Build planner-driven agent workflows with Microsoft Semantic Kernel
Compose prompts, plugins, memory, planners, and connectors into repeatable Python, .NET, or Java agent workflows with Microsoft Semantic Kernel.
npx skills add agentskillexchange/skills --skill build-planner-driven-agent-workflows-with-microsoft-semantic-kernel
Use Microsoft Semantic Kernel when an agent or operator needs to turn a one-off assistant idea into a repeatable orchestration workflow: define native functions or plugins, connect model providers, add memory/vector-store integrations, and coordinate single-agent or multi-agent plans across business systems. Invoke this skill when the desired output is a maintainable workflow scaffold or orchestration pattern, not when a user simply wants to chat with a model or browse Semantic Kernel documentation. The scope boundary is implementation planning for agent orchestration using Semantic Kernel primitives—tools/plugins, planners, memory, connectors, and multi-runtime setup—not a general Microsoft product card, generic SDK listing, or blanket recommendation to adopt the framework.
What this skill actually does
Inputs and prerequisites: Microsoft Semantic Kernel; Python 3.10+, .NET 10.0+, or Java 17+ depending on runtime; LLM provider credentials as required by the chosen connector.
Setup notes: Install the Semantic Kernel package for the target runtime, then configure a model connector and define plugins/functions, memory, and planner or agent orchestration components following the official documentation.
Source and verification boundary: use https://learn.microsoft.com/semantic-kernel/ as the canonical reference before running the workflow; keep commands, API calls, CLI usage, and generated outputs reviewable against that upstream source.
Framework fit: publish this as a Multi-Framework workflow only when the operator can invoke the documented toolchain directly, rather than treating the upstream project as a generic product listing.