memory-summarization

# Memory Summarization Skill ## Capabilities - Implement conversation summarization strategies - Configure rolling summary updates - Design hierarchical summarization - Implement token-aware summarization - Create extractive and abstractive summaries - Design summary quality evaluation ## Target Processes - conversational-memory-system - long-term-memory-management ## Implementation Details ### Summarization Strategies 1. **Rolling Summary**: Update summary with new messages 2. **Hierarchical**: Multi-level summarization 3. **Token-Budget**: Fit within token limits 4. **Extractive**: Key message selection 5. **Abstractive**: LLM-generated summaries ### Configuration Options - LLM for summarization - Summary token budget - Update frequency - Summary template - Quality thresholds ### Best Practices - Balance detail vs compression - Preserve key information - Monitor summary quality - Test with long conversations - Handle context window limits ### Dependencies - langchain-core - LLM provider