Skip to content
Kitbay
Skill

agent-memory-systems

MITUpdated on Nov 17, 202649,917GitHub stars

A skill in the agentic-awesome-skills collection, adapted from the Apache-licensed vibeship-spawner-skills project, for designing memory, retrieval and retention for AI agents.

What the skill teaches

A skill in the agentic-awesome-skills collection, adapted from the Apache-licensed vibeship-spawner-skills project, for designing memory, retrieval and retention for AI agents. It sorts memory into facts and preferences, events and reusable procedures, and asks that every stored record be scoped to an authenticated tenant or user, kept with its source, timestamp and version, and retained only with permission and within limits. The agent confirms scope and existing stores first, chooses an adapter the installed SDK supports, specifies conflict and versioning rules, and tests isolation between users, stale records, bounded retrieval and failed writes before delivering a design and change plan.

Capabilities

  • Design semantic, episodic and procedural memory
  • Scope stored memory to a tenant or user
  • Set retention, conflict and versioning rules
  • Test isolation and bounded retrieval