Recalls, stores, and searches persistent memories across agent sessions.
npx skills add https://github.com/QuixiAI/Hexis --skill core-memory
Please help me install this Agent Skill. Skill: core-memory Source: https://github.com/QuixiAI/Hexis/tree/main/skills/installed/core-memory Command: npx skills add https://github.com/QuixiAI/Hexis --skill core-memory
LLM agents lose all context between sessions, forcing users to repeat themselves and preventing any real continuity. This Skill gives the agent durable memory operations so it can recall past conversations, store important facts, and maintain goals and commitments over time.
recall for targeted semantic retrieval and search_history for exact names, phrases, or details from earlier sessions, with recall_at_time and diff_memory_history for point-in-time and change-over-time queries.remember, and manage goals, schedules, backlogs, and watch-and-notify responsibilities with dedicated management tools.search_documents and search_document_chunks to open_document and desk tools to cite ingested source material directly.Ask the agent to recall what was discussed in previous sessions about a specific topic and store any new decisions from the current conversation.
High-intent search queries and answers about installing and using this skill.
Use the remember tool to store durable facts, events, preferences, and decisions, then retrieve them later with recall for semantic queries or search_history for exact phrases. The agent checks memory before claiming it does not know something.
Use search_history with the exact name, phrase, or detail you need from earlier sessions. It works even when semantic recall is weak or embeddings are unavailable, making it the fallback for precise lookups.
Recall performs targeted semantic retrieval and works best with specific queries about topics or concepts. Search_history finds exact strings from prior sessions, so use it for names, quotes, and precise details.
Yes, recall_at_time answers what was known as of a past instant, and diff_memory_history shows what changed between two points in time. This supports auditing how beliefs and memories evolved.
Do not store secrets unless the user explicitly asks, and avoid persisting every minor sentence. Memory should hold facts, promises, and decisions that will matter later, keeping tool use proportional to the task.
SalesTeamToolboxRetrieve and persist cross-task knowledge using the memory MCP tool.
QuixiAIAudit memory, goals, and identity during heartbeat cycles.
TylerGarlickTrack and audit belief changes with AI influence detection.
aisa-groupBuild retrieval-augmented generation systems with embeddings, vector stores, and reranking.