SAP AI Skills for AI Coding Assistants

SAP AI Core is SAP's enterprise-grade service for running AI and machine learning workloads on SAP BTP. It provides infrastructure for generative AI deployments, orchestration workflows, retrieval-augmented generation (RAG) scenarios, and ML training pipelines. Through the Generative AI Hub, developers can access supported foundation models through SAP-managed APIs, subject to tenant, region, and commercial availability.

Building against this stack means understanding several layers: the AI Core service itself (deployment, configuration, content filtering, data masking), the SDKs that wrap its APIs, and the orchestration service that handles templating, grounding, and tool calling. Coding assistants without this context often hallucinate SDK methods, confuse the JavaScript and Python API surfaces, and miss the service-plan-specific limitations that affect production deployments.

These skills are built for the AI coding harnesses you already write code in: Claude Code, ChatGPT and Codex, Cursor, GitHub Copilot, and Windsurf, plus the other agent-based editors supported by skills.sh. One distinction worth knowing: SAP's own assistant, Joule, has its own skills ecosystem — SAP's AI Skills Library lists SKILL.md-format skills for Joule and for coding assistants alike — while this catalog ships through skills.sh for your harness. The guide to the SAP AI Skills Library explains the overlap, and if you are evaluating Joule for SAP work, the article on orchestrating Joule with Claude Code shows how the two complement each other.

The AI skills in this category cover the SAP AI development surface. The sap-ai-core skill guides deployment of generative AI models, orchestration workflows, RAG implementations, prompt registry management, and content filtering. The sap-cloud-sdk-ai skill integrates the SAP Cloud SDK for AI into JavaScript/TypeScript and Java applications, covering chat completion, embeddings, streaming, function calling, and LangChain/Spring AI integration. The sap-cloud-sdk-ai-python skill provides the equivalent for Python via the sap-ai-sdk-gen library, and local-machine-learning tooling such as sap-rpt1 rounds out the set for prediction experiments on SAP data.

Together they let your assistant — whichever harness you run — work on SAP's AI services through documented SDK and service patterns instead of inventing APIs that fail at runtime.