Search EDITH memory layers with hybrid vector and full-text retrieval.
npx skills add https://github.com/knsiuss/orion --skill memory-search-knsiuss
Please help me install this Agent Skill. Skill: memory-search Source: https://github.com/knsiuss/orion/tree/main/workspace/skills/memory-search Command: npx skills add https://github.com/knsiuss/orion --skill memory-search-knsiuss
This Skill allows you to search through EDITH's entire memory, including past conversations, decisions, and factual information, to retrieve specific details or context from any point in time.
Use the memory-search skill to find out when we last discussed the project's budget.
High-intent search queries and answers about installing and using this skill.
To recall past conversations and facts in EDITH, the memory-search skill conducts deep searches across vector stores, episodic logs, and pinned facts, retrieving specific past decisions and contextual information.
Hybrid retrieval for memory logs works by combining semantic vector similarity search with exact phrase matching using FTS5, returning ranked and grouped results for accurate contextual recall.
Yes, you can retrieve a specific project discussion from months ago using vector search. The skill searches across episodic logs and vector stores to pinpoint exact past conversations and relevant contextual details.
Supported memory layers for context retrieval include vector stores, episodic logs, pinned facts, user profiles, and feedback records, enabling comprehensive deep search across all available EDITH memory data.
The best way to find when a specific topic was last discussed in episodic logs is to use the memory-search skill, which employs hybrid retrieval to scan and rank past contextual conversations by relevance.
Yes, exact phrase matching works alongside semantic similarity for memory recall. The skill utilizes a hybrid retrieval mechanism combining FTS5 full-text search and vector similarity to ensure accurate ranked results.
designnotdrumStore structured memories with facts, reasons, sentiment, and implications.
abhattacherjeeSearch Obsidian vault notes by keyword, tag, or frontmatter fields.
markus41Implement JWT, OAuth2, sessions, and RBAC for authentication and authorization.
LGuguiImplement semantic vector search with AgentDB for document retrieval.
