cognitive-memory

To enable persistent memory for multi-agent systems, you need centralized memory management that stores and indexes long-term knowledge across sessions. This Skill automates recall, archiving, and governance using modular memory stores to maintain reliable agent context.

The best way to manage long-term memory routing and retrieval is through a routing classifier that directs knowledge to core, episodic, semantic, procedural, or vault stores, enabling deterministic recall and archiving for multi-agent workflows.

Yes, you can maintain memory hygiene in multi-agent workflows by applying a decay-based forgetting model. This model automates the removal of outdated knowledge, while reflection cycles and an audit trail ensure proper governance of retained memory.

Yes, this memory governance approach supports knowledge graph indexing by categorizing data into semantic, procedural, and episodic layers. It provides deterministic retrieval and archiving across multiple agents and sessions.

To bootstrap a centralized memory system in your workspace, you initialize the memory stores and start the routing and reflection cycles. This setup enables agents to coordinate identity, context, and learned workflows across sessions.