into a polished report
   • supervisor_agent returns the final report to the user

5. Technical Implementation Notes

The codebase includes JSON-based functional definitions for each agent, which contain:

1. Agent name and unique identifier
2. Function descriptions
3. Expected input parameters
4. Usage examples

These agent definition files enable the supervisor_agent to:

• Discover available agents
• Understand their capabilities
• Properly invoke them with structured input