Semantic Kernel

In this catalog it is the Frameworks option when you need to compose multi-step agent loops and tool handoffs. Keep Semantic Kernel on the observe/propose side of the loop: useful signals and drafts, with human or policy authority still required before anything irreversible runs.

Semantic Kernel embeds LLMs into conventional apps via plugins/skills, planners, and memory connectors (.NET / Python / Java).

1. Register plugins (skills) and memory connectors.
2. Planner or prompt invokes plugins.
3. Memory read/write via connectors when configured.
4. Return results into the host application flow.

Related: autogen · microsoft-agent-framework · topics/12-central-nervous-system

• Use when: Semantic Kernel (Microsoft) embeds LLMs into conventional applications via plugins/skills, planners, and memory connectors across matches the bottleneck and you specifically need to compose multi-step agent loops and tool handoffs.
• Skip when: another pack already covers the need with less ops surface, or Semantic Kernel would become a second place people retype truth.
• Prefer instead: compare Autogen, Microsoft Agent Framework before locking the cell.

• Ops: confirm license, hosting, and release cadence for Semantic Kernel (microsoft/semantic-kernel) before relying on it in a live corridor.
• Integration: bad credentials, scopes, or schema drift produce empty retrievals or runaway tool/model spend without an obvious UI alarm.
• Scope: Frameworks components support agents; they are not the company ledger or the approval authority.