MCP and Agent Skills Move to Vendor-Neutral Governance Under Linux Foundation
What happened
According to the Google Open Source Blog: November 2026, Google reported that two foundational agentic AI specifications are being co-maintained under vendor-neutral governance. These are MCP (the protocol that lets AI agents connect to external tools and data sources) and Agent Skills (the mechanism by which agents load sets of procedural instructions). Both are now under the Linux Foundation's Agentic AI Foundation. The shift takes both specifications out of single-vendor control and places them under a multi-party stewardship model. Changes to the rules governing what agents can do and what they can connect to will now be determined by a broader group of contributors. Specification changes can alter agent permissions, authentication methods, and tool access boundaries. No single vendor will issue a conventional product update or patch notice for these changes. Teams that have deployed MCP-connected agents or integrated Agent Skills into production workflows will need to monitor the Agentic AI Foundation's release cycle. It should be treated as a monitored dependency in change management and security review programs.
Why it matters
- ·Protocol governance moving to a multi-party body means specification changes affecting what agents can access or do may arrive faster and from more directions. Organizations with no process for monitoring standards body output will miss changes that require security re-testing or control updates before continued deployment is safe.
- ·Existing security reviews and vendor due diligence assessments for MCP-connected agents were performed against a specification controlled by a single vendor. Under multi-party governance, the approval basis for those deployments may shift. This creates a gap that procurement and risk teams need to close by updating their vendor and tool assessment criteria.
- ·Regulators reviewing agentic AI deployments, including under frameworks such as the Five Eyes Guidance on the Careful Adoption of Agentic AI Services, expect organizations to demonstrate that agents operate within defined and auditable boundaries. A protocol governance change is a material event that affects the evidence base for those controls. It should be reflected in the AI model registry and incident readiness programs.
Stay ahead of stories like this
Get every Global AI governance development like this one, plus the rest of the week's developments. Every Thursday.
Recent issues
- The next quarter will test whether compliance programs can keep pace with AI systems that breach containment, reshape capital markets, and enable fraud at a scale no prior control framework anticipated.12 Nov
- State-linked attackers and criminal operators are actively weaponizing AI tools as entry points into enterprise systems, turning internal platforms and open-source frameworks into live attack vectors.5 Nov
Free every Thursday. Unsubscribe anytime.
Governance controls affected
What to do now
- ☐Identify every AI agent in your deployment that connects to external tools or data sources using MCP, and confirm whether Agent Skills are loaded in any of those agents, so you know which systems are affected by governance changes at the protocol level.
- ☐Assign a named owner to monitor the Linux Foundation's Agentic AI Foundation for specification releases and require that owner to trigger a security review of affected agent deployments before any updated protocol version is adopted in production.
- ☐Review existing security approvals for MCP-connected agents and document whether those approvals referenced the previous single-vendor specification. Update the approval basis to reflect multi-party governance and set a timeline for re-assessment.
- ☐Ask your security or engineering team whether your current MCP server inventory (the list of tool connection points agents can reach) is complete and whether each server has been reviewed against the current specification version.
- ☐Update procurement and vendor contract templates to require disclosure of MCP or Agent Skills specification version dependencies, so your team is notified when a vendor adopts a new version that may change agent behavior or permissions.
What to watch next
Compliance teams should monitor the Linux Foundation's Agentic AI Foundation for its first specification releases and any associated security advisories. Multi-party governance bodies can iterate on specifications faster than enterprise change management cycles can follow. The 253 MCP Server Bugs and Six Framework Bypasses Expose Agent Authorization Gap and 15,000 Public MCP Servers Found Vulnerable, Abandoned, or Sovereignty-Risky incidents already show that MCP infrastructure carries significant unmanaged risk. Regulatory guidance on agentic AI tool connections is expected to reference recognized standards bodies. This includes the Singapore Model AI Governance Framework for Agentic AI. Early engagement with the Agentic AI Foundation's output will be relevant to future conformity assessments.