maigp-microsoft-agent-framework

AIGP governance for the Microsoft Agent Framework (MAF).

Part of MAIGP — the Mediated AI Governance Protocol: a runtime consent / scope / evidence layer for AI systems. Every governed call is CHECKed before it runs (policy, scope, jurisdiction, circuit-breaker) and RECORDed + TRACEd as tamper-evident evidence after.

What it governs

Governs MAF agents (ChatAgent runs, function tools, and the middleware chain).

Governance model

This adapter maps the framework's lifecycle onto the AIGP agent-governance loop provided by maigp-agent-core:

Install

pip install maigp-microsoft-agent-framework

Pulls in maigp-agent-core (shared lifecycle) and, transitively, maigp-client.

Quick start

from agent_framework import ChatAgent            # MAF
from aigp_microsoft_agent_framework import AigpMafGovernance

gov = AigpMafGovernance(gov_url="https://gov.example.com",
                        app_id="my-maf-agent", hmac_secret="secret")

# Style 1 — wrap a run explicitly:
result = await gov.run(agent, "Analyze this finding", agent_name="analyst")

# Style 2 — register as MAF middleware for automatic coverage:
agent = ChatAgent(chat_client=..., tools=[...], middleware=[gov.middleware()])

Context

MAF 1.0 (GA 2026-04-03) is the converged successor to Semantic Kernel + AutoGen. This adapter supersedes maigp-autogen and maigp-semantic-kernel; prefer it for new projects. The framework object is duck-typed (usage read from the response), so the adapter works across MAF point releases without importing MAF directly.

Links

• Protocol, specs & full adapter catalog: https://github.com/owner-spec/aigp-protocol
• Issues: https://github.com/owner-spec/aigp-protocol/issues

License: Proprietary. © Kanjani AI Research & Causum.