Daniel Walsh examines Googles Open Knowledge Format, a portable structure for Markdown knowledge bundles inspired by the LLM wiki pattern. He uses a coding agent to generate a linked knowledge layerAI agent memory is stored information that an agent can retrieve and use across steps, sessions, or changing contexts. for an existing application.
Walsh explains how indexes, links and references help agents navigate a bundleContext engineering designs how relevant instructions, evidence, state and tool information are assembled for a model's task. and access current sources. He compares declarative knowledge with procedural agent skillsAn AI agent skill is a reusable package of instructions, resources, and tool guidance for performing a bounded kind of work. and argues that curated wikis and retrieval can complement each otherRetrieval is the process of selecting relevant stored information and returning it to an AI system for the current task..
The video examines producer-agent validation and cautions that open formatting does not standardize the meaning of entity types or remove the need to maintain and review stale knowledge.
Watch on YouTube
