Your Knowledgebase Should Learn and Grow

AI agents should feed discoveries and decisions back into the knowledge base during task execution. When an agent learns something valuable—like a required environment variable or a working remediation step—that knowledge should be persisted so every future interaction benefits. The knowledge base should be treated as a living, evolving system rather than a static document loaded once at startup.