RISC
Risk Incident Sharing and Coordination · Risk and Incident Sharing and Coordination Profile
By Boston Identity · Explained · 5 min read
The short answer
RISC is an OpenID Foundation set of account-risk event types, such as an account disabled after hijacking or a credential found compromised, that one service sends another as security event tokens over the Shared Signals Framework.
- What it is
- A standard
- Defined by
- OpenID Foundation, RISC 1.0 (Final)
- Who it's for
- Teams linking identity providers and apps
- In our briefings
- 70by month, last 12
What is RISC?
RISC, short for Risk Incident Sharing and Coordination, defines the event types one system needs to tell another that something has happened to an account.§1 The events are based on the OpenID Shared Signals Framework (SSF), and each one travels as a Security Event Token (SET), a message format for security events defined in its own specification.S1
The list covers an account's life as one provider sees it. A credential was found compromised, the account was disabled, enabled or permanently deleted, the user started a recovery flow, or an email address or phone number changed or passed to a new owner.§2 We read the point as letting a service hear about a risk change on a shared account that it did not see itself.
The Shared Signals working group published RISC 1.0 on 29 August 2025,S1 and the OpenID Foundation announced its approval as a Final Specification on 16 September 2025.1
Not the same as CAEP, the Continuous Access Evaluation Profile, which carries signals about sessions. RISC's own event for revoking all of an account's sessions is deprecated, and new implementations must use CAEP's session-revoked event instead.§2.11
How does RISC work?
The specification names two parties, a Transmitter that sends events and a Receiver that gets them.§2.8 In its examples the transmitter is an identity provider at idp.example.com.§2.1
Figure steps
- The transmitter picks an event type
- The transmitter writes a security event token
- The transmitter adds detail where the event allows it
- Identifier events name an email or phone
- The user asks to opt out
- The transmitter reports where the opt-out ended
The transmitter names one RISC event type in a token addressed to the receiver, and an account that asks to opt out keeps exchanging events for a while before the opt-out takes effect.
Drawn from RISC 1.0, sections 2 and 2.8
Swipe for the full figure →
The transmitter picks an event type
Each RISC event type is a web address under one base address, https://schemas.openid.net/secevent/risc/event-type/, with a name such as account-disabled or credential-compromise on the end.§2
The transmitter writes a security event token
The specification's examples carry an issuer (iss), a unique token id (jti), an issue time (iat), an audience (aud), the subject's identifier (sub_id) and an events claim keyed by the event type's address.§2.1
The transmitter adds detail where the event allows it
Account Disabled may carry an optional reason, hijacking or bulk-account.§2.3 Credential Compromise must name the credential type, using the values CAEP defines, and may add when the transmitter discovered it and separate reasons for administrators and for users.§2.7
Identifier events name an email or phone
For Identifier Changed and Identifier Recycled, the subject must be an email address or phone number.§2.5, §2.6 Identifier Changed must give the old value and may give the new one, and it should come only from the provider that is authoritative over the identifier.§2.5
The transmitter reports where the opt-out ended
Opt Out Effective reports that the account is now in the opt-out state and no longer takes part, and Opt Out Cancelled reports that it is back in opt-in.S1 Transmitters should send these events when the opt-out state changes, and the state diagram also shows an opted-out account opting in again.§2.8
Why does RISC matter now?
The OpenID Foundation approved Shared Signals 1.0, CAEP 1.0 and RISC 1.0 as Final Specifications in September 2025, aimed at the time after login.1 Its release says sessions often last days or weeks while a user's access can change.1 It also says Apple, IBM and Okta have adopted these protocols, without naming a product or reporting any deployment.1
One agent-authorization draft lets any participant subscribe to Shared Signals events under the CAEP or RISC profiles.2 It says a recipient of a revocation or risk signal should remediate, and must enforce revoked or downgraded authorization without undue delay.2 Shared Signals 1.0 itself never mentions agents, workloads or AI, and account-disabled is a RISC event rather than one the framework defines.3
We forecast that by March 2028 the Shared Signals working group publishes a draft or profile with agent-specific events, such as an agent's credential being revoked.4 That forecast settles wrong if CAEP and RISC still describe only human sessions and devices by then.4
Where does it fit?
| Compared with | How RISC differs |
|---|---|
| SSF | SSF is the framework the events travel over. RISC only defines account-risk event types, and its security section says implementations should comply with SSF.§4 |
| CAEP | CAEP carries session signals, and RISC now defers to CAEP's session-revoked event.§2.11 RISC's Credential Compromise event also takes its list of credential types from CAEP.§2.7 |
What goes wrong?
- The specification warns that exchanging these events without complying with SSF may result in security issues.§4
- An event says what happened, not what the receiver must do. Shared Signals 1.0 lets each deployment set how events are processed, from informational input to mandatory enforcement.5
- Users can leave the exchange, since users should be allowed to opt out.§2.8 Parecki's Global Token Revocation draft points to that opt-out when it contrasts RISC with its own approach, which is framed as a command with defined outcomes.6
- Google's existing RISC implementation uses the field name subject_type instead of format. New services must not use that name, and receivers of Google's events need code to work around it, which should be written so it does not break if Google updates its implementation.§3
What should a security leader ask?
- 01
Which of our identity providers and apps can send or receive RISC events today, and over which Shared Signals stream?
- 02
When an account-disabled or credential-compromise event arrives, what does each receiving system do, and how quickly?
- 03
Which accounts have opted out of RISC events, and how long does an opt-out request wait before it takes effect?
- 04
If we take events from Google, does our receiver handle both subject_type and format?
- 05
How would our receivers tell an agent's account from a person's if RISC events arrive for both?
Sources
- S1
OpenID RISC Profile Specification 1.0OpenID Foundation · Final, 29 August 2025
›Evidence: 6 claims from 6 briefings
- 1BI-FIND-2025-0282 · OpenID Foundation finalized Shared Signals, CAEP and RISC in September 2025, naming adopters without evidenceOn 16 September 2025 the OpenID Foundation announced that it had approved three Final Specifications: the Shared Signals Framework 1.0, the Continuous Access Evaluation Profile (CAEP) 1.0 and Risk Information Sharing and Coordination (RISC) 1.0. They are aimed at the time after login. The release says sessions "often last days or weeks" while a user's device compliance or access can change, and that systems relying on federated identity had no way to hear about it. The case it makes for adopting them now rests on its own word: it reports no deployment, interoperability result or implementer count. Its line that "Apple, IBM, Okta, and others, have already adopted these protocols" names no product, version or link, and Executive Director Gail Hodges says Final status "unlocks the ability of many governments to adopt the specifications" without naming a government.
- 2BI-FIND-2026-0182 · Six companies' agent-auth draft, since adopted by the WIMSE group, keeps credentials out of the model's reachThe draft calls observability "a security control, not solely an operational feature". Any participant MAY subscribe to Shared Signals Framework events under the CAEP or RISC profiles. A recipient of a revocation or risk signal SHOULD remediate and MUST ensure revoked or downgraded authorization is enforced without undue delay, and "Cached authorization decisions and tokens that are no longer valid MUST NOT continue to be used after a revocation or risk notification is received". Audit records MUST be tamper-evident and MUST record at least the agent identifier, the delegated subject when there is one, the resource, the action and decision, a timestamp with a correlation identifier, the posture or risk state, and any remediation. The draft does not say how the logs are made tamper-evident, and it leaves the policy model, its format and compliance criteria out of scope as specific to each deployment.
- 3BI-FIND-2025-0151 · Under Shared Signals 1.0 a sender may accept a subscription and send nothing, so agent revocation needs a backupThe framework never mentions agents, workloads or AI. Its subjects are "human or robotic principals, devices, customer tenants in a multi-tenanted service, organizational units within a tenant, groups of subject principals, or other entities", and its complex subject has members for user, device, session, application, tenant, org_unit and group. A complex subject must refer to exactly one principal, which we read as ruling out a single subject that names both an agent and the person it acts for. That pairing could travel as an extra subject member defined by an event type, which the specification allows, or in a proprietary identifier format the parties agree between them. The revocation events themselves come from elsewhere: session-revoked is a CAEP event and account-disabled a RISC one, and the only event types SSF defines itself are stream verification and stream-updated.
- 4BI-CALL-2026-0044 · The panel expects agent lifecycle management to fail first; Eve Maler wants assurance levels for agents in 2027By March 2028 the OpenID Foundation's Shared Signals working group publishes a draft or profile that carries agent-specific events, such as an agent's credential being revoked, its delegation withdrawn, or its session terminated, so an identity provider can push an agent's change of state to the systems it touches. Settles wrong if CAEP and RISC still describe only human sessions and devices by then.
- 5BI-FIND-2018-0008 · Under RFC 8417 a security event token reports a fact, not an order, and the receiver decides what to doRFC 8417 defines the token's structure and leaves the events, and much of how they are checked, to profiling specifications. Of the claims it takes from the JWT standard, iss, iat and jti are REQUIRED, aud is RECOMMENDED and sub is OPTIONAL. A profile MUST say how the subject is identified and which steps a recipient MUST run to check that a SET is valid, and it may decide whether the events claim can hold more than one member. The RFC names OpenID RISC and back-channel logout as example profiles "at the time of this writing". Shared Signals 1.0, a later OpenID profile of this RFC that we have read separately, requires explicit typing and forbids the sub and exp claims. It also lets each deployment or integration set how events are processed, from informational input to mandatory enforcement.
- 6BI-FIND-2026-1877 · Global Token Revocation must end a user's refresh tokens; their access tokens can outlive the callParecki's appendix compares the draft with RFC 7009, OpenID Connect's two logout specifications and the Shared Signals events, and says how each differs. RFC 7009 takes the token itself and is called by the OAuth client that holds it. OpenID Connect Front-Channel Logout works only while the user is in a web browser. Back-Channel Logout identifies the user only by an ID token's sub, assumes the application uses OpenID Connect rather than SAML, and, the draft notes, says refresh tokens issued with the offline_access scope "SHOULD NOT be revoked". A CAEP event is, in his words, "a signal that may or may not be acted upon by the receiver", and RISC lets users opt out. Global Token Revocation instead takes a user identifier in RFC 9493 formats such as an email address and is framed as a command with defined outcomes, whatever protocol the user signed in with.
›In our briefings
The claim each of the newest briefings on RISC would have you leave with.
Much of the new text in -01 says how an agent's access ends, in plain prose with no MUST or SHOULD. Section 8 now says a posture check at provisioning "establishes trust at the moment a credential is issued or renewed" and does not by itself show that posture still holds at a later request. It offers two existing remedies: shorter credential lifetimes, so posture is re-assessed at each renewal, or the event mechanisms of Section 11, so a posture change after issuance revokes or down-scopes the agent's credentials and tokens. Deprovisioning becomes a third phase after provisioning and rotation. At a minimum renewal stops and the short-lived credentials lapse; where the change must take effect before expiry, the issuer signals relying parties and the issuers of any derived credentials. A new Section 9.3 bounds a long-lived session or stream by the credential that opened it: the accepting party re-validates when that credential expires, and both sides end or down-scope the session on a revocation or risk signal. A retired agent's identifier must not be reassigned to a different agent.
WIMSE's agent draft adds how an agent's access ends and where policy engines fit, without a new requirementBI-FIND-2026-2029
A successful Global Token Revocation call does not promise that the user's access has ended. Parecki's draft requires the authorization server to revoke all of the user's active refresh tokens and to re-authenticate the user before issuing new tokens, but it only says the server SHOULD invalidate access tokens. Section 4 then makes access-token revocation optional, because self-contained access tokens may be impossible to revoke without extra measures such as a Token Status List, and it lets a server return success without revoking them. The 204 reply says revocation "has begun" and does not guarantee every token is gone; the draft's way to confirm completion is an optional Shared Signals event, CAEP Session Revoked.
Global Token Revocation must end a user's refresh tokens; their access tokens can outlive the callBI-FIND-2026-1876
The draft's sharpest rule reads "The Large Language Model MUST NOT have access to an agent's credentials or to credentials that may be needed to access tools and services", which it says keeps prompt injection from getting the model to use, expose or disclose them. Around it sit the identity rules. An agent MUST be assigned exactly one WIMSE identifier and MUST hold credentials cryptographically bound to it; those credentials SHOULD be short-lived and MUST carry an explicit expiry, and static API keys are "an antipattern for agent identity". The draft names WIT, X.509 WIC and SPIFFE SVIDs as formats and leaves the choice to the deployment. Our briefing on Sierra's cloud agent platform found a similar boundary built in: a proxy approves each call and swaps in the credential, so the agent never holds it.
Six companies' agent-auth draft, since adopted by the WIMSE group, keeps credentials out of the model's reachBI-FIND-2026-0179
›Cited in 7 briefings
Every briefing that discusses this term, newest first.
- 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
- Global Token Revocation must end a user's refresh tokens; their access tokens can outlive the call
Draft · · IETF Datatracker · Aaron Parecki, Okta · Feb 25, 2026
- 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
- A Shenzhen paper ties each AI agent to a verified owner and its exact code, at 15 to 40 seconds of proving a step
Paper · · arXiv · Zibin Lin et al. · Dec 19, 2025
- OpenID Foundation finalized Shared Signals, CAEP and RISC in September 2025, naming adopters without evidence
Article · · OpenID Foundation · Serj Hallam · Sep 16, 2025
- Under Shared Signals 1.0 a sender may accept a subscription and send nothing, so agent revocation needs a backup
Article · · OpenID Foundation · Atul Tulshibagwale et al. · Aug 29, 2025
- Under RFC 8417 a security event token reports a fact, not an order, and the receiver decides what to do
Article · · RFC Editor · Phil Hunt et al. · Jul 2018
Related terms
What appears alongside RISC in the same briefings.
SSFAgent Identity Management SystemCAEPSGNLRFC 8417 (Security Event Token)Zscaler
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.