Claude Code and Oracle AI Database: agent tutorial

This tutorial presents Limitless, a learning and research workflow built around Claude Code and Oracle AI Database. The presenter uses a terminal conversation to explore agent memory, then inspects the stored learning records with SQL. Following along requires access to Claude Code and an Oracle database; the session points to a repository with setup instructions.

The workflow separates research, tutoring and assessment into skills. An AI researcher gathers material into a reference document. XP Builder asks diagnostic questions and prompts the learner to explain concepts in their own words. XP Debrief evaluates the conversation and saves evidence of strengths and gaps to the database. The presenter describes this division as a way to keep each AI agent task manageable, while warning that generated research still needs checking.

Oracle AI Database holds session history, concept assessments and vector embeddings, with LangChain Oracle DB connecting the workflow to storage. The demonstrated database runs on OCI. The presenter also describes a local container option, but says repository instructions for it still need an update.

Obsidian is an optional front end for scores, review suggestions and links between concepts. The content framework builder remains a roadmap item rather than a demonstrated feature. Audience questions cover persistent memory, skill portability, command-line access and token usage.