Chapter 2. New features and enhancements

This release improves the installation program functionality in several ways:

This enhancement introduces a new agentic interface that transforms the trusted software supply chain (tssc) CLI into a first-class model context protocol (MCP) server. This enhancement enables large language model (LLM)-driven workflows for platform engineering, allowing AI agents to orchestrate automated, self-service installations, configurations, and diagnostics on OpenShift clusters. By leveraging MCP, the server provides structured tools and guidance to LLMs, promoting efficient, stateful interactions while integrating with Kubernetes-native mechanisms for scalability and maintainability.

This architecture draws on Mixture of Experts (MoE) principles in LLMs, where specialized "experts" handle distinct tasks via a router guided by system prompts, enabling efficient scaling without proportional computational resource utilization. This means the MCP server acts as a workflow orchestrator, managing installation states and tool calls to ensure idempotent, low-frequency operations, which is ideal for initial Red Hat Advanced Developer Suite setups.

In this release, for customized pipelines in GitLab the following changes apply: