Linux Foundation Takes On AI Agent IDs

The Linux Foundation on Tuesday, Nov. 10, took Agent Name Service (ANS), an open standard that uses the internet's Domain Name System to give AI agents verifiable identities, under its wing as an official project. The announcement came at KubeCon + CloudNativeCon North America in Salt Lake City.

ANS tries to answer a basic question that grows harder as software agents call each other across company lines: who is this agent, and who stands behind it? Instead of a new central directory, it ties each agent to a domain its operator already controls.

Key takeaways

• ANS is now a Linux Foundation project, governed by a technical steering committee, with a specification made up of seven documents.
• Eleven organizations, including Cisco, Cloudflare, DigiCert, GoDaddy, Infoblox and Salesforce, have contributed to or adopted the standard since June.
• The project's own site still labels the published specifications as drafts, and it stresses that ANS proves identity, not what an agent is allowed to do.

How Agent Name Service works

Each agent gets a versioned, DNS-style name such as ans://v1.0.0.payments.example.com, according to the project's GitHub organization. An identity certificate is issued only after the operator proves control of the domain through DNS and ACME validation, the same kind of check used for website certificates.

Every registration event is written to an append-only, Merkle-tree transparency log modeled on Certificate Transparency (RFC 6962 and RFC 9162). An offline tool, ans-verify, can check an agent's record and proofs without contacting a live service. The project sums up the design as "DNS + Certificate Transparency, built for agents."

According to the Linux Foundation's announcement, the seven specification documents cover domain-anchored naming, registration and versioning, DNS publication, public log auditing, integrity monitoring and agent-to-agent authentication.

Who supports the AI agent identity standard

The foundation lists 11 organizations that have contributed to or adopted ANS since its intent to launch was announced on June 23: Cisco, Cloudflare, DigiCert, DistributedApps.ai, GoDaddy, HOL, Infoblox, Radix, Salesforce, OWASP and Trustmodel.ai.

Two discovery tools already use it. The AGNTCY Agent Directory Service, another Linux Foundation project, can verify ANS identities. ANS Finder implements the Agentic Resource Discovery (ARD) specification and passes its registry conformance tests.

"Autonomous AI agents are transacting and operating across global systems, but they can't succeed without verifiable identity," said Jim Zemlin, the Linux Foundation's CEO. Brian Trzupek, a senior vice president at DigiCert, put it more simply: "Think of it as a passport for AI agents."

Context: why agent identity matters now

Companies are giving agents access to internal systems, payment flows and other agents. Cloudflare CTO Dane Knecht said in the release that organizations "need to know who an agent represents before trusting it to act." Vendors have been building their own answers, such as IBM's Agent Identity product, while incidents involving agents that turned to SQL injection when ordinary requests failed keep the issue in view.

The reference implementation is written in Go under the MIT license and builds on an IETF draft. SDKs are available in Go, Rust and Java.

What's next

The press release says the specification is "now formalized," but the project's official site says the published specifications "are drafts under active development." Ken Huang, CEO of DistributedApps.ai, who led ANS's initial development, said he looks forward to joining the technical steering committee. The site also notes that ANS establishes identity, while applications still decide what a verified agent may do.

FAQ

Does ANS replace DNS?
No. It builds on DNS and existing certificate checks, publishing agent identity information through domains that operators already own.

Can developers try it today?
Yes. The project publishes a Go reference implementation with a registry, transparency log and offline verifier, plus Go, Rust and Java SDKs.

Does a verified ANS identity mean an agent is safe?
No. ANS confirms who operates an agent. Permissions and trust decisions remain with the applications that interact with it.

References

• Linux Foundation Welcomes Agent Name Service to Bring Verifiable Identity and Discovery to AI Agents — The Linux Foundation (Nov. 10, 2026)
• Verifiable identity for AI agents · Agent Name Service — Agent Name Service
• Agent Name Service · GitHub — Agent Name Service