| Internet-Draft | WIMSE Verifier Evaluation Semantics | October 2026 |
| Jackson | Expires 13 April 2027 | [Page] |
Delegation chain specifications describe the shape of conveyed authority. They leave the verifier's half of the exchange underdetermined. Two verifiers can check the same chain, both report success, and enforce different policy. This document states what a verifier must do: the explicit inputs evaluation depends on, how those inputs behave when their sources are stale or unavailable, and four rules that keep evaluation fail-closed. The rules are drawn from the Grant & Autonomy Lifecycle (GAL) and Provenance & Trust Context (PTC) specifications and from a public reference implementation.¶
This Internet-Draft is submitted in full conformance with the provisions of BCP 78 and BCP 79.¶
Internet-Drafts are working documents of the Internet Engineering Task Force (IETF). Note that other groups may also distribute working documents as Internet-Drafts. The list of current Internet-Drafts is at https://datatracker.ietf.org/drafts/current/.¶
Internet-Drafts are draft documents valid for a maximum of six months and may be updated, replaced, or obsoleted by other documents at any time. It is inappropriate to use Internet-Drafts as reference material or to cite them other than as "work in progress."¶
This Internet-Draft will expire on 13 April 2027.¶
Copyright (c) 2026 IETF Trust and the persons identified as the document authors. All rights reserved.¶
This document is subject to BCP 78 and the IETF Trust's Legal Provisions Relating to IETF Documents (https://trustee.ietf.org/license-info) in effect on the date of publication of this document. Please review these documents carefully, as they describe your rights and restrictions with respect to this document. Code Components extracted from this document must include Revised BSD License text as described in Section 4.e of the Trust Legal Provisions and are provided without warranty as described in the Revised BSD License.¶
A delegation chain is a claim about authority. A verifier turns that claim into a decision. Chain formats standardize the claim. This document standardizes the decision.¶
The failure this prevents is silent divergence. A verifier that drops an authorization entry it cannot parse and still approves has enforced nothing while reporting success. A verifier that derives "now" from the records it is judging keeps expired authority alive during quiet periods. A verifier that lets one broken path poison a valid one hands any attacker a veto over legitimate authority. Each of these was observed in running code during WIMSE implementation work. Each is fixed here with a stated rule.¶
This document does not define a chain format, a token envelope, or a policy language. It defines the evaluation function those formats are judged by. It is complementary to chain-format documents such as [DELEGATION-CHAIN] and [CONNECTED-FLIGHT]: those define what is conveyed, this defines how the receiver judges it.¶
The verifier specified here is deterministic. Its decision is a function of the inputs in Section 3, with no model and no judgment of content anywhere in it. In the terms of [PTC] it is the receiving boundary, which that specification calls the airlock: one party's enforcement point presents to another party's airlock, and how the receiver classifies the sender is the receiver's own deterministic decision ([PTC] Section 6.1). It is never a claim carried in the message and never the sending agent's to make.¶
[CONNECTED-FLIGHT] Section 4.1 maps its credentials onto the inputs in Section 3 and the rules in Section 4. It records the consumption rule of Section 4.3 as only partially addressed by those credentials, since scope and expiry bound a credential's validity without making it single-use. This document does not restate those profiles, which are that document's contribution.¶
This document is the verifier-side counterpart to executor-side work. [AEB] defines the boundary at which an executor admits, consumes authority for, and classifies the outcome of a consequential action, and [CAID] defines how two representations of one action are compared. Where a rule here needs those mechanisms, it states the property and points to them rather than defining its own.¶
Two of the rules in Section 4 are exercised in the public reference implementation [RI]: consumption keyed to the frozen call, and verification under role-separated keys. The other two are not. That implementation consumes no [RFC9396] authorization_details entries, and it has no derived-grant concept, so no call reaches it by a path; the lifecycle half of the per-path rule rests on GAL-39, which [GAL] marks normative and not yet implemented there. A rule's force comes from the specification that states it, and these notes record what has been exercised.¶
Several requirements here are stated ahead of that implementation. Issuer standing (Section 3.5), the direction in which clock skew is applied (Section 3.2), and the as-of time and maximum age of a fetched input (Section 3) are clauses of [GAL] (GAL-41, GAL-42, GAL-43) and of [PTC] (PTC-49, PTC-50), and each is marked there as not yet implemented. So is the caller-facing classification of a refusal (Section 5, PTC-44). Of what this revision adds to the fourth rule (Section 4.4), [GAL] marks continuity for records the issuer signs as not yet implemented. This document states each of these rules and makes no claim that the reference implementation demonstrates them. The key-bootstrapping mechanism discussed in Section 5 is the known open item; it is named as a gap rather than specified.¶
The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "NOT RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in BCP 14 [RFC2119] [RFC8174] when, and only when, they appear in all capitals, as shown here.¶
Frozen call: the exact action a human was shown and approved: capability, arguments, recipient, principal, and the authority the call was evaluated under, as presented at approval time.¶
Authorization instance: one approval bound to one frozen call, from the moment the call is held until a terminal disposition closes it.¶
Refuse: what a verifier does to a token or record that fails evaluation under Section 4. A refusal is the verifier's decision and involves no human.¶
Reject: what an approver does to a held frozen call as a terminal disposition. A rejection closes the authorization instance without executing the call. The expiry of an approval is not a rejection (Section 4.3).¶
Evaluation instant: the instant against which validity, including expiry, is judged.¶
On-behalf-of principal: the principal the chain acts under, distinct from the agent acting at each hop.¶
Principal: the identity tuple a grant is issued to: agent identity, skill, user, and trust tier. The unit of authority is the (principal, action class) pair. Those four attributes and the action class together name who acts, in what capacity, on whose behalf, at what trust, and over what class of action. The semantics of the trust tier are deployment-defined; this document standardizes the tier's position in the tuple, not its values.¶
Level: the position a grant holds on an ordered scale of autonomy. [GAL] Section 4.1 defines one such scale.¶
Record: a signed statement of one write to a grant.¶
Ledger: the append-only sequence of records for one (principal, action class) pair.¶
Issuer key: the key that signs the records an issuing ceremony writes. Those records raise a level (issuance, promotion), lower it voluntarily (tightening), or re-issue a grant at the level it already holds (re-attestation). It is the only key under which a level is raised.¶
Evaluator key: the key that signs the records the automatic evaluator writes (demotion, lapse), which may only lower a level. It is held by a deliberately separate identity.¶
As-of time: the instant as of which a verifier established an input it fetched from a source. It is the instant the verifier obtained the input from the source, or, where the source states when the input was current, the earlier of that time and the instant the verifier obtained it. Taking the earlier of the two resolves doubt toward less authority: a replayed answer carries its old time and ages out, and a source whose clock runs fast cannot make its answer look fresher than the moment it was fetched. (GAL-43, PTC-50)¶
Maximum age: the longest interval after its as-of time for which a fetched input may be used, pinned by the relying party for each input. The as-of time plus the maximum age is the input's horizon.¶
Delegation path: one chain of delegations from a root of trust to the acting agent.¶
Issuer standing: whether the party an issuer key speaks for still holds the role it issued under.¶
A verifier's decision MUST be a function of explicit inputs. Five inputs are required. Anything the verifier needs that is not in this list MUST be supplied explicitly; the verifier does not guess.¶
Some inputs come from sources the verifier does not control: status, keys, and records. Each fails in a direction. An input that confers authority only while fresh, such as a short-lived grant, fails toward less authority when its source is unavailable, because nothing renews it. An input that removes authority, such as a revocation entry or a demotion record, fails toward more, because its absence reads the same as good news. That property is what makes a lease [LEASES] safe under failure. The requirements in this section are written so that every fetched input fails toward less authority.¶
A verifier MUST NOT accept a status assertion, key set, or record set from a source after it has accepted a later one from the same source, where "later" is by that source's own ordering, such as a sequence number or an issuance time. Otherwise an attacker who can replay an earlier signed response restores a key or an issuer that the source has since withdrawn.¶
Every fetched input also has an age. For each input a decision depended on that was fetched from a source, the verifier MUST record the input's as-of time. A verifier MUST NOT use a fetched input at an evaluation instant at or past its horizon, which is its as-of time plus the maximum age the relying party pins for it. Past the horizon the input is not established. The verifier obtains it again, or treats it as unavailable, and an evaluation that needs an unavailable input ends in a transient refusal (Section 5). Without a maximum age, an input that was true when fetched stays in force for as long as the process that fetched it keeps running, so a key withdrawn or a standing lost after the fetch is still honored. Without a recorded as-of time, a later reader of the decision cannot tell one made on current inputs from one made on stale ones. (GAL-43, [GAL] Section 6.15; PTC-50, [PTC] Section 6.13)¶
Each input is judged on its own horizon. A decision that depended on several fetched inputs holds only while every one of them is inside its maximum age, and the as-of times are recorded per input, never as one time for the decision. The decision does have a horizon of its own, which is the earliest of its inputs' horizons: a decision is good no longer than its shortest-lived input. A verifier MUST record that instant, and a decision MUST NOT be relied on past it. A call held on the decision, or carried on from it to another party, is evaluated again. Otherwise a call is permitted and held, the key that verified its chain is withdrawn while it waits, and the hold is released on the strength of the earlier decision: every input was in date when it was read, and the action still runs on a withdrawn key. A record that keeps a resolved input without its horizon cannot separate a live admission from one that outlived its window. The horizon is a validity condition judged at the evaluation instant (Section 3.2), in the way an expiry is. Passing it is not a refusal by the source and is not evidence that the input changed. The as-of time and the maximum age are properties of how the inputs below are held; they are not a sixth input. How a record encodes them belongs to the record format, for which see [AUDIT-RECORD]. This document defines no encoding.¶
The frozen call is the exact action the human was shown and approved. Evaluation and consumption bind to that call as presented, not to a re-serialization, a paraphrase, or an identifier that names it. (GAL-36, GAL Section 4.1)¶
"As presented" needs a basis for comparison once the same action can arrive in more than one representation; otherwise two artifacts describing one action carry digests that cannot be compared without guessing. A verifier MUST compare a candidate call to the frozen call over the material fields declared for the action's type, under a canonicalization or mapping profile pinned by the relying party. It MUST NOT infer equivalence from field names, descriptions, or identifiers, and a comparison the pinned profile cannot make is a mismatch. [CAID] supplies both halves: action types that declare their material fields (Section 4.2), and an Action-Mapping Profile that the relying party pins and that yields INDETERMINATE, never a match, when the mapping would lose or cannot cover a material field (Section 8). This document requires the property and does not define a canonicalization of its own.¶
The authority a call runs under is part of what the approver approved. An approval covers the visible coordinates of the call under the authority shown at approval time, so two calls with the same visible fields under different authority are different calls. A release that runs under a different delegation than the one shown breaks what-you-see-is-what-you-execute even when every other coordinate matches. A verifier MUST be able to verify that the authority under which a call is released is equivalent to the authority approved. This document does not prescribe whether that equivalence is carried explicitly with the approval or reconstructed from immutable records of the frozen call.¶
The instant against which validity, including expiry, is judged is an explicit input to evaluation. It MUST NOT be derived from the timestamps of the records under evaluation. Deriving "now" from the latest record keeps expired authority alive through quiet periods, because time passes without producing records. (GAL-34)¶
The instant MUST come from a clock the verifier trusts and MUST NOT be taken from data the presenter controls. A presenter that supplies the instant chooses which grants are live. The verifier's own clock is also an attack surface: set back, it revives expired authority; set forward, it refuses everything. Authenticated time, such as Network Time Security [RFC8915], narrows that attack.¶
When the verifier compares the instant with a time asserted under another party's clock, it MUST apply the skew bound (Section 5) in the direction that confers less authority. An expiry or a term, and an authority-lowering record such as a demotion, a lapse, or a withdrawal of issuer standing, is treated as effective up to the bound earlier. An authority-raising record, and a not-before time, is treated as effective up to the bound later. A symmetric leeway lets skew extend authority past the point where it ended, and whoever benefits from a slow clock gets that window for nothing. (GAL-42, [GAL] Section 6.7.6; PTC-49, [PTC] Section 6.13)¶
A record carries more than one time, so which one the verifier compares needs stating. A record's timestamp is the instant its writer appended it, and nothing else. The instant a transition took effect is taken from the record's own typed field for that instant where the record type has one, and never from the timestamp. A lapse is the case that matters. Authority fell when the grant's term expired, and the lapse record is written at or after that instant, possibly much later. The effective instant of a lapse is the expired term, which the grant carries and the lapse record may carry beside its timestamp. A verifier MUST NOT infer from a record's timestamp the time of any event the record does not carry as a field of its own. A verifier that reads a lapse record's timestamp as the moment of the lapse honors the grant for the interval between the term and the write, and a verifier that waits for the record honors it until somebody writes one. (GAL-34, GAL-40, [GAL] Section 5.2, [GAL] Section 6.7.6)¶
Expiry is a validity condition judged at the evaluation instant: an approval, a grant, or a term has expired when the instant is at or after the time it states. Expiry is never a disposition, and it is never the approver's rejection (Section 4.3).¶
Conveyed evaluation material MUST include the on-behalf-of principal. A verifier that does not share the issuer's log cannot recover the principal from the action chain, so material that omits it authenticates an action without saying whom it was taken for. ([PTC] Section 6.6)¶
Evaluation is against the whole tuple, never the agent identity alone: a verifier that matches on agent identity while ignoring skill, user, or tier has evaluated a different grant. (GAL-36, [GAL] Section 5.1)¶
A party that assigns an identifier used in the tuple, or an identifier for an issuer, MUST NOT reassign it to a different subject. The relying party MUST pin, for each identifier namespace it joins on, an assigning party that meets this requirement. Standing, lapse, and audit records join to a chain by identifier. Where an identifier can be reassigned, a check of a departed issuer's standing returns the standing of whoever holds the identifier now. The subject identifier of [OIDC-CORE] Section 2 carries the same requirement.¶
Records are signed under one of two roles, and a record's type names its role. The issuer key signs the records an issuing ceremony writes: issuance and promotion, which raise a level; tightening, which lowers one voluntarily; and re-attestation, which re-issues a grant at the level it already holds after the declared bounds the grant is bound to have changed. The evaluator key signs the records the automatic evaluator writes, demotion and lapse, which may only lower a level. The evaluator is a deliberately separate identity. An identity MUST NOT hold both roles' signing keys. The separation is structural, not conventional: a grant-store write the agent can reach is a promotion bypass.¶
The two roles divide records by who writes them, and the direction of the change does not decide the role. Every record that raises a level is signed by the issuer key. Two kinds of issuer-signed record raise nothing: a tightening lowers, and a re-attestation changes no level at all. (GAL-37, [GAL] Section 4.3, [GAL] Section 6.7.2, [GAL] Section 6.10)¶
Whether a chain was valid when issued and whether it is valid when evaluated are different questions. A chain whose root or intermediate issuer has been disabled, or has lost the role it issued under, can carry valid signatures and unexpired grants after its authority is gone.¶
A verifier MUST establish the standing of every issuer on a path, at the evaluation instant, from a source the relying party pins and within a maximum age the relying party pins. A path through an issuer whose standing is withdrawn, or cannot be established within that age, is not fully valid, and Section 4.2 decides the token. Standing carried only in the presented token MUST NOT be accepted as its own source.¶
Standing is established by the verifier, at the evaluation instant, from sources the verifier holds and the relying party has pinned. The chain supplies the identities of the issuers on the path, and that is all the standing check takes from it. The check for each issuer is a lookup made beside the chain against those identities, and a lookup that cannot complete fails closed the way a timeout does. Standing therefore needs no claim surface in any chain format. A chain may name the actor at each hop, as the act claim of [RFC8693] Section 4.1 does. Those names are attribution: they tell the verifier which parties acted and confer no standing on any of them. A statement of standing that a chain does carry is not an input to this check.¶
Attribution is attested at each hop and composed afterward. A verifier records the actor it was presented with and the as-of time of that observation. It establishes standing for every issuer on the path at its own evaluation instant. It does not re-derive the attribution that an upstream verifier attested, and it does not vouch for it. Reconstructing a chain across hops composes the per-hop records.¶
The source is a deployment choice. A status list [STATUS-LIST] or a directory query satisfies it. So do grant lifetimes no longer than the pinned maximum age, where each renewal re-establishes the issuer's standing; that form needs no fetch at evaluation and fails toward less authority by construction. [DELEGATION-CHAIN] Section 9.5 describes the window between a withdrawal and a verifier's next fetch. The pinned maximum age is the bound on that window.¶
Standing is the one input that can require a live source at evaluation time, which is why it is listed rather than left to deployment.¶
[GAL] Section 6.10 states the rule for a stored grant (GAL-41): a verifier establishes that the key which signed the record a grant's level rests on has not been withdrawn, and a grant whose issuer has lost standing is enforced at the fallback level the grant itself declares. That fall belongs to the enforcer inside the grant's own domain, where the fallback level is known. The same section states the other case: a verifier in another domain holds no such level for the grant, because a conveyed chain declares no fallback, so a path through an issuer without standing confers nothing.¶
A verifier MUST process every authorization_details entry [RFC9396] presented or refuse the token. An entry of a type the verifier cannot evaluate is an entry it cannot process, and the token MUST be refused. A verifier that silently drops an entry enforces an incomplete policy while reporting success. That is fail-open. (A verifier advertises the types it supports with authorization_details_types_supported [RFC9728].)¶
Such a refusal is permanent for the presented token: no retry of the same token can succeed, because the caller cannot change which types the verifier supports. What the caller is told is stated in Section 5.¶
Where several delegation paths reach the same agent, the verifier MUST evaluate each path independently along its whole length. One fully valid path suffices for authority. A broken path MUST NOT neutralize a valid one, and a valid path MUST NOT excuse a broken one. Where no presented path is fully valid, the token confers no authority. Discovering which paths exist is the delegation mechanism's work; judging each presented path is the verifier's.¶
The lifecycle half of this rule is stated normatively in [GAL] as GAL-39: a derived grant does not outlive or out-rank the grant it derives from, and where a grant is demoted or lapses, every derived grant descending from it ceases to confer authority from that evaluation instant, whether or not a record of the change has been written. [GAL] Section 1.3 separates that half, which the grant standard owns, from path resolution, which it does not. The scope half, that one fully valid path suffices and a broken path never neutralizes a valid one, is this document's.¶
An approval MUST be consumed only by release or rejection of the frozen call it binds, by the whole principal the call was frozen for. Consumption MUST NOT be inferred from the approval's identifier appearing in any record. Otherwise any party able to write a record can burn another party's approval without using it. The release executes the stored call and nothing re-sent. (GAL-36, [GAL] Section 4.1)¶
Rejection here means the approver's terminal disposition, the one that closes the authorization instance. A non-terminal disposition, such as a deferral or a request for changes, is neither release nor rejection: the approval stays pending and nothing is consumed. Each authorization instance has at most one consumption event, and it is the release or the rejection of its frozen call.¶
This document bounds release and does not bound pending. A held call confers no authority while it is pending, so an approval that stays pending without limit fails closed by construction. Where an approval carries an expiry, the expiry is a validity condition judged at the evaluation instant (Section 3.2): a release presented at or after it is refused. Expiry is not a disposition. It consumes nothing, and it is not the approver's rejection, because the approver decided nothing. The verifier's record of an expired approval says that it expired ([PTC] Section 6.12 states the record-side rule as PTC-53). Whether a pending instance is given a lifetime at all belongs to the lifecycle of the instance, at the authorization server or in the grant standard, and this document sets none. Appendix A.1 gives the cases.¶
Consumption keys on the closure of the authorization instance, never on the outcome of an execution. Whether a released call executed, failed, or ended in an indeterminate state is a separate dimension, which executor-side documents classify ([AEB] Section 5.13). Once a release has begun execution, an indeterminate outcome MUST NOT return the approval to pending, because a retry under it could duplicate an effect that did occur. [AEB] Section 5.11 describes the executor-side counterpart of this rule, an atomic consume-or-reserve before invocation.¶
Where the verifier cannot establish that an approval is unconsumed, because its consumption state is unavailable, it MUST refuse. Restoring consumption state from an earlier point MUST NOT make a consumed approval presentable again. A deployment that restores either carries the consumed set forward from a source the restore did not roll back, or refuses every unexpired approval issued before the restore point. [AEB] Section 10 describes the same effect for a change of authority namespace, which makes consumed grants look unused.¶
This document does not standardize approval interaction. It standardizes what the verifier must check about consumption.¶
A verifier MUST select the acceptable signing keys from the record type, MUST refuse a record signed by the other role's key, and MUST refuse a key resolvable under both roles. The key that grants authority is never the key that judges it. (GAL-37)¶
A signature under the right key says who wrote a record. It does not say the record is one that writer may write. Two further checks apply to a correctly signed record, and each closes a way of raising authority that the key check passes.¶
Direction: a verifier MUST refuse a record signed under the evaluator role whose resulting level ranks above its starting level. The evaluator runs with no human in its path, and its key exists to lower. A demotion or a lapse that raises a level is authority minted by the side that must never mint it, and it verifies under the evaluator key exactly as an honest demotion does. An honest evaluator writes no such record, so this check only ever meets one that was planted in the store or written with a stolen evaluator key. (GAL-17, [GAL] Section 4.3)¶
Continuity: a verifier that reads a grant's ledger MUST refuse a record, under either role, that does not start from the level the ledger held immediately before it. The first record of a ledger starts from no level. The direction check does not cover this. A demotion from the highest level to the middle one lowers on its own terms, and placed on a ledger that stood at the lowest level it leaves the derived level raised under the evaluator's signature. For the issuer's records, the ceremony reads the level it moves from, so a break in continuity marks a record that did not come through the ceremony, and a promotion that restates its starting level can hide a skipped step. (GAL-37, [GAL] Section 6.10)¶
A re-attestation record changes no level. A verifier that derives a grant's level from its ledger, or that looks for the record a level rests on, MUST pass over a re-attestation record: the level and the record that earned it are the ones the ledger held immediately before it. Without this, two verifiers read one ledger and name different records as the one that earned the level. The one that names the re-attestation checks the standing of a different signer (Section 3.5), and it finds no term on that record, where the promotion carries the term its ratifier agreed to. ([GAL] Section 4.3, [GAL] Section 6.6)¶
A grant whose ledger contains a record refused under this section has no level the verifier can derive. A path through that grant is not fully valid, and Section 4.2 decides the token.¶
The level a verifier evaluates a grant at is the level its records derive. [GAL] Section 6.6 requires that no write to a grant go unrecorded and that a re-attestation always be signed under the issuer role (GAL-15), so grant state that no record accounts for is the trace of a write made outside the ceremony. It is not evidence of authority. A verifier that holds both a grant and its ledger MUST evaluate the grant at the lower of the level the grant states and the level its records derive, and MUST record the difference where there is one ([GAL] Section 6.11). Without this rule, one verifier trusts the stated level and another the derived one, and the two reach different decisions on the same evidence, which is the divergence this document exists to remove.¶
Clock skew: the evaluation instant is supplied by the evaluating party as an explicit input. When two parties do not share a clock, the party evaluating supplies the instant. The clock source and the skew bound are deployment details, and a deployment MUST document which clock supplies the evaluation instant and the maximum skew it tolerates. This document sets no default bound; Section 3.2 fixes the direction in which the bound is applied.¶
Key bootstrapping across boundaries: a receiving verifier needs the sending boundary's issuer key before any rule in Section 4 can run.¶
Open issue: this document does not specify how the verifier comes to hold the sending boundary's issuer key. That is three questions, and they are separable:¶
A discovery mechanism can answer the first without answering the second, and a binding that answers the second still leaves the third to the relying party's own policy. The reference implementation answers all three at once by provisioning keys out of band, which requires a prior arrangement with each peer. This document requires only that the verifier hold a key it has decided to trust before any rule in Section 4 runs; the mechanism for each question is left for the working group.¶
Refusals: a refusal is either permanent for the presented token or transient. It is permanent when no retry of the same token can succeed: an unsupported entry type (Section 4.1), a signature that does not verify, an expired grant or approval, a record refused under Section 4.4, an issuer whose standing is withdrawn. It is transient when the verifier could not evaluate because an input it needs was unavailable: a standing source, a key source, or consumption state. An input past its maximum age that the verifier could not obtain again is unavailable in the same sense (Section 3).¶
Toward a caller the verifier authenticated before evaluating, and that it serves, a refusal MUST be distinguishable as permanent or transient. A transient refusal reported as permanent stops a caller whose token would succeed later, and a generic failure is retried at length. Toward any other caller the verifier MUST NOT make the two distinguishable. A refusal on the content of a message is not an outcome of evaluation under this document.¶
The first requirement was a recommendation in earlier revisions. A recommendation is how a transient refusal comes to be reported as permanent in practice, and the harm of that misreport falls on a caller that did nothing wrong: it abandons a token that would have succeeded, or it retries a token that never will, which costs both parties availability. The second requirement exists because of who else can ask. To a party probing the verifier without being one of its callers, any difference between two refusals is a map of the verifier's configuration, and an answer of "transient" reports that an attack on a key source or a status source is working. [PTC] Section 6.1 states the same rule for the receiving boundary (PTC-44), together with the rule this document leaves to it: a refusal on content looks the same as acceptance to every sender.¶
Where the transport already defines a code, that code is the one to use toward a caller the verifier serves. At the authorization server, [RFC9396] Section 5 defines invalid_authorization_details for an unknown authorization details type, and [RFC6749] Section 4.1.2.1 defines temporarily_unavailable. A resource server using bearer tokens has invalid_token [RFC6750] for a permanent refusal, and status code 503 with a Retry-After field [RFC9110] for a transient one. This document defines no new code and no response format.¶
Whether a refusal is permanent or transient describes the verifier's state, not the caller's chain, so disclosing it to a caller the verifier serves is consistent with the rule below. The verifier's own record of a refusal MUST distinguish a token it evaluated and refused from a token it could not evaluate, and MUST name the input that was unavailable. Collapsing the two makes every missing input look like a policy decision to whoever reads the record later.¶
The record member that carries this distinction is the evaluation member of [AUDIT-RECORD] Section 5.4, and this document uses that vocabulary and defines none of its own. A transient refusal is recorded with a status of not-evaluated and names each unavailable input from the closed set that document enumerates. A permanent refusal is recorded with a status of evaluated and a reported decision of deny. The names this section gives to unavailable inputs (standing source, key source, consumption state) are members of that set. The requirement in the preceding paragraph can be checked in a record only because that member exists: without it, the record of a verifier that lost an input and refused is identical to the record of a refusal it evaluated. [EVAL-STATE] states the same distinction as requirements independent of encoding, for agent protocol decisions beyond verifier evaluation.¶
The refusal MUST NOT disclose which delegation path or grant failed evaluation, or the standing of any issuer. A verifier consumes tokens from another trust domain, and at that boundary an informative error is an oracle: naming the path that broke, or the parent grant that lapsed, lets the caller map the delegation graph and enumerate grant state it cannot otherwise see. The rule is to disclose properties of the verifier, which are public, and never the evaluated state of the caller's chain. The detail belongs in the verifier's audit record, where the operator reads it. A caller may still believe it holds authority after a grant above it has lapsed; the verifier does not owe it a correction.¶
An unsupported entry is a different case. The entry types are in the caller's own token and the supported types are published (Section 4.1), so whenever one entry carries a type the verifier does not advertise, the caller can already tell which entry it was. Naming that entry discloses nothing further, and this document neither requires nor forbids it.¶
Timing carries the same information as content. A verifier that stops at the first broken path answers sooner for some chains than for others, and the difference can locate the break. Where chain state is sensitive, a verifier evaluates every path before it answers. The same holds for the difference between permanent and transient toward a caller that is not to learn it: a refusal that waited on a source answers later than one that did not.¶
Dependency availability: every source a verifier fetches from is also a way to stop it. An attacker who can make a standing source, a key source, or consumption state unavailable turns fail-closed evaluation into a denial of service for every chain that depends on that source. This document does not relax fail-closed to avoid that. Whether an agent that cannot act is the safe outcome depends on what the agent does, so the response (redundant sources, caching within the pinned maximum age, alerting) is deployment policy.¶
Audit availability: where the verifier cannot write its own record, the detail this section sends there is lost. A deployment MUST document whether a verifier that cannot write its record continues to admit.¶
Standing and human presence: issuer standing establishes that an issuer still holds its role. It does not establish that a human was present when the issuer key signed; a signature establishes control of a key ([AEB] Section 10). A deployment that needs to know how the issuer authenticated carries that separately, for example as the acr and amr claims of [OIDC-CORE] Section 2.¶
Key scope: a key the verifier knows is not thereby a key that may speak for every subject. Which party and which identities a key may sign for is the verifier's own configuration, established before evaluation and never taken from the chain. (PTC-46, [PTC] Section 6.7)¶
Key custody: a signature shows that a key was used and does not show where the key is held. A chain looks the same whether or not the signer's agent can reach the signing key, so what a verifier knows about custody is a record in its own configuration, and a record its operator declared is not a check performed on the signer. (PTC-48, [PTC] Section 7.1)¶
Structural separation: the agent MUST NOT hold signing keys and MUST NOT have write access to the grant store. Verification authenticates lineage; it does not clean taint and does not substitute for the receiver's own trust map. ([GAL] Section 6.7.2, [PTC] Section 6.6, [PTC] Section 6.7)¶
Misconfigured verification, such as a configured but unusable key source, MUST fail closed loudly and MUST NOT silently degrade to no verification. ([PTC] Section 6.7)¶
This document makes no request of IANA.¶
These cases restate requirements of Sections 3, 4 and 5 as checks. They add no requirement. Each gives the expected verdict and the expected reason, and a verifier that returns the verdict for a different reason has not passed the case: a verifier that refuses everything returns "refuse" wherever it is expected. The reason is what the verifier's own record carries (Section 5). The caller learns at most whether a refusal is permanent or transient. The reasons are written as prose; this document defines no reason codes.¶
One held call is pending under an approval that expires at E. Every other input is valid. B is the pinned skew bound where E was asserted under another party's clock, and zero where it was asserted under the verifier's own. The evaluation instant t comes from the verifier's trusted clock and from nothing the presenter supplies (Section 3.2). The same call is presented for release in each case.¶
| Case | Condition | Verdict | Reason the record carries |
|---|---|---|---|
| E1 | t earlier than E minus B | release | The approval is valid at t and unconsumed. The release is the one consumption event of the instance. |
| E2 | t at or after E | refuse, permanent | The approval expired at the evaluation instant. Recorded as evaluated. Not a rejection: no approver disposition exists, and nothing is consumed. |
| E3 | B above zero; t at or after E minus B and earlier than E | refuse, permanent | As E2. The bound is applied toward less authority. |
| E4 | As E2; every timestamp in the presented records is earlier than E minus B | refuse, permanent | As E2. A verifier that releases derived the instant from the records. |
| E5 | As E2; the presenter asserts a current time earlier than E minus B | refuse, permanent | As E2. A verifier that releases took the instant from the presenter. |
| E6 | The E1 presentation repeated after the E1 release, t still earlier than E minus B | refuse, permanent | The approval was consumed by the release. Not expiry. |
| E7 | As E1; consumption state unavailable | refuse, transient | Could not evaluate; consumption state unavailable. Recorded as not-evaluated. Not expiry and not consumption. |
E1 and E2 are the pair that matters: the same call, just before and just after its approval expires, against an instant supplied independently of the call. E4 and E5 hold the instant after expiry while everything the presenter controls says otherwise.¶
A grant has a ledger. Levels are written L1, L2 and L3, lowest first. Every record in these cases carries a signature that verifies under the key named. In R1 to R4 the grant has no level the verifier can derive, a path through it is not fully valid, and a refusal that follows is permanent (Section 4.4).¶
| Case | Record | Verdict | Reason the record carries |
|---|---|---|---|
| R1 | A demotion from L1 to L2, signed by the evaluator key | refuse the record | Direction: a record under the evaluator role raises a level. Not a signature failure. |
| R2 | The ledger stands at L1. A demotion from L3 to L2, signed by the evaluator key | refuse the record | Continuity: the record does not start from the level the ledger held. Not direction, since the record lowers on its own terms. |
| R3 | The ledger stands at L1. A promotion from L2 to L3, signed by the issuer key | refuse the record | Continuity, as R2. |
| R4 | A promotion signed by the evaluator key | refuse the record | The record is signed by the other role's key. |
| R5 | The ledger holds a promotion to L2, then a re-attestation at L2 by a different signer | level L2 | The record the level rests on is the promotion. The re-attestation is passed over. A verifier that names the re-attestation fails the case although it reports L2. |
| R6 | The grant states L3. Its records derive L2, and none is refused | evaluate at L2 | No record accounts for L3. The difference is recorded. |
The standing of one issuer was fetched from a pinned source with an as-of time of A. The pinned maximum age is M, so the horizon is A plus M. Every other input is valid.¶
| Case | Condition | Verdict | Reason the record carries |
|---|---|---|---|
| G1 | t earlier than A plus M; standing held | admit | Standing established as of A. The record carries A for this input. |
| G2 | t at or after A plus M; the source answers, standing held | admit | Standing established as of the new fetch. The record carries the new as-of time, not A. |
| G3 | t at or after A plus M; the source does not answer | refuse, transient | Could not evaluate; standing source unavailable. Recorded as not-evaluated. Not a withdrawal of standing. |
| G4 | t earlier than A plus M; standing withdrawn as of A | refuse, permanent | Issuer standing withdrawn. Recorded as evaluated. |
| G5 | The source's answer is earlier, by the source's own ordering, than one already accepted from it | answer not accepted | Earlier than an accepted answer from the same source. The accepted answer stays in force until its own horizon. |
Theo Adam contributed the test vector for the consumption rule and the argument that the authority a call runs under belongs in the frozen call, and separated closure of an authorization instance from the outcome of its execution. Iman Schrock raised the terminal meaning of rejection, the comparison basis for "as presented", and the three-way split of key bootstrapping. Girish Konda raised what a refused caller is told, the collision between refusal and rejection, and the difference between disclosing a failed entry and disclosing a failed path. Cliff Singletary raised issuer standing as an evaluation input, the difference between a denial and an evaluation that could not complete, and identifiers that survive re-provisioning.¶
For this revision: Errol Seymour raised whether a pending approval is bounded and whether issuer standing needs a claim surface in the chain. Iman Schrock raised that expiry is never the approver's rejection, and the review case of one held call presented on either side of its approval's expiry. Girish Konda raised that the caller-facing classification of a refusal should be a requirement, and that the protocol side and the record side of a decision with unresolved inputs have to change together, and the horizon for which a resolved input is good. Sankalp Gilda defined the evaluation member of the audit record, on a construction Cliff Singletary proposed. Kieran Sweeney raised the boundary between attribution carried in a chain and authority. Jijie Wei raised a third outcome for a decision with unresolved inputs, and an as-of time for each input established at a different time, and that a decision has a horizon of its own, no later than that of its shortest-lived input.¶
This document was prepared with AI assistance for drafting, source verification and editorial review. The technical positions are the author's.¶