Expand memory entries via semantic vector seeds and graph-neighbor recall.
npx skills add https://github.com/spqian/dreamweave --skill graph-recall
Please help me install this Agent Skill. Skill: graph-recall Source: https://github.com/spqian/dreamweave/tree/main/skills/graph-recall Command: npx skills add https://github.com/spqian/dreamweave --skill graph-recall
This Skill addresses the limitations of flat memory lookups by providing a graph-based search and neighbor expansion for enhanced memory recall.
Run the graph-recall skill with the query 'last week's project updates'.
High-intent search queries and answers about installing and using this skill.
Graph-based memory recall improves flat lookups by using semantic vector seeds and neighbor expansion to find connected information. It provides a comprehensive retrieval experience by mapping relationships between people, systems, and incidents.
The three-tiered memory structure models data using instinct, recall, and archive tiers. This approach manages complex memory structures efficiently, enabling in-depth analysis and targeted data retrieval for analytical workflows.
To retrieve connected project updates, execute a query using graph-based vector search. The system expands memory entries via semantic seeds and graph neighbors, returning comprehensive and related context.
Use graph search instead of flat memory lookups when querying complex relationships involving people, systems, incidents, or dates. It is designed for administrative and analytical workflows requiring in-depth, connected memory analysis.
No, graph memory recall operates without external dependencies. It leverages an internal three-tiered approach and scripts to process complex memory structures and deliver enhanced vector search results.
Neighbor expansion is limited to analyzing structured memory tiers of instinct, recall, and archive. It is designed specifically for administrative and analytical workflows rather than unstructured flat data retrieval.
ovachieverGenerate Gemini embeddings and index them for vector search.
812lclVisualize Readwise highlights and their cross-source connections in a 2D graph.
CommunityToolkitImplement three-tier Hot, Cold, and Wiki memory for Squad agents.
shinprIngest local documents and query them with hybrid semantic and keyword search.
