rhoai-gen-ai-playground

name	rhoai-gen-ai-playground
metadata	{"author":"rhoai3-demo","version":"1.0.0","platform-family":"rhoai","platform-baseline":"repo","ocp-baseline":"repo","skill-group":"RHOAI Platform"}
description	Use when documenting, reviewing, or operating Red Hat OpenShift AI Gen AI studio playground workflows: enabling the playground dashboard feature, preparing AI asset endpoints, configuring project playgrounds, testing models, comparing models side by side, using playground RAG uploads and RAG chunk settings, saving reusable prompts in MLflow, configuring and testing MCP servers, creating custom internal or external endpoints, exporting playground Python templates, updating or deleting playgrounds, and troubleshooting playground response, RAG, or MCP behavior. Do NOT use for Llama Stack server/provider/vector-store configuration (use rhoai-llama-stack), model-serving runtime configuration (use rhoai-model-serving-platform), dashboard-wide feature flag authoring beyond playground-related fields (use rhoai-dashboard-customization), custom app chatbot code (use rhoai-chatbot-customization), NeMo/FMS Guardrails product safety controls (use rhoai-guardrails-safety), or live cluster changes without the OpenShift safety guard.

RHOAI Gen AI Playground

Use this skill for OpenShift AI Gen AI studio playground work on the active product baseline in docs/PLATFORM_BASELINE.md.

Source Grounding

Read references/source-capture.md before using product workflow details. Official Red Hat documentation is product authority. This skill adapts the official gen AI playground guide to this repo's demo workflow and GitOps review model.

Scope

This skill covers:

• Gen AI studio playground Technology Preview support posture
• dashboard and project prerequisites for playground use
• OdhDashboardConfig fields specific to Gen AI studio and custom endpoints
• AI asset endpoints for project models, custom endpoints, MaaS models, and MCP servers
• model and vLLM runtime requirements for RAG and MCP tool-calling behavior
• configuring a playground from Gen AI studio -> Playground or AI asset endpoints
• creating internal and external custom endpoints
• model experimentation, temperature, streaming, system instructions, and side-by-side comparison
• playground RAG uploads, inline vector database limitations, and chunking settings
• reusable prompt save/load workflows backed by MLflow
• MCP server configuration, browser-session authorization, and tool testing
• exporting playground configuration as a Python template
• updating or deleting project playgrounds
• troubleshooting playground response, RAG, and MCP issues

Use other skills for adjacent work:

• rhoai-llama-stack for Llama Stack Operator activation, LlamaStackDistribution, providers, vector stores, Responses API behavior, OAuth/ABAC, CA trust, and HA/autoscaling
• rhoai-autorag for Gen AI studio AutoRAG optimization runs, remote Milvus prerequisites, RAG pattern leaderboard review, and generated notebooks
• rhoai-model-serving-platform for KServe, vLLM, model-serving runtimes, runtime arguments, and model deployment settings
• rhoai-model-deployment for deploying project models as AI asset endpoints, routes, token authentication, and runtime-specific endpoint smoke tests
• rhoai-dashboard-customization for broader OdhDashboardConfig behavior and dashboard feature flags
• rhoai-model-catalog-workflows for catalog discovery and deployment into an AI asset endpoint before playground testing
• rhoai-model-management-monitoring for operating deployed models behind AI asset endpoints
• rhoai-mlflow for MLflow availability, workspace access, and prompt persistence storage checks
• rhoai-users-groups-access and rhoai-access-group-selection for user and group access prerequisites
• rhoai-chatbot-customization for custom demo chatbot application code, prompts, and UI behavior outside the product playground
• rhoai-guardrails-safety for product guardrails, detector services, validation-only checks, and guarded generation around playground endpoints
• rhoai-evaluation for official EvalHub, LM-Eval, and automated risk assessment workflows beyond exploratory playground testing
• rhoai-model-evaluation for legacy repo-specific RAGAS, custom judge, and Step 08 evaluation rebuild details
• rhoai-maas-governance for governed MaaS endpoint access and external model provider access

Demo Policy

For this repo:

• Treat Gen AI studio playground, AI asset endpoints, custom endpoints, multi-model comparison, and prompt management as Technology Preview unless the active baseline changes.
• Use the playground to demonstrate exploratory model selection and prompt/RAG iteration, not as formal evaluation evidence by itself.
• Prefer project-local AI asset endpoints for private model demos.
• For external OpenAI or other third-party endpoints, document that Responses API data, RAG context, MCP tool results, and user input can leave the cluster.
• Keep endpoint tokens and provider API keys in Kubernetes Secrets; never commit secret values.
• Do not claim the playground RAG upload path supports external or remote vector databases. The official guide says this path uses an inline vector database.
• Before relying on RAG or MCP behavior, verify the selected model supports tool calling and that the vLLM runtime has the required tool-calling arguments.
• Treat exported Python as a starting template, not a runnable application.
• Warn users that updating a playground deletes the inline vector database for all users in the project.
• Warn users that deleting a playground removes it for all users in the project.

Workflow

1. Confirm the active baseline in docs/PLATFORM_BASELINE.md.
2. Read references/source-capture.md and references/official-doc-extraction.md.
3. Decide whether the task is:
   • cluster or dashboard prerequisite review
   • AI asset endpoint review
   • custom endpoint enablement
   • playground creation or model selection
   • RAG upload and chunking
   • reusable prompt management
   • MCP server setup or testing
   • export, update, delete, or troubleshooting
4. Use examples/gen-ai-playground-patterns.md for focused review patterns.
5. Validate with references/validation-checklist.md.

References

• references/source-capture.md
• references/official-doc-extraction.md
• references/validation-checklist.md
• examples/gen-ai-playground-patterns.md