Skip to content
IAM Insights

Explainers · Glossary

Dynamic Client Registration

DCR · RFC 7591 · RFC7591

By Boston Identity · Explained · 7 min read

The short answer

Dynamic Client Registration (RFC 7591) lets an app or agent register itself with an OAuth authorization server at runtime. It sends a description of itself and gets back a client ID, with no admin setting it up by hand.

What it is
A protocol
Defined by
IETF, RFC 7591 (Standards Track, 2015)
Who it's for
Teams connecting agents and MCP clients to authorization servers
In our briefings
69By month of the source: 2025-12: 0, 2026-01: 1, 2026-02: 4, 2026-03: 3, 2026-04: 2, 2026-05: 2, 2026-06: 1, 2026-07: 9, 2026-08: 1, 2026-09: 11, 2026-10: 6, 2026-11: 00by month, last 12
01

What is Dynamic Client Registration?

Before an app can use OAuth, the open standard behind most delegated sign-ins and API permissions, it needs a client identifier at the authorization server, the service that hands out tokens. Traditionally someone creates that identifier by hand.§1 RFC 7591, published by the IETF (Internet Engineering Task Force) in July 2015, lets it happen over the network instead. The client or its developer sends the server a set of facts about the client, called metadata, and gets back a client ID plus the metadata the server actually registered.Abstract

It exists because a developer often cannot know, while building the software, which organizations will run the API it talks to, so the software cannot be set up in advance at each of them.§1.2 The metadata covers the redirect URIs (the addresses the server may send a user back to after sign-in), the client's name and logo, and how it will prove itself at the token endpoint.§2 The RFC generalizes the registration methods that OpenID Connect and UMA (User-Managed Access) already used, and stays compatible with both.§1

DCR is not the same as Client ID Metadata Documents (CIMD), where the client's ID is a web address pointing to its own description and no prior registration happens.1 With DCR, the server issues the client ID itself, in reply to a registration request.§3.2.1

02

How does Dynamic Client Registration work?

Two parties take part: the client (an app, an agent, or a developer acting for it) and the authorization server's client registration endpoint. An optional token and an optional signed statement can come before the request. How the client learns the endpoint's address is left outside the RFC.§1.2

Figure steps
  1. The server may issue an initial access token
  2. A publisher may sign a software statement
  3. The client posts its metadata
  4. The server weighs the request
  5. The server issues the client ID
  6. The client can check the reply, then starts OAuth
Fig. 1

How a client registers itself with an authorization server and gets a client ID under RFC 7591.

Drawn from RFC 7591

Swipe for the full figure →

  1. The server may issue an initial access token

    A server that wants to limit who can register gives the developer or client an initial access token, an OAuth token used to authorize calls to the registration endpoint. Its type and format are up to each service, and the RFC leaves out how it is issued;§1.2 a common approach is a manual portal where the developer signs up.S1 Without the token, registration is open, and the RFC says the endpoint should accept such requests to support interoperability.§3

  2. A publisher may sign a software statement

    The client can also carry a software statement. This is a JWT (JSON Web Token) signed by the developer or a third party, asserting facts about the software such as its name and a software ID.§2.3 It must be signed, or protected with a message authentication code, and it must name its issuer. Each server decides for itself which issuers to trust.§2.3

  3. The client posts its metadata

    The client or developer sends an HTTP POST to the registration endpoint. The body is a JSON object listing its redirect URIs, name, logo and token-endpoint authentication method, plus the software statement if it has one.S1 In the RFC's example, an initial access token travels separately, in the request's Authorization header as a bearer token.§3.1 The endpoint must accept JSON and must sit behind TLS (Transport Layer Security), because credentials travel in clear text inside the request and response.§3, §5

  4. The server weighs the request

    The server must treat every metadata value as self-asserted, except values inside a software statement it trusts, which override the plain JSON.§5 It must consider the whole request together: statement, token and metadata.§5 It may replace requested values with its own, or refuse with an HTTP 400 error such as invalid_redirect_uri or invalid_client_metadata.S1

  5. The server issues the client ID

    On success the server answers 201 Created. The reply carries a client_id, a client secret if the client is confidential, and every metadata value the server registered, including values it filled in itself.§3.2.1 The client never chooses its own identifier. The RFC forbids that so that one client cannot be tracked across colluding servers.§6

  6. The client can check the reply, then starts OAuth

    The client or its developer can compare what came back with what it asked for, for example whether the server kept an authentication method the software supports.§3.2.1 It then uses the new client ID in ordinary OAuth flows with that server.Abstract

03

Why does Dynamic Client Registration matter now?

The Model Context Protocol (MCP), the standard that connects AI assistants to tools, has demoted DCR. Its 2026-07-28 revision deprecated DCR as the way an MCP client obtains a client ID, made CIMD the mechanism servers and clients should support, and kept DCR only for servers that lack CIMD.2 MCP's authorization page now says DCR is retained for backwards compatibility.3 The same revision requires every DCR request to carry an application_type. It also adopted a lifecycle rule that keeps a deprecated feature for at least twelve months before removal.4

Vendors still ship it. Okta and Ping Identity released CIMD in September and changed their DCR support in the same releases, so neither has dropped it. CyberArk's September MCP feature relies on DCR.1 Cisco Duo's MCP integration, in beta, lets agents register by DCR and does not yet support CIMD.5 A view recorded in November 2025 held that DCR in MCP would persist even after client ID metadata arrived, because so much was already built on it.6

A paper on agent identity says DCR "creates a large number of anonymous clients".7 One study tested 119 remote MCP servers that register clients dynamically, and 114 of them accepted any redirect address from an anonymous registrant.8 That rate describes only the servers open to outside registration.

04

Where does it fit?

Compared withHow Dynamic Client Registration differs
Client ID Metadata DocumentsWith CIMD the client ID is a web address the client controls, and nothing is registered in advance.1 MCP now prefers it and keeps DCR for servers that cannot do CIMD.3
OAuth 2.0 Protected Resource MetadataProtected resource metadata is how an MCP server names the authorization server that guards it, and MCP requires every server to publish it.3 Getting a client ID at that server is a separate step, done by CIMD, pre-registration or DCR.3
OpenID ConnectOpenID Connect had its own dynamic registration spec first. RFC 7591 generalizes it to OAuth uses beyond sign-in and stays compatible with it.§1
05

What goes wrong?

  • Everything a client says about itself is self-asserted. A rogue client can send a real app's name and logo, or reuse its software ID, to pass as that app.§5 The RFC asks servers to check that the logo and policy links share a host with the redirect URIs, and suggests warning users about newly registered clients.§5 When a client is not meant to have several instances, a server should treat as suspect, and reject, a new registration that reuses an existing client's software ID and version.§5
  • The RFC makes redirect URIs mandatory for redirect-based flows, to stop an impostor intercepting a registered client's code.§5 In our reading, that rule cannot help when the attacker is the one registering. In the MCP study, an anonymous registrant could name their own callback, get a victim to click a genuine sign-in link and collect the code; that flaw sat behind seven of the nine CVEs (published vulnerability records) the authors obtained.8
  • A software statement is signed, yet the RFC still counts it as self-asserted, because the client presents it. In most cases it is not enough on its own to fully identify the software.§5
  • A client ID is public and can be used to impersonate the client at the authorization endpoint. A server that gives one client ID to many instances has to be careful about when it does so, and it should not give them the same client secret.§5
  • Open registration can be flooded. The RFC lets servers rate-limit registration against denial of service.§3 It also notes that browser apps without storage may re-register on every page load, leaving dead client IDs the server may expire.§5
  • CIMD leaves part of this gap open. In our reading of the CIMD draft, an attacker can host a document naming their own callback, and at least 62 of the 119 consent pages in the MCP study did not show where the code would go.9
06

What should a security leader ask?

  1. 01

    Which of our authorization servers accept client registrations with no initial access token, and who reviews the clients created that way?

  2. 02

    Do our servers accept any redirect URI from a dynamically registered client, or only ones on a host tied to that client?

  3. 03

    Which of our MCP connections still depend on DCR, and what is the plan for each before MCP removes it?

  4. 04

    What access does a self-registered agent get by default, and who must approve anything more?

  5. 05

    For each identity vendor we use, is CIMD supported today, or DCR only?

07

Sources

›Evidence: 9 claims from 8 briefings
  1. 1BI-FIND-2026-0704 · Three vendors now route MCP tool calls through an identity check, and only Cloudflare's is generally availableOkta and Ping both shipped client ID metadata documents (CIMD), and CyberArk went the other way. With CIMD, an agent or app names itself with a URL that points to its own metadata, and needs no prior registration. Okta made CIMD a beta for AI agents and GA in preview for OIDC apps on September 16. Ping's Advanced Identity Cloud shipped it on September 29. CyberArk's September MCP feature relies on dynamic client registration (DCR) instead. Okta and Ping also changed their DCR support in the same releases, so neither has dropped it.
  2. 2BI-FIND-2026-0111 · An MCP gateway can be demoed in a week; Sierra's rollout waited on the auth layer underneath itThe Model Context Protocol revision of 2026-07-28 deprecated OAuth Dynamic Client Registration as the way an MCP client obtains a client ID and made Client ID Metadata Documents the mechanism authorization servers and clients should support, with DCR kept only for servers that lack CIMD. We checked the changelog, which confirms WorkOS's reading. The same revision still defines no gateway or proxy role and keeps the rule that an MCP server must not accept or pass on tokens issued for anything else, which is the rule a token-passthrough gateway breaks. Anyone building the client-identity layer of a gateway this quarter is building against a moving target with one fixed point.
  3. 3BI-FIND-2026-0190 · MCP's authorization spec makes the agent a plain OAuth client and leaves delegation to an extensionWe had this from WorkOS's reading of the changelog; the page confirms it first-hand. Every MCP server MUST publish protected-resource metadata naming its authorization server, and that server MUST expose its own metadata or OpenID discovery. A client obtains its identity by a Client ID Metadata Document, by pre-registration, or by dynamic client registration, and the last is now the fallback: 'Dynamic Client Registration is deprecated and retained for backwards compatibility with authorization servers that do not support Client ID Metadata Documents.' Servers should name the scopes an operation needs in the 401 challenge. ZeroID implements both registration paths; the specification has now chosen between them, and any gateway that onboards agents by dynamic registration is building on the deprecated path.
  4. 4BI-FIND-2026-0395 · MCP's July changelog: a stateless protocol and clients named by web address, but no mention of gatewaysThe identity changes in the 2026-07-28 revision are filed under Minor and Deprecated, and the minor ones are the new ones. Clients 'MUST key persisted credentials by the issuer identifier, MUST NOT reuse them with a different authorization server, and MUST re-register when the authorization server changes' (SEP-2352); clients must validate a present RFC 9207 iss parameter before redeeming a code (SEP-2468); a Dynamic Client Registration request must carry an application_type (SEP-837). Above them, DCR is deprecated 'as a client registration mechanism in favor of Client ID Metadata Documents' (PR #2858) and kept for authorization servers without CIMD, which we had already confirmed first-hand on the authorization page itself; what the changelog adds is the clock, since the same revision adopts a lifecycle policy with a twelve-month minimum before a deprecated feature can be removed. None of it is marked breaking. The CIMD draft the revision points at never mentions MCP; WorkOS and the OpenID whitepaper both read DCR's anonymous client as the risk this closes.
  5. 5BI-FIND-2026-1194 · Cisco Duo's MCP beta lets agents register themselves; Duo calls client ID metadata documents the futureCisco Duo added two OAuth standards to a new beta integration for the Model Context Protocol (MCP) and to its OAuth 2.1/OIDC integration, also in beta, to match the MCP specification released on November 25. Dynamic client registration (RFC 7591) lets agents register themselves as OAuth clients, and Duo says admins keep full control over the access level of clients registered that way. Resource indicators (RFC 8707) let an agent name the resource it wants so Duo issues a token valid only there, which Medfisch says ensures a token for one service "cannot be replayed or accepted by another." Client ID metadata documents are not in the beta: Medfisch calls them "the future" and says Duo is building support, with no date.
  6. 6BI-FIND-2025-0085 · WorkOS's Tobin South, in November 2025: agents will get a new kind of account, not a new identity standardAsked in November 2025 how an assistant could hand its access to downstream agents, South said on-behalf-of flows and delegated authority existed as proposals "but they're not widely supported". Separately, he told of a developer whose MCP-connected agent, running with his super-admin privileges, took down Netlify's core website, and warned that "just because you as a person are extremely trustworthy, that does not necessarily mean Claude code using your permissions is extremely trustworthy". He also expected dynamic client registration in MCP to persist by path dependency even after client ID metadata arrived.
  7. 7BI-FIND-2025-0057 · OAuth handles an agent in one company, working while you watch. The OpenID paper says that is where it stops.The paper finds every existing way of identifying an agent as a client incomplete. Dynamic client registration 'creates a large number of anonymous clients'. Client ID Metadata Documents fix that and are 'still insufficient as a workload identity'. SPIFFE gives a strong workload identity that depends on controlling the infrastructure, so it does not cross into another organization. Vendor systems such as Microsoft Entra Agent ID and Okta's are named as proprietary, with limited interoperability. In the year since, the Model Context Protocol has made metadata documents its preferred method and an IETF draft has joined them to SPIFFE.
  8. 8BI-FIND-2026-0944 · 114 of 119 MCP servers that let outsiders register an OAuth client let them choose where sign-in codes are sentOf the 119 remote MCP servers whose OAuth setups Zhou, Zhang and colleagues could test end to end, all of which register clients dynamically, 114 accepted any redirect address from an anonymous registrant. That lets an attacker register a client with a callback of their own, get the victim to click a genuine sign-in link, and collect the authorization code; seven of the nine CVEs the authors obtained trace to this flaw. Every server tested had at least one flaw, 325 in all. The 119 are what was left of 1,118 servers advertising a registration endpoint after the authors dropped duplicates, unreachable hosts and 316 servers behind enterprise access controls, so the rate describes servers open to outside registration, not MCP as a whole.
  9. 9BI-FIND-2026-0947 · 114 of 119 MCP servers that let outsiders register an OAuth client let them choose where sign-in codes are sentOne of the authors' fixes for open registration is a move to Client ID Metadata Documents (CIMD), which they say pin client identity to a "cryptographically verifiable" document, and the MCP specification's 2026-07-28 revision did deprecate dynamic registration in favour of CIMD, keeping it only for authorization servers that lack CIMD support. We think the fix is weaker than the paper makes it sound. In the CIMD draft, which we read separately, a client's identity is control of a web address, its redirect addresses come from the document it publishes, and tying them to the client's own hostname is optional for the server. An attacker can host a document naming their own callback. What is left to stop them is the consent page telling the user who is asking, and in this study at least 62 of the 119 consent pages did not show where the code would go.

›In our briefings

The claim each of the newest briefings on Dynamic Client Registration would have you leave with.

  1. Linx's Goldenberg calls one agent acting for one user "in a way" a solved challenge, because the user's token usually already carries what downstream applications need to control data access. Once agents delegate to other agents across several hops, he says, "you lose track of who is the initial actor", and he sees toxic combinations and violations of segregation of duties. Aembit's Goldschlag answers with token exchange at every hop, from user to agent to gateway to MCP server to resource, each hop holding only the credential for the next step. Godha points to IETF drafts on chained delegation tokens and to transaction tokens, which he says are coming.

    When AI agents hand work to other agents, the person who started the job can get lostBI-FIND-2026-0070

  2. None of the authorization changes in MCP's 2026-07-28 revision is listed under Major changes; they sit under Minor changes and Deprecated. SEP-2352 says clients 'MUST key persisted credentials by the issuer identifier, MUST NOT reuse them with a different authorization server, and MUST re-register when the authorization server changes'. Authorization servers should include the RFC 9207 iss (issuer) parameter in authorization responses, and clients must check a present iss against the recorded issuer before redeeming the authorization code (SEP-2468). SEP-837 requires clients to specify an appropriate application_type during Dynamic Client Registration, to avoid OpenID Connect redirect URI conflicts. Dynamic Client Registration itself is deprecated 'as a client registration mechanism in favor of Client ID Metadata Documents' (PR #2858) and remains available for authorization servers that do not support them. The same revision adopts a lifecycle policy with 'a minimum twelve-month deprecation window'; the changelog gives no removal date for Dynamic Client Registration and does not say when its window starts.

    MCP's July spec ties client credentials to the authorization server that issued themBI-FIND-2026-0395

  3. AAuth's premise is that an OAuth client or an API key has no identity of its own and that SPIFFE and WIMSE stop at one trust domain, so it gives each agent one: an identifier of the form aauth:local@domain issued by an Agent Provider and bound to the agent's own signing key by a short-lived agent token (the draft recommends Ed25519, which every party must support), so that 'every agent has its own cryptographic identity' and 'the first API call to a resource is the registration'. No credential in the protocol is a bearer credential; every request is signed, and a resource can challenge for more at any time. In its lightest mode the draft pitches this as a drop-in for API keys: 'this replaces API keys with cryptographic identity.' What an agent token attests about the software behind the key is left to companion specifications.

    AAuth gives every agent its own key, then keeps the delegation chain out of the token on purposeBI-FIND-2026-0365

›Cited in 69 briefings

Every briefing that discusses this term, newest first.

  1. When AI agents hand work to other agents, the person who started the job can get lost

    Podcast · · The Non-Human & AI Identity Podcast · NHI & AI Identity Summit : Governing AI · Sep 7, 2026

  2. MCP's July spec ties client credentials to the authorization server that issued them

    Article · · Model Context Protocol project · Jul 28, 2026

  3. AAuth gives every agent its own key, then keeps the delegation chain out of the token on purpose

    Draft · · IETF Datatracker · Dick Hardt, Hellō · Sep 25, 2026

  4. The OAuth working group split over a token-binding draft that never says what DPoP lacks

    Draft · · IETF Datatracker · Justin Richer et al. · Jul 28, 2026

  5. MCP's November 2025 security guide makes proxy servers ask consent for each client before forwarding a login

    Article · · Model Context Protocol · Model Context Protocol project · Nov 25, 2025

  6. Okta and Google tie an agent's calls to the user, while GitHub's shared Copilot automations now act as Copilot

    Release tracker · · 11 vendors' release notes · Oct 2026

  7. MCP's client credentials extension puts the secret in the form body, while its server rule names HTTP Basic

    Article · · Model Context Protocol project · Oct 23, 2026

  8. An OAuth client gets to name who vouches for it, and removing one waits on server caches

    Draft · · IETF Datatracker · Karl McGuinness · Sep 28, 2026

  9. Register a platform's identity provider once, not each agent: trust rests on its issuer, with no revocation

    Draft · · IETF Datatracker · Paul Carleton et al. · Sep 22, 2026

  10. WIMSE's agent draft adds how an agent's access ends and where policy engines fit, without a new requirement

    Draft · · IETF WIMSE Working Group · Pieter Kasselman et al. · Sep 15, 2026

  11. OAuth's URL client ID draft tightens server rules in -03; the client is still whoever controls the URL

    Draft · · IETF Datatracker · Aaron Parecki et al. · Jul 6, 2026

  12. OAuth's SPIFFE draft opens metadata URLs to all three credentials and says what a self-served key cannot prove

    Draft · · IETF Datatracker · Arndt Schwenkschuster et al. · Jun 15, 2026

  13. WorkOS's agent checklist limits a delegated agent to what it and its user may both do, checked on every call

    Article · · WorkOS Blog · Maria Paktiti · Sep 30, 2026

  14. OAuth's draft for token requests that wait on a person now gives the client a link for the user to approve

    Draft · · IETF Datatracker · Guilherme de Oliveira Niero et al. · Sep 23, 2026

  15. Cross-App Access draft warns that one stolen sign-in token could reach every app the identity provider permits

    Draft · · IETF Datatracker · Aaron Parecki et al. · May 21, 2026

  16. MCP's November 2025 authorization spec puts metadata documents ahead of dynamic client registration

    Article · · Model Context Protocol · Nov 25, 2025

  17. OAuth tokens prove an agent's rights, not a person's consent; revision 05 adds a day-long grant with spend caps

    Draft · · IETF Datatracker · Meiling Chen et al. · Oct 27, 2026

  18. MCP client setup moves from registration to metadata checks in a Spring walkthrough

    Video · · OWASP Foundation · Oct 24, 2026

  19. Six companies' agent-auth draft, since adopted by the WIMSE group, keeps credentials out of the model's reach

    Draft · · IETF Internet-Draft (individual submission, since adopted by the WIMSE working group) · Pieter Kasselman et al. · Jul 6, 2026

  20. OpenID Federation 1.0 lets each level of a trust chain only tighten the rules, and defines no revocation process

    Article · · OpenID Foundation · Roland Hedberg et al. · Feb 17, 2026

  21. An attested OAuth client is one somebody vouched for; revision 12 says to trust its claims only as far as the voucher

    Draft · · IETF Datatracker · Tobias Looker et al. · Oct 13, 2026

  22. Eve Maler says the push to give every AI agent a responsible human misses agents a company does not own

    Podcast · · Identity at the Center · #453 - Treating Identity Like a Product with Eve Maler · Oct 12, 2026

  23. An MCP server picks the login URL its client opens, and that was enough to run code on users' machines

    Video · · Black Hat · Aug 20, 2026

  24. TPX replaces an app's embedded AI provider key with a capped, metered grant the user approves at their provider

    Article · · tokenpony.dev · Alex Polvi · Jul 20, 2026

  25. Three vendors now route MCP tool calls through an identity check, and only Cloudflare's is generally available

    Release tracker · · 11 vendors' release notes · Sep 2026

  26. DIF's agent spec names every AI agent by a DID; the IETF's Agent Name Service switched to domain names

    Article · · Decentralized Identity Foundation (KYA-OS) · Dylan Hobbs et al. · Jul 29, 2026

  27. Registering an MCP server on Bifrost ran its command, and the admin API that allowed it shipped with no password

    Article · · WorkOS Blog · Maria Paktiti · Sep 29, 2026

  28. WorkOS's Tobin South, in November 2025: agents will get a new kind of account, not a new identity standard

    Podcast · · CodingCat.dev Podcast · 5.10 - How to Secure AI - Identity Access Management for Agents · Nov 21, 2025

  29. OAuth handles an agent in one company, working while you watch. The OpenID paper says that is where it stops.

    Article · · OpenID Foundation · Tobin South et al. · Oct 2025

  30. Two ways to give an AI agent an OAuth identity clashed in April; a draft now lets them work together

    Podcast · · Hybrid Identity Protection Podcast · Securing Non-Human Identities in the Age of Agentic AI with Sarah Cecchetti, Director of Product Management at Semperis · Apr 28, 2026

  31. MCP's authorization spec makes the agent a plain OAuth client and leaves delegation to an extension

    Article · · Model Context Protocol project · Jul 28, 2026

  32. An MCP gateway can be demoed in a week; Sierra's rollout waited on the auth layer underneath it

    Article · · WorkOS Blog · Maria Paktiti · Sep 11, 2026

  33. Highflame's open-source ZeroID narrows an agent's token at every hop, a step we did not find in Uber's write-up

    Article · · Highflame (GitHub) · Sep 25, 2026

  34. The browser OAuth draft favors a backend that keeps tokens out of JavaScript

    Draft · · IETF Datatracker · Aaron Parecki et al. · Jul 6, 2026

  35. A July OAuth draft keeps agent identity stable as keys rotate

    Draft · · IETF Datatracker · Karl McGuinness · Jul 5, 2026

  36. 114 of 119 MCP servers that let outsiders register an OAuth client let them choose where sign-in codes are sent

    Paper · · arXiv · Huijun Zhou et al. · May 21, 2026

  37. Microsoft and Google made AI agent identity generally available in April; Okta's notes put its GA in May

    Release tracker · · 8 vendors' release notes · Apr 2026

  38. OpenID agent-claims draft adds a provider-set user, application or agent label; scores stay each provider's own

    Draft · · IETF Datatracker · Raza Sharif, CyberSecAI · Mar 26, 2026

  39. A KubeCon demo moves MCP servers' OAuth checks into a gateway, safe only if nothing can reach a server directly

    Video · · CNCF [Cloud Native Computing Foundation] · Uploaded March 24, 2026

  40. OAuth 2.1's October draft bans plain PKCE and adds browser-app rules; 2.1-only clients can fail on older servers

    Draft · · IETF Datatracker · Dick Hardt et al. · Mar 2, 2026

  41. OAuth's in-app login draft lets a native app collect the password itself, and leaves each server to vet the app

    Draft · · IETF Datatracker · Aaron Parecki et al. · Feb 28, 2026

  42. A Cisco-led agent draft puts the verifiable credential at a public URL, with selective disclosure a SHOULD it does not yet show

    Draft · · IETF Datatracker · Jean Diaconu et al. · Feb 6, 2026

  43. In February Okta let AI agents reach its management APIs over MCP; Cloudflare added controls on MCP connections

    Release tracker · · 7 vendors' release notes · Feb 2026

  44. Cisco Duo's MCP beta lets agents register themselves; Duo calls client ID metadata documents the future

    Article · · Cisco Duo · Colin Medfisch · Jan 7, 2026

  45. On its authors' tasks, an LLM check of agent scopes works at one tool and refuses most valid three-tool requests

    Paper · · arXiv · Majed El Helou et al. · Oct 30, 2025

  46. An agent OAuth proposal has the helper forward grants before receiving access tokens

    Draft · · IETF Datatracker · Kehan Yao, China Mobile · Oct 20, 2025

  47. SPIFFE can authenticate an OAuth client; trust and registration still come first

    Draft · · IETF Datatracker · Arndt Schwenkschuster et al. · Oct 3, 2025

  48. One author's OpenID Connect for Agents proposal checks each delegation hop but makes the chain optional

    Paper · · arXiv · Subramanya Nagabhushanaradhya · Sep 30, 2025

  49. An OAuth client that never registers: the July 2025 draft makes its URL the identity and leaves most server checks as SHOULDs

    Draft · · IETF Datatracker · Aaron Parecki et al. · Jul 23, 2025

  50. The July 2025 Cross App Access draft lets the IdP decide app-to-app access and treats agents as plain clients

    Draft · · IETF Datatracker · Aaron Parecki et al. · Jul 2, 2025

  51. MCP's June 2025 authorization spec: a server accepts only tokens issued for it and passes none upstream

    Article · · Model Context Protocol project · Jun 18, 2025

  52. An MCP gateway can do OAuth for servers that have none; this one passes them a user header

    Paper · · arXiv · Ivo Brett · Apr 28, 2025

  53. RFC 9728 tells clients where an API's tokens come from and leaves most of the trust decision to them

    Article · · RFC Editor · Michael B. Jones et al. · Apr 2025

  54. RFC 9700 bans OAuth's password grant and loose redirect matching; binding access tokens stays a SHOULD

    Article · · IETF (RFC Editor) · Torsten Lodderstedt et al. · Jan 2025

  55. DPoP token binding depends on checks at the resource server

    Article · · RFC Editor · Daniel Fett et al. · Sep 2023

  56. Rich Authorization Requests need API-specific permission rules

    Article · · RFC Editor · Torsten Lodderstedt et al. · May 1, 2023

  57. RFC 9207 has OAuth servers name themselves in each response; clients must refuse a missing name only where they expect one

    Article · · RFC Editor · Karsten Meyer zu Selhausen et al. · Mar 1, 2022

  58. Signed OAuth requests can be bypassed until the server refuses unsigned ones, and RFC 9101 leaves that off by default

    Article · · RFC Editor · Nat Sakimura et al. · Aug 1, 2021

  59. RFC 8705 lets OAuth bind tokens to a client certificate, and leaves the proxy in front of the API to the deployer

    Article · · RFC Editor · Brian Campbell et al. · Feb 28, 2020

  60. RFC 8414 tells an OAuth client to discard server metadata whose issuer differs from the URL it started from

    Article · · RFC Editor · Michael B. Jones et al. · Jun 1, 2018

  61. RFC 8252 puts native-app OAuth authorization outside the host app’s browser control

    Article · · IETF / RFC Editor · William Denniss et al. · Oct 1, 2017

  62. A 2017 security proof for OpenID Connect assumes fixes the spec did not yet require

    Paper · · arXiv · Daniel Fett et al. · Apr 27, 2017

  63. Trying to prove OAuth 2.0 secure, a 2016 Trier paper found four attacks, the IdP mix-up among them

    Paper · · arXiv · Daniel Fett et al. · Jan 6, 2016

  64. RFC 7592 defines replacement updates for managed OAuth registrations

    Article · · RFC Editor · Justin Richer et al. · Jul 9, 2015

  65. RFC 7591 asks OAuth servers to accept clients with no token, and calls their signed statements self-asserted

    Article · · RFC Editor · Justin Richer et al. · Jul 9, 2015

  66. OAuth's assertion framework is built for bearer assertions, and one can buy fresh tokens until it expires

    Article · · RFC Editor · B. Campbell et al. · May 1, 2015

  67. RFC 7523 lets a signed JWT authenticate an OAuth client or serve as a grant, and leaves replay checks optional

    Article · · RFC Editor · Michael B. Jones et al. · May 1, 2015

  68. RFC 6819 warned in 2013 that servers receiving a token may reuse it, and offered audience binding and client-bound tokens

    Article · · RFC Editor · T. Lodderstedt et al. · Jan 2013

  69. RFC 6749 admits a server can't tell who presents a token; the fixes arrived in 2020 and 2023

    Article · · RFC Editor · D. Hardt et al. · Oct 2012

In our columns

  1. 114 of 119 MCP servers tested accepted any redirect. The registration debate misses it

    Column · Nov 13, 2026

  2. By default, an MCP agent's name tells you only who owns a web address

    Column · Oct 3, 2026

Related terms

What appears alongside Dynamic Client Registration in the same briefings.

Client ID Metadata DocumentsConsent phishingMix-up attackOpen redirectorOAuth 2.0 Authorization Server MetadataPKCEOAuth 2.0 Protected Resource MetadataJWT Bearer Grant

Was this explainer useful?

Share this explainer

Pass it on. Every claim has a permanent number, so what you share today still points at the same words next year.