remember

The remember skill is designed to save user-provided information into long-term memory storage for future retrieval. It addresses the need to preserve important insights, decisions, facts, or learnings across sessions when a user explicitly asks to "remember this" or "save this." Instead of relying on temporary conversation context, the skill extracts the essential information and stores it using a connected memory system, making it easier to reference important knowledge later.

The skill analyzes the content the user wants to preserve, identifies the core insight or fact, and prepares it for storage. It generates a set of searchable concepts consisting of specific, lowercased keyword phrases that improve future retrieval. It can also identify and record any referenced file paths, whether absolute or repository-relative. The skill then saves the full content, associated concepts, and file references through the memory storage tool and confirms the successful save operation to the user, including the concepts used for indexing.

This skill is useful for developers, AI agents, researchers, knowledge workers, and anyone using long-running workflows that benefit from persistent memory. Common use cases include saving project decisions, documenting important discoveries, preserving technical notes, tracking implementation details, and storing user preferences or other information intended to remain available across future interactions. It is especially valuable in environments where long-term memory functionality is provided through the connected agentmemory infrastructure.