Build Paths Advanced
Scaffold an agent system where Oracle AI DB is the *only* state store, composed from...
Install
npx skills add oracle-devrel/oracle-ai-developer-hub --skill build-paths-advanced -g -a claude-code -y- Repository
- oracle-devrel/oracle-ai-developer-hub4.5k
- Category
- Design
- Updated
- 5d ago
- License
- See repository
- Context cost
- Mediumabout 3,000 tokens
- Runs scripts
- No, instructions only
What it does
Scaffold an agent system where Oracle AI DB is the *only* state store, composed from the build-paths/skills/ building-block library. Stack, langchain-oracledb + oracle-database-mcp-server + in-DB ONNX embeddings + OCI GenAI Grok 4 + Open WebUI. Three projects, production-feeling NL2SQL+RAG hybrid analyst, self-improving research agent, conversational schema designer. For users who want a real DB-as-only-store agent demo.
The user picked the advanced path. Two non-negotiable rules at this tier:
- Oracle AI DB is the only state store. No Redis, no Postgres, no SQLite, no Chroma/FAISS/Qdrant/Pinecone, no JSON or pickle on disk for runtime state.
verify.pygrepssrc/for forbidden imports and fails the build. - You compose, you don't write boilerplate. The Oracle layer comes from
skills/oracle-aidb-docker-setup,skills/langchain-oracledb-helper,skills/oracle-mcp-server-helper. Your job is to invoke those, then write the application logic (the agent loop, the memory adapters, the tools, the adapter, the notebook). ~500-700 LOC of project code per idea, not 1500.
Step 0, Read these references first
skills/README.md, the composition pattern.skills/oracle-aidb-docker-setup/SKILL.mdskills/langchain-oracledb-helper/SKILL.mdskills/oracle-mcp-server-helper/SKILL.mdshared/references/oamp.md, load-bearing for the conversational + per-user durable memory layer (ideas 1 + 2).- All of
shared/references/, yes all of them. The advanced tier can touch every feature. advanced/project-ideas.md.
Step 1, Interview
Run shared/interview.md plus the advanced-only questions.
