Store and retrieve persistent memories across agent sessions via API.
npx skills add https://github.com/ncz-os/mnemos --skill mnemos-memory
Please help me install this Agent Skill. Skill: mnemos-memory Source: https://github.com/ncz-os/mnemos/tree/main/integrations/openclaw/.agents/skills/mnemos-memory Command: npx skills add https://github.com/ncz-os/mnemos --skill mnemos-memory
MNEMOS Memory provides a persistent memory service accessible across agent sessions. It enables reading and writing memories to capture decisions, context, patterns, and standards, with full-text and semantic search and support for organizing knowledge via category and knowledge-graph triples.
Store a memory about a recent decision, then retrieve it later by category using the mnemos memory API.
High-intent search queries and answers about installing and using this skill.
To persist agent context across sessions, you need a memory store that captures decisions and standards. This Skill provides an API to store and retrieve persistent memories using full-text search and knowledge-graph triples.
A knowledge graph triple is a relational fact structure used to organize memories. It allows persistent memory services to store relational facts alongside categorized content, enabling agents to retrieve complex project context and architectural decisions.
To store architectural decisions for long-running tasks, use a memory API to save content with a specific category like decisions. You can retrieve these memories later through full-text or semantic search by filtering for that category.
Yes, persistent memory across agent sessions requires an API to store content, categories, subcategories, and metadata. The API must support retrieving these memories through text and category search and managing knowledge-graph triples.
Yes, you can search memories by category and text content. The persistent memory store supports full-text and semantic search, allowing you to filter memories by categories like solutions, patterns, infrastructure, projects, and standards.
Knowledge graph triples for agent memory are limited to storing relational facts. They must be paired with categorized content and metadata to fully capture complex project context, and retrieval depends on available full-text and semantic search capabilities.
gujincheng1128Provide persistent memory and recall for AI agents across conversations.
jmylchreestSearch stored memories and decisions to answer questions about past context.
forranySave and load cross-session context under .claude/memory/sessions.
roy2392Retrieve long-term memory via vector search and recursive graph traversal on SQLite.
