AI/ML

Multi-Agent Reference Architecture

Reference architecture for multi-agent AI systems spanning user interaction, agent gateway, agent runtime, knowledge, LLM, MCP, evaluation, and observability layers.

Version	Release Manager	Authors	Controls	Threats	Capabilities
DEV	DB Development Build	DT Development Team	23	32	25

Mapping References

FINOS AI Governance Framework0.2.0

AIGF mitigations expressed as Gemara guidelines; each control implements one of these guidelines.

AI Governance Framework Risk Vectors0.2.0

FINOS AIGF risks expressed as Gemara vectors; the grounding source for every threat in this catalog.

Multi-Agent Reference Architecture Capability Catalog1.0.0

Subsystem-level capabilities of the multi-agent reference architecture that threats endanger and controls protect.

Multi-Agent Reference Architecture Threat Catalog1.0.0

Architecture threats that controls in this catalog mitigate.

Contribute to the Next Release

This catalog is maintained as versioned YAML files. Generated artifacts are published here as each release is cut.

View on GitHub →