Saved in:
| Main Author: | |
|---|---|
| Format: | Recurso digital |
| Language: | |
| Published: |
Zenodo
2026
|
| Online Access: | https://doi.org/10.5281/zenodo.18525121 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866901796222402560 |
|---|---|
| author | Rahul Jain |
| author_facet | Rahul Jain |
| contents | <p>The emergence of autonomous agent systems driven by large language models has brought about the need to have secure architectural frameworks that can help to balance operational autonomy with organizational control. The Model Context Protocol is an abstract base layer that allows standard interactions between intelligent agents and enterprise infrastructure and ensures security boundaries and governance concerns. Multi-agent systems that involve dedicated computing agents show greater capabilities in complex task performance under the collaborative workflow but present serious problems concerning the prevention of unauthorized access, policy compliance, and the maintenance of regulatory compliance. The security controls, such as sandboxed execution practices, access controls, attribute-based authorization systems, and multi-layered defense measures, all create protective barriers to the emergent risks related to the autonomous system behaviors. With extensive audit infrastructure coupled with security information and event management platforms, real-time use of anomalies and the ability to perform forensic analysis are available that are critical in establishing enterprise trust. The issue of scalability requires advanced orchestration, resource allocation, and transaction management solutions that are distributed in nature and support heterogeneous enterprise infrastructure. New modalities in workflow graph representations, secure memory architectures, and the ability to work in the few-shot learning regime provide avenues to more autonomous and yet manageable agent systems that can provide support to mission-critical organizational functions and yet stay within security posture and compliance requirements.</p> |
| format | Recurso digital |
| id | zenodo_https___doi_org_10_5281_zenodo_18525121 |
| institution | Zenodo |
| language | |
| publishDate | 2026 |
| publisher | Zenodo |
| record_format | zenodo |
| spellingShingle | Secure Multi-Agent MCP Architectures: A Framework For Enterprise AI Governance Rahul Jain <p>The emergence of autonomous agent systems driven by large language models has brought about the need to have secure architectural frameworks that can help to balance operational autonomy with organizational control. The Model Context Protocol is an abstract base layer that allows standard interactions between intelligent agents and enterprise infrastructure and ensures security boundaries and governance concerns. Multi-agent systems that involve dedicated computing agents show greater capabilities in complex task performance under the collaborative workflow but present serious problems concerning the prevention of unauthorized access, policy compliance, and the maintenance of regulatory compliance. The security controls, such as sandboxed execution practices, access controls, attribute-based authorization systems, and multi-layered defense measures, all create protective barriers to the emergent risks related to the autonomous system behaviors. With extensive audit infrastructure coupled with security information and event management platforms, real-time use of anomalies and the ability to perform forensic analysis are available that are critical in establishing enterprise trust. The issue of scalability requires advanced orchestration, resource allocation, and transaction management solutions that are distributed in nature and support heterogeneous enterprise infrastructure. New modalities in workflow graph representations, secure memory architectures, and the ability to work in the few-shot learning regime provide avenues to more autonomous and yet manageable agent systems that can provide support to mission-critical organizational functions and yet stay within security posture and compliance requirements.</p> |
| title | Secure Multi-Agent MCP Architectures: A Framework For Enterprise AI Governance |
| url | https://doi.org/10.5281/zenodo.18525121 |