build-paths

Scaffolding an Oracle AI Database RAG chatbot requires selecting a difficulty level, which routes you to the correct build-path sub-workflow for your project, ensuring the right app shape and safety constraints.

Beginner build paths target RAG chat demos, intermediate paths build NL2SQL tool-calling agents with in-database ONNX embeddings, and advanced paths create DB-as-only-store agent systems.

Yes, setting up an Oracle AI Database project requires Docker and Python 3.11 or higher, alongside OCI Generative AI bearer-token credentials, before the router dispatches you to a scaffold workflow.

Ask your coding agent to run the build-paths router, select the intermediate difficulty level, provide your target directory, and confirm when prompted to enter the NL2SQL tool-calling agent sub-workflow.

No, the build-paths router does not generate project files itself; it dispatches you to the matching SKILL.md sub-workflow that handles the actual file generation for your Oracle AI Database project.

You should not use this router when requested capabilities fall outside the v1 scope or when prerequisites like Docker, Python 3.11+, and OCI Generative AI bearer-token credentials are missing, as it enforces strict decision guardrails.