Microsoft Semantic Kernel

Definition

Microsoft's SDK for building AI agents in C# and Python with a plugin-based architecture. Integrates with Azure OpenAI, supports planners for multi-step reasoning, and provides plugin-based tool access. Plugin permissions exist but are static, developer-configured — not dynamically enforced at runtime.

Why It Matters

Plugin permissions are trust-based: developers configure what plugins can do, and the system trusts the model to use them correctly. This is not Zero Trust. Dynamic, runtime enforcement evaluates every action against policy rules based on current state — not on static configuration.

How Exogram Addresses This

Exogram provides execution governance that Semantic Kernel's plugin system lacks. Wrap plugin execution with exogram.evaluate(). Each plugin call gets independently validated by the deterministic policy engine.

Related Terms

medium severityProduction Risk Level

Key Takeaways

  • This concept is part of the broader AI governance landscape
  • Production AI requires multiple layers of protection
  • Deterministic enforcement provides zero-error-rate guarantees

Governance Checklist

0/4Vulnerable

Frequently Asked Questions