Manage vectorized memory with heat-based decay using PostgreSQL and pgvector.
npx skills add https://github.com/Geralt1983/Thanos --skill memory-v2-skill
Please help me install this Agent Skill. Skill: Memory V2 Skill Source: https://github.com/Geralt1983/Thanos/tree/main/.claude/skills/memory-v2 Command: npx skills add https://github.com/Geralt1983/Thanos --skill memory-v2-skill
This Skill provides a dynamic, heat-based memory system designed to combat information overload and improve focus for individuals with ADHD, ensuring relevant information is surfaced when needed.
whats_hot and whats_cold to quickly identify current focus and neglected items.memory_router for consistent memory operations.Use the memory router to search for memories related to the 'Orlando project status'.
High-intent search queries and answers about installing and using this skill.
Heat-based decay prioritizes recently accessed or important context management information, gradually fading less relevant data over time. This ensures information retrieval surfaces current focus areas while suppressing outdated items automatically.
Vectorized memory storage requires PostgreSQL with pgvector for database infrastructure and OpenAI embeddings for vectorization. You store facts, documents, and context, enabling efficient semantic search through the unified memory router API.
Yes, semantic search finds related project discussions, documents, and client information by querying the vectorized memory system. It automatically surfaces relevant past context, reducing manual searching and keeping you focused on current tasks.
ADHD-specific helpers `whats_hot` and `whats_cold` identify current focus and neglected items respectively. `whats_hot` surfaces actively relevant information, while `whats_cold` highlights neglected data needing attention for improved context management.
Yes, the memory router API provides a unified interface for consistent memory operations while also supporting direct service access. This allows advanced users to bypass the router and perform complex vectorized memory operations directly.
markus41Manage vector embeddings for semantic search across Pinecone, Weaviate, Chroma, and pgvector.
kevinlinSynchronize tasks and memory from local workspace activity into TASKS.md and memory directories.
koasis89Create and manage notes using OMX persistent memory features.
smith-hornPerform vector-based semantic search over document collections with AgentDB.