cncf
[Initiative]: MCP Server Authentication and Authorization Standards - Whitepaper
Name
MCP Server Authentication and Authorization Standards
Short description
Development of a whitepaper exploring authentication and authorization considerations and reference patterns for Model Control Protocol (MCP) servers in cloud native environments
Responsible group
TOC
Does the initiative belong to a subproject?
Yes
Subproject name
TOC Artificial Intelligence Initiatives
Primary contact
Satarupa Deb (GitHub: @Satarupa22-SD, email: [email protected])
Additional contacts
Yujia Lin (GitHub: @blueandhack, email: [email protected])
Initiative description
Objective: The objective of this initiative is to explore comprehensive authentication and authorization considerations for Model Control Protocol (MCP) servers, examining security patterns relevant to AI model deployment and management infrastructure. As AI systems become increasingly distributed and integrated into enterprise environments, understanding standardized security mechanisms for MCP servers warrants careful evaluation. This initiative builds upon the growing adoption of MCP servers in production environments and examines security frameworks that may be compatible with existing cloud native security approaches. The work will focus on documenting practical, implementable security patterns that organizations may consider when deploying MCP infrastructure within cloud native ecosystems.
Key areas of exploration include: Authentication mechanisms that organizations might evaluate for MCP server access (multi-factor, service-to-service, token-based approaches) Authorization frameworks including RBAC and ABAC considerations for model operations Integration patterns that could be evaluated with existing cloud native security infrastructure Threat modeling approaches specific to MCP server environments Monitoring and auditing capabilities for compliance scenarios Considerations for secure deployment and operational management The initiative will produce a framework documenting security considerations that address the unique challenges of AI model serving infrastructure while examining compatibility with existing security patterns commonly found in cloud native environments
Deliverable(s) or exit criteria
1. Whitepaper (≤ 10 pp): "MCP Security Framework: Authorization Considerations for Cloud-Native AI Systems"
- Documentation of security considerations, reference patterns, and evaluation criteria
- Coverage of authentication mechanisms, authorization frameworks (RBAC/ABAC), threat modeling approaches, and integration patterns organizations might consider for MCP servers in cloud native environments
2. Reference patterns:
- Kubernetes-compatible examples for secure MCP server deployment scenarios
- Integration patterns organizations might evaluate with ingress controllers, service mesh security, SPIFFE/SPIRE identity, Open Policy Agent (OPA) authorization, Sigstore code signing, and OpenTelemetry (OTel) observability components
- Documentation presented as reference implementations rather than prescriptive requirements
3. Community alignment:
- Review with TAG-Security for security considerations validation
- Cross-working group coordination with CNCF AI initiatives and adjacent workstreams to ensure alignment with broader cloud native AI security approaches
The objective is to conclude deliverables by December 2025 (3-month development timeline).
Tracking document for meeting and progress
TBD
I agree the involvement of TAG Security and Compliance would help here. I'd like to be involved.
@JustinCappos Would TAG Security and Compliance agree to be the home for this Initiative?
Hi I also would like to get involved and contribute to this initiative :) edit: I am in AEST timezone (GMT+10) I’m interested in contributing as an author as well as a reviewer.
I'd be interested in contributing! (EST timezone, open to either being an author for a section or reviewing!)
Hi everyone,
Thank you for your interest in contributing to this project. It would be great if you could please mention in what capacity you would be interested in collaborating ( Author / Reviewer), also please specify your timezone, it would help us to decide the timings for the weekly meetings.
Kindly note that the tentative duration for this project is from Oct - Dec.
I'd like to contribute as an author (open to also helping review), PST timezone
I'd be happy to serve as a TAG Security and Compliance leadership representative for this work. US Central timezone.
I’m interested in contributing as an author as well as a reviewer in the areas I’m interested in. I’m in the IST timezone, thanks!
I would like to contribute to this, too. Time zone is JST. I will be a little busy from October to December, so I may mainly contribute as a reviewer.
@JustinCappos Would TAG Security and Compliance agree to be the home for this Initiative?
We've discussed and given the interest of @jkjell @y-tabata and others, we certainly could be the home.
Thanks @JustinCappos ! As soon as the TAG Security and Compliance charter merges, the TAG can add this to their board. No vote required. cc: @riaankleinhans
Approved this initiative at the 1 Oct 2025 TAG-SC North America meeting.
Initiative Lead: @Satarupa22-SD TAG-SC Liason: @jkjell
Stuff that you get with a TAG liason:
- They can create meetings on the CNCF calendar through PCC
- They can create projects, labels, and manage issues in the TOC repo to track sub-elements of the initiative
- You can get a directory e.g. under https://github.com/cncf/toc/tree/main/tags/tag-security-and-compliance/initiatives
You probably want to also create a dedicated slack channel for the initiative, e.g. #initiative-mcp-authX-whitepaper
@Satarupa22-SD if you want to coordinate a time for the first meeting, I can set it up on the TAG Security and Compliance calendar. If it's easier too, we can start a 🧵 on Slack in #tag-security-and-compliance for scheduling.
I’m interested in contributing as an author as well as a reviewer. IST/JST timezones.
Thank you all for your interest in joining this project! We have created a Slack channel to help us coordinate and communicate. I have already added some members, but I wasn’t able to find everyone. If you haven’t been added yet, please feel free to join using the link below: mcp-authentication-standards-whitepaper
@Satarupa22-SD which slack workspace? I tried that link in the CNCF slack and got the following
@JonZeolla Probably because the channel is private now. I have added you now.
Hi, if anyone is unable to join the Slack channel please dm me or Yujia on slack, we will add you. Thanks!
@Satarupa22-SD Hello, I’m interested in contributing as a reviewer. Time zone is JST.
Gentle Reminder : Please connect with Yujia or me, if you are not added on the slack channel yet. Thanks!
@Satarupa22-SD could you also add or invite me to the slack channel pls? I don't think I have an account there.
@Satarupa22-SD could you also add or invite me to the slack channel pls? I don't think I have an account there.
Hi @jeremychoi I did not find you in the CNCF Slack workspace. I sent an invitation to your GitHub profile email address. Could you please check it out? After you register, I will add you. Or you can use any email address to register by the link: https://communityinviter.com/apps/cloud-native/cncf
Short description
Development of a Whitepaper exploring authentication and authorization considerations and reference patterns for Model Control Protocol (MCP) servers in cloud native environments
I assume this initiative is about Model Context Protocol https://modelcontextprotocol.io/ ?
