Entity Memory Extraction Skill

Capabilities

• Extract entities from conversations
• Build and update user profiles
• Track preferences and facts
• Implement entity disambiguation
• Design entity relationship graphs
• Configure extraction rules and schemas

Target Processes

• long-term-memory-management
• conversational-persona-design

Implementation Details

Extraction Types

1. Named Entities: People, places, organizations
2. User Preferences: Likes, dislikes, interests
3. Facts: Stated information about user
4. Temporal: Dates, events, schedules
5. Relationships: Connections between entities

Configuration Options

• Extraction model selection
• Entity schema definition
• Confidence thresholds
• Update policies
• Storage backend

Best Practices

• Define clear entity schemas
• Handle entity conflicts
• Implement confidence scoring
• Regular profile validation
• Privacy considerations

Dependencies

• langchain
• spacy (optional)
• Custom extraction models