<?xml version='1.0' encoding='utf-8'?>
<!DOCTYPE rfc [
  <!ENTITY nbsp    "&#160;">
  <!ENTITY zwsp   "&#8203;">
  <!ENTITY nbhy   "&#8209;">
  <!ENTITY wj     "&#8288;">
]>
<?xml-stylesheet type="text/xsl" href="rfc2629.xslt" ?>
<!-- generated by https://github.com/cabo/kramdown-rfc version 1.7.43 (Ruby 3.4.12) -->
<rfc xmlns:xi="http://www.w3.org/2001/XInclude" ipr="trust200902" docName="draft-meunier-webbotauth-capacity-management-00" category="info" submissionType="IETF" tocInclude="true" sortRefs="true" symRefs="true" version="3">
  <!-- xml2rfc v2v3 conversion 3.35.0 -->
  <front>
    <title abbrev="PACE">PACE: Capability Management for Agents</title>
    <seriesInfo name="Internet-Draft" value="draft-meunier-webbotauth-capacity-management-00"/>
    <author fullname="Thibault Meunier">
      <organization>Cloudflare</organization>
      <address>
        <email>ot-ietf@thibault.uk</email>
      </address>
    </author>
    <author fullname="Christopher A. Wood">
      <organization>Cloudflare</organization>
      <address>
        <email>caw@heapingbits.net</email>
      </address>
    </author>
    <date year="2026" month="November" day="15"/>
    <keyword>anonymous rate limiting</keyword>
    <abstract>
      <?line 80?>

<t>This document addresses the noisy neighbour problem without requiring a
stable client identifier at the origin. Separately allocated budgets let
origins limit consumption beneath a shared platform identity. It describes
PACE: a framework for managing budget continuity across requests using
a composition of the Web Bot Auth with the Moderation of unLinkable Endorsements
(MOLE) architecture <xref target="MOLE-ARCH"/> and protocols <xref target="MOLE-PROTOCOLS"/>. This composition allows for
per-client capacity management in a way that allows origins to serve clients
above an otherwise constrained public baseline.</t>
    </abstract>
    <note removeInRFC="true">
      <name>About This Document</name>
      <t>
        The latest revision of this draft can be found at <eref target="https://thibmeu.github.io/draft-capacity-management/draft-meunier-webbotauth-capacity-management.html"/>.
        Status information for this document may be found at <eref target="https://datatracker.ietf.org/doc/draft-meunier-webbotauth-capacity-management/"/>.
      </t>
      <t>Source for this draft and an issue tracker can be found at
        <eref target="https://github.com/thibmeu/draft-capacity-management"/>.</t>
    </note>
  </front>
  <middle>
    <?line 91?>

<section anchor="intro">
      <name>Introduction</name>
      <t>Agents are increasingly used in business and user workflows, including AI
assistants, search indexing, content aggregation, and automated testing. Their
relationship with origins is symbiotic. They need information from origins, and
origins need signals to apply rate-limit policies. These can be binary, like
allow or block when an origin cannot serve traffic, or resource-based to
allocate capacity among clients.</t>
      <t>Historically, most of these signals have been enshrined into deployments of
internet protocols. An agent needs an IP address to receive a response from an
origin. Its TLS stack and browser behavior provide further signals. It provides
a familiar <tt>User-Agent</tt> because middleboxes might otherwise filter its traffic
(see <xref target="USER-AGENT-HISTORY"/>).</t>
      <t>Recent developments break these patterns. Agents use intermediaries to drive
traffic, including running workloads in cloud platforms or by using connection
proxies (<xref target="CONNECT-UDP"/>, <xref target="CONNECT-IP"/>, <xref target="OHTTP"/>, or plain <xref target="HTTP"/>).
As workloads shift to platforms making requests on behalf of users, traditional
signals used for moderating traffic are no longer viable.</t>
      <t>An address, including an <xref target="IPv6"/> address, is not necessarily a user. Blocking
a shared IP range can affect everyone behind it. A platform signing requests under one
<xref target="WBA-HTTPSIG"/> identity creates a similar shared fate: when one tenant's
agent overloads an origin, limiting the platform also limits its other
tenants. <xref target="SIGNED-AGENT-METADATA"/> describes this problem and proposes
signed product, project, and session identifiers to let an origin isolate
traffic below the platform level.</t>
      <t>More granular identifiers are useful when the origin needs to recognize a
particular product or project, investigate abuse with its platform, or follow
a task within a session. <xref target="SIGNED-AGENT-METADATA"/> advocates for such
identifiers. While they are pseudonymous, the receiving origin can link
requests within the scope and lifetime of those identifiers. This has several
privacy costs:</t>
      <ul spacing="normal">
        <li>
          <t>Discriminatory treatment: Explicit identifiers allow for discriminatory
treatment of traffic based on those identifiers. An origin can give
preferential treatment to some products or projects, or block them
regardless of their requests' behavior.</t>
        </li>
        <li>
          <t>Linkability: Explicit identifiers allow for linking traffic to that identifier.
The scope and lifetime of the identifier, together with other request data,
determine how much activity the origin can associate with it.</t>
        </li>
        <li>
          <t>Relationship disclosure: Explicit identifiers reveal information about the
relationship between an intermediary and its clients, such as which requests
belong to the same project. Clients may have no clear view of how that
information will be processed, stored, or shared.</t>
        </li>
      </ul>
      <t>This presents a natural tension between the needs of an origin and the needs
of a user. Identifiers give an origin a way to apply capacity and abuse
controls at the granularity provided. This depends on the platform assigning
them reliably. Agents using those controls bear the cost of attribution. An origin
has to be able to apply policies to traffic whether it finds it abusive or
acceptable. Bounding consumption does not always require more fine-grained
identifiers.</t>
      <t>This document describes PACE, a framework for managing capacity for agents with
a combination of Web Bot Auth and MOLE, an architecture for private moderation <xref target="MOLE-ARCH"/>,
in a way that does not reduce to the shared fate, noisy neighbor problem.
In PACE, budgets are allocated to clients through MOLE, where a client can can
be a tenant on a platform, a tenant or agent workload, or a user behind any of
these, and origins moderate capacity based on those budgets. Traffic is similarly
attributed to platforms through Web Bot Auth signatures. Composing these two
technologies balances the needs of all parties: origins can establish policies on
a per-platform and then per-client basis within that platform, choosing the
granularity of client using their anchor set.</t>
      <t>Avoiding unnecessary disclosure supports the end-user interests discussed in
<xref target="INTERNET-FOR-END-USER"/>. For these capacity decisions, the origin need not
receive product, project, or session identifiers. When used, a platform
signature remains visible, along with public spending terms and other request
data. This approach does not prevent discrimination based on that information.</t>
      <t>Such a mechanism is useful in a number of cases. Some examples follow:</t>
      <ol spacing="normal" type="1"><li>
          <t>An AI agent service that hosts remote agents and concentrates their traffic
on a small set of IP addresses. One user's activity can affect others when
an origin can only limit the shared address or the platform as a whole.</t>
        </li>
        <li>
          <t>Network operators carry traffic from device-resident agents behind shared
IP ranges. Policies applied to an address range or an entire ASN give these
users a shared fate.</t>
        </li>
        <li>
          <t>VPN providers and enterprise zero trust gateways have a noisy neighbour
problem. They may know their users, but should not have to expose them to
each origin. An origin able to tell employees apart could single out an
executive's activity for instance.</t>
        </li>
      </ol>
      <section anchor="related-work">
        <name>Related work</name>
        <t><xref target="SIGNED-AGENT-METADATA"/> addresses shared fate under a platform's Web Bot
Auth signature with signed, pseudonymous sub-identifiers. It lets an origin
apply policy to a product, project, or session while the platform can map
those identifiers back to its customers. Origin-scoped session values
reduce cross-origin correlation (<xref section="7.2" sectionFormat="of" target="SIGNED-AGENT-METADATA"/>). However,
each identifier still links requests within its own scope and lifetime.
As described in <xref target="scenarios"/>, this document does not specify tenant-specific
abuse reporting. An Anchor can constrain project allocations, but the origin
cannot recognize
and selectively block all of a project's independent Credentials from their
presentations alone.</t>
        <t><xref target="WBA-ABA"/> also uses MOLE for anonymous bot access and rate limiting.
Its deployment relies on Anchor vetting and origin acceptance of issuance
policies (Sections <xref target="WBA-ABA" section="3" sectionFormat="bare"/> and <xref target="WBA-ABA" section="4.1" sectionFormat="bare"/> of <xref target="WBA-ABA"/>). It also discusses convergence on
accepted policies (<xref section="4.3" sectionFormat="of" target="WBA-ABA"/>) and the risk of cheap registration
yielding multiple credentials (<xref section="8.2" sectionFormat="of" target="WBA-ABA"/>). These are governance
dependencies between Anchors and sites. This approach shares those dependencies
and focuses on allocations beneath a platform's identity, where the platform
can use its existing tenant relationships to constrain new budgets.</t>
        <t>Poppy <xref target="POPPY"/> uses User IDs and sessions, including for anonymous browsing.
<xref target="AAP"/> uses anonymous admission tokens and recommends a session afterwards,
though it also allows a fresh token per request. Here, the aim is to maintain a
scarce, independently allocated budget without linking its presentations.</t>
        <t><xref target="WIMSE-AIMS"/> profiles WIMSE and OAuth for agents within a defined trust domain.
It treats an agent as a workload with its own identifier and credential. It
carries the user, when there is one, only as the subject of an OAuth access token, so that the two
identities are separate and compose on one request. <xref target="WAG"/> lets a
platform-hosted workload obtain such a token for itself without per-workload
registration. This document places both among the options an origin can ask
for, next to attribution and anonymous capacity.</t>
      </section>
    </section>
    <section anchor="terminology">
      <name>Conventions and Definitions</name>
      <t>This document uses the HTTP terminology of <xref target="HTTP"/> and the MOLE terminology
of <xref target="MOLE-ARCH"/>. The roles can be operated by separate parties or combined.</t>
      <section anchor="parties">
        <name>Parties and hosting relationships</name>
        <dl>
          <dt><strong>Platform:</strong></dt>
          <dd>
            <t>A service that hosts workloads for tenants.</t>
          </dd>
          <dt><strong>Tenant:</strong></dt>
          <dd>
            <t>A customer organization, developer account, or billing workspace that runs
workloads on a platform, often organized as a project.</t>
          </dd>
          <dt><strong>Workload:</strong></dt>
          <dd>
            <t>A deployed application or service, as described in <xref target="WIMSE-ARCH"/>.
A software product or service can comprise several workloads.</t>
          </dd>
          <dt><strong>Agent:</strong></dt>
          <dd>
            <t>Software that performs tasks, potentially on behalf of a user.
A session is a bounded task execution or workflow run, not the agent itself.</t>
          </dd>
          <dt><strong>User:</strong></dt>
          <dd>
            <t>A person on whose behalf an agent may act.</t>
          </dd>
        </dl>
        <t>A typical hosted deployment has this hierarchy:</t>
        <artwork><![CDATA[
Platform
  Tenant
    Workload
      Agent
]]></artwork>
        <t>A tenant runs workloads that execute agents. Users can use several agents,
and agents can serve several users or run unattended. Device-resident agents
need no hosting tenant.</t>
      </section>
      <section anchor="mole-roles-and-budget-scopes">
        <name>MOLE roles and budget scopes</name>
        <dl>
          <dt><strong>Client:</strong></dt>
          <dd>
            <t>The MOLE participant that obtains an Endorsement, redeems it for a
Credential, and presents that Credential when requesting a resource.
An agent or a shared workload component can implement this protocol role.</t>
          </dd>
          <dt><strong>Anchor:</strong></dt>
          <dd>
            <t>The party that grants Endorsements under a policy the Moderator accepts.
Its relationship with a Client constrains access to new budgets.</t>
          </dd>
          <dt><strong>Moderator:</strong></dt>
          <dd>
            <t>The party that issues Credentials and verifies their presentations. It
sets the budget policy and may be operated by the origin or on its behalf.</t>
          </dd>
          <dt><strong>Origin:</strong></dt>
          <dd>
            <t>The server that controls access to the requested resource.</t>
          </dd>
          <dt><strong>Intermediary:</strong></dt>
          <dd>
            <t>A proxy, gateway, or tunnel on the request path. An intermediary that
terminates HTTP can sign forwarded requests under its own identity.</t>
          </dd>
        </dl>
        <t>A Client may hold several budgets, and an account or workload may receive
budgets for several Clients. Those allocations depend on Anchor and Moderator
policy. A Credential does not establish a unique account, workload, device,
or other principal.</t>
        <t>MOLE(X) denotes a presentation for budget scope X. MOLE(C) is a Client's
budget, while MOLE(T), MOLE(W), and MOLE(U) denote tenant, workload, and
account scopes. These labels do not appear as identifiers in presentations.
WBA(Y) denotes a Web Bot Auth signature from intermediary Y.</t>
      </section>
      <section anchor="endorsement-examples">
        <name>Endorsement examples</name>
        <t>An accepted Anchor could allocate Endorsements at each scope:</t>
        <ul spacing="normal">
          <li>
            <t>Platform: A vetting service grants a hosting provider a finite pool.</t>
          </li>
          <li>
            <t>Tenant: The platform caps a retailer's workspace allocation under its
subscription, as in <xref target="scenarios"/>.</t>
          </li>
          <li>
            <t>Workload: A deployment service authenticates the shopping-assistant workload
and caps its allocation across replicas and restarts.</t>
          </li>
          <li>
            <t>Agent: The workload operator allocates a shopping agent's task budget
against the workload's allowance. New tasks do not reset that allowance.</t>
          </li>
          <li>
            <t>User: An account provider authenticates a customer and caps allocations
across that account's agents. This limits accounts, not unique people.</t>
          </li>
        </ul>
        <t>Clients redeem these Endorsements for Credentials. The Anchor accounts for the
allocation scope, which presentations do not reveal. Capacity evidence does
not grant permission. Combining budget proofs remains undefined (<xref target="composition"/>).</t>
      </section>
    </section>
    <section anchor="scenarios">
      <name>Budget limits and abuse</name>
      <t>Consider the following scenario. A platform hosts agents for two tenants. Tenant A sends excessive requests
while tenant B runs a shopping agent within its allocation. Both tenants'
requests carry the same platform signature. Blocking that signature would
interrupt both tenants. The origin needs to bound A's consumption and stop
abusive requests without requiring either tenant to disclose a stable
identifier.</t>
      <t>Without per-tenant budgets, the origin sees only the shared platform signature
as attribution evidence. If A exhausts the platform's aggregate allowance, the
origin also limits B, even though B remains within its intended allocation.</t>
      <t>Now suppose the platform also acts as an Anchor accepted by the origin. It grants
at most two Endorsements per tenant per epoch. Each Endorsement can be redeemed
once at this Moderator for an EC-ACT Credential with ten credits. The Moderator
charges one credit per accepted operation attempt and returns an updated
Credential with the remaining balance. It returns no credits and offers no
top-ups. Credentials expire at the epoch boundary.</t>
      <t>In this scenario, credits represent capacity granted by the Moderator under its
own access policy. Twenty credits correspond to the capacity allocated to a
tenant per epoch. A system based on fingerprints might instead apply a rate limit
per IP address, <tt>User-Agent</tt>, or TLS fingerprint.</t>
      <t>Each tenant's allocation therefore permits at most twenty budgeted operation
attempts in the epoch. Starting another task does not reset the allowance.
This bound depends on the Anchor enforcing the allocation and on other
accepted Anchors not supplying additional budgets. The origin verifies each
spend without learning which tenant's allocation it consumes.</t>
      <t>If the origin detects abuse on A's presented request, its Moderator can record
<tt>NoUpdate</tt> and end that credential chain. The chain is the sequence of
Credentials obtained by updating one initial Credential. Its remaining credits
are forfeited. B's independently allocated Credentials remain usable.
<xref target="fig-budget-ban"/> shows this decision. The tenant labels explain the example and
are not identifiers carried in presentations.</t>
      <figure anchor="fig-budget-ban">
        <name>Ending an abusive chain while an independent budget remains usable</name>
        <artset>
          <artwork type="svg"><svg xmlns="http://www.w3.org/2000/svg" version="1.1" height="480" width="496" viewBox="0 0 496 480" class="diagram" text-anchor="middle" font-family="monospace" font-size="13px" stroke-linecap="round">
              <path d="M 8,32 L 8,80" fill="none" stroke="black"/>
              <path d="M 8,128 L 8,176" fill="none" stroke="black"/>
              <path d="M 8,272 L 8,336" fill="none" stroke="black"/>
              <path d="M 8,416 L 8,464" fill="none" stroke="black"/>
              <path d="M 120,80 L 120,120" fill="none" stroke="black"/>
              <path d="M 120,176 L 120,192" fill="none" stroke="black"/>
              <path d="M 120,240 L 120,264" fill="none" stroke="black"/>
              <path d="M 120,336 L 120,352" fill="none" stroke="black"/>
              <path d="M 120,384 L 120,408" fill="none" stroke="black"/>
              <path d="M 232,128 L 232,176" fill="none" stroke="black"/>
              <path d="M 232,416 L 232,464" fill="none" stroke="black"/>
              <path d="M 264,128 L 264,176" fill="none" stroke="black"/>
              <path d="M 264,416 L 264,464" fill="none" stroke="black"/>
              <path d="M 376,80 L 376,120" fill="none" stroke="black"/>
              <path d="M 376,176 L 376,192" fill="none" stroke="black"/>
              <path d="M 376,240 L 376,264" fill="none" stroke="black"/>
              <path d="M 376,336 L 376,352" fill="none" stroke="black"/>
              <path d="M 376,384 L 376,408" fill="none" stroke="black"/>
              <path d="M 488,32 L 488,80" fill="none" stroke="black"/>
              <path d="M 488,128 L 488,176" fill="none" stroke="black"/>
              <path d="M 488,272 L 488,336" fill="none" stroke="black"/>
              <path d="M 488,416 L 488,464" fill="none" stroke="black"/>
              <path d="M 8,32 L 488,32" fill="none" stroke="black"/>
              <path d="M 8,80 L 488,80" fill="none" stroke="black"/>
              <path d="M 8,128 L 232,128" fill="none" stroke="black"/>
              <path d="M 264,128 L 488,128" fill="none" stroke="black"/>
              <path d="M 8,176 L 232,176" fill="none" stroke="black"/>
              <path d="M 264,176 L 488,176" fill="none" stroke="black"/>
              <path d="M 8,272 L 488,272" fill="none" stroke="black"/>
              <path d="M 8,336 L 488,336" fill="none" stroke="black"/>
              <path d="M 8,416 L 232,416" fill="none" stroke="black"/>
              <path d="M 264,416 L 488,416" fill="none" stroke="black"/>
              <path d="M 8,464 L 232,464" fill="none" stroke="black"/>
              <path d="M 264,464 L 488,464" fill="none" stroke="black"/>
              <polygon class="arrowhead" points="384,408 372,402.4 372,413.6" fill="black" transform="rotate(90,376,408)"/>
              <polygon class="arrowhead" points="384,264 372,258.4 372,269.6" fill="black" transform="rotate(90,376,264)"/>
              <polygon class="arrowhead" points="384,120 372,114.4 372,125.6" fill="black" transform="rotate(90,376,120)"/>
              <polygon class="arrowhead" points="128,408 116,402.4 116,413.6" fill="black" transform="rotate(90,120,408)"/>
              <polygon class="arrowhead" points="128,264 116,258.4 116,269.6" fill="black" transform="rotate(90,120,264)"/>
              <polygon class="arrowhead" points="128,120 116,114.4 116,125.6" fill="black" transform="rotate(90,120,120)"/>
              <g class="text">
                <text x="180" y="52">Platform</text>
                <text x="244" y="52">acting</text>
                <text x="284" y="52">as</text>
                <text x="324" y="52">Anchor</text>
                <text x="108" y="68">At</text>
                <text x="140" y="68">most</text>
                <text x="176" y="68">two</text>
                <text x="244" y="68">Endorsements</text>
                <text x="312" y="68">per</text>
                <text x="356" y="68">tenant</text>
                <text x="108" y="148">Tenant</text>
                <text x="144" y="148">A</text>
                <text x="364" y="148">Tenant</text>
                <text x="400" y="148">B</text>
                <text x="56" y="164">Two</text>
                <text x="112" y="164">10-credit</text>
                <text x="184" y="164">budgets</text>
                <text x="312" y="164">Two</text>
                <text x="368" y="164">10-credit</text>
                <text x="440" y="164">budgets</text>
                <text x="88" y="212">Abusive</text>
                <text x="152" y="212">request</text>
                <text x="344" y="212">Allowed</text>
                <text x="408" y="212">request</text>
                <text x="64" y="228">+</text>
                <text x="124" y="228">presentation</text>
                <text x="320" y="228">+</text>
                <text x="380" y="228">presentation</text>
                <text x="188" y="292">Origin</text>
                <text x="232" y="292">and</text>
                <text x="288" y="292">Moderator</text>
                <text x="108" y="308">Both</text>
                <text x="164" y="308">requests</text>
                <text x="224" y="308">carry</text>
                <text x="264" y="308">the</text>
                <text x="316" y="308">platform</text>
                <text x="392" y="308">signature</text>
                <text x="108" y="324">Each</text>
                <text x="180" y="324">presentation</text>
                <text x="260" y="324">spends</text>
                <text x="296" y="324">a</text>
                <text x="340" y="324">separate</text>
                <text x="404" y="324">budget</text>
                <text x="116" y="372">NoUpdate</text>
                <text x="348" y="372">Update</text>
                <text x="400" y="372">(t=0)</text>
                <text x="72" y="436">Presented</text>
                <text x="136" y="436">chain</text>
                <text x="180" y="436">ends</text>
                <text x="336" y="436">Chain</text>
                <text x="400" y="436">continues</text>
                <text x="72" y="452">Remaining</text>
                <text x="144" y="452">credits</text>
                <text x="196" y="452">lost</text>
                <text x="328" y="452">One</text>
                <text x="372" y="452">credit</text>
                <text x="424" y="452">spent</text>
              </g>
            </svg>
          </artwork>
          <artwork type="ascii-art"><![CDATA[
+-----------------------------------------------------------+
|                 Platform acting as Anchor                 |
|           At most two Endorsements per tenant             |
+-------------+-------------------------------+-------------+
              |                               |
              v                               v
+---------------------------+   +---------------------------+
|         Tenant A          |   |         Tenant B          |
|    Two 10-credit budgets  |   |    Two 10-credit budgets  |
+-------------+-------------+   +-------------+-------------+
              |                               |
       Abusive request                 Allowed request
       + presentation                  + presentation
              |                               |
              v                               v
+-----------------------------------------------------------+
|                   Origin and Moderator                    |
|          Both requests carry the platform signature       |
|          Each presentation spends a separate budget       |
+-------------+-------------------------------+-------------+
              |                               |
          NoUpdate                      Update (t=0)
              |                               |
              v                               v
+---------------------------+   +---------------------------+
|   Presented chain ends    |   |      Chain continues      |
|   Remaining credits lost  |   |      One credit spent     |
+---------------------------+   +---------------------------+
]]></artwork>
        </artset>
      </figure>
      <t>This ban applies to the presented chain. It does not revoke A's other
Credential or prevent A from obtaining a later allocation. A replacement must
count against the tenant's allowance. Exhaustion also prevents further spending
from a budget, but an update returning zero credits does not end a chain that
still has a sufficient balance. Neither a task boundary nor a tenant boundary
is established by the presentation itself.</t>
      <t>The platform's aggregate limit and the origin's available capacity still apply
to both tenants. Public-baseline requests are outside the illustrated budget.
Missing budget evidence does not itself establish abuse.</t>
      <t>For the origin to ask the platform to investigate a tenant or change its future
allocations, they need a feedback arrangement. One possible extension would
require a fresh platform Endorsement before accepting an origin-issued
Privacy Pass token, using a reverse-flow exchange. Feedback could also occur out of band. These
are possible deployment arrangements, not exchanges specified here. Their
binding to the affected allocation, enforcement, and disclosure to each party
remain to be defined. They need not require a persistent tenant identifier at
the origin.</t>
    </section>
    <section anchor="principles">
      <name>Scope and principles</name>
      <t>The goal of PACE is to bound consumption on a per-platform and per-client basis
through independently allocated and managed budgets. Exhausting one budget
should not exhaust another allocation, provided Anchor issuance and Moderator
spending policies keep them separate.</t>
      <t>The purpose is to allow origins to offer better public baseline services for its
users. Without per-client enforcement, origins are left with overly restrictive
baselines (e.g., block everything) because honest clients cannot prove their capacity.
In other words, the deployment goal is a baseline level of access without a credential,
signature, or some other mechanism that demonstrates capacity. Absence of an active
signal does not itself establish abuse <xref target="CIVIL"/>.</t>
      <t>Permissioned (non-public) resources may require evidence of a right before
serving any content. A capability can establish that right without identifying
its holder. Where a resource's policy depends on a particular principal and
its permissions, authenticating the principal with a stable identifier is
appropriate. Capacity evidence alone establishes neither identity nor a
right to perform a permissioned operation.</t>
      <t>Clients choose which evidence to offer based on their needs and the needs of
the origin. Origins should ask only for evidence needed by the requested operation.
A verified identity does not establish that a request is harmless, and a valid
budget does not grant every permission. Moreover, client evidence should reveal
only what the origin needs to apply its policy. Network metadata and other
request fields can still identify or correlate clients.</t>
    </section>
    <section anchor="system">
      <name>Attribution, permission, and capacity</name>
      <t>PACE is centered around moderation through attribution and authorization.
Specifically, an origin decides whether to serve an operation and how much
capacity to allocate to it using evidence presented from the client. This evidence
can establish attribution (who is responsible for this traffic?), permission
(is the party responsible for this traffic allowed to access the resource?),
or capacity (does the party responsible for this traffic have budget for
accessing the resource?). A request may need more than one form of evidence.</t>
      <section anchor="attribution">
        <name>Attribution</name>
        <t>Attribution addresses the question of who is responsible for traffic,
and there are many layers at which this attribution can occur (see <xref target="parties"/>).
For example, a verified WBA(P) signature attributes its covered components to
a key published by P <xref target="WBA-HTTPSIG"/>. The origin can apply an aggregate limit to P
without receiving an identifier for each Client behind it. The signature does
not establish permission unless origin policy grants it to that signer.
A workload can authenticate using a variety of schemes, e.g., with a WIMSE
Proof Tokens, HTTP message signatures, or even transport-layer authentication
(see <xref target="WIMSE-AIMS"/>). Users can authentiate using <xref target="OIDC"/> or related techniques.</t>
      </section>
      <section anchor="permission">
        <name>Permission</name>
        <t>Permission addresses the question of whether the attributed party can access
the resource in question. As such, permission depends on the resource's policy.
For example, an OAuth access token can convey an authorization grant <xref target="OAUTH"/>.
The origin can check the token as its authorization server requires, including
its audience, validity, permissions, and any proof of possession. Identity
obtained through <xref target="OIDC"/> or workload authentication can also be an input
to access policy.</t>
        <t>A workload can obtain a token using <xref target="WAG"/>, and an existing federation can
support delegated access through <xref target="OAUTH-XAA"/>. These grants are exchanged at
the authorization server, which decides what access token to issue. They do
not define a common token claim layout. <xref target="AAUTH"/> defines a separate signed
authorization protocol with its own challenges and credentials.</t>
        <t>Notably, authentication alone does not determine the
permitted operations. In fact, an origin can accept a capability without
identifying its holder, including a Privacy Pass token when the issuer's
attestation satisfies that policy <xref target="PRIVACYPASS"/>.</t>
      </section>
      <section anchor="capacity">
        <name>Capacity</name>
        <t>Capacity addresses the question of whether the attributed party has available
budget for accessing a resource. Budget can be allocated and managed using
granular identities such as a session ID, user ID, or platform ID. Budget
can also be allocated and managed using MOLE, which permits clients to hold
a hidden balance <tt>c</tt> and then update it through interaction with the origin
and its moderator. (This is referred to as presenting a MOLE credential.)
The origin learns the public amounts and that the spend
was possible, without learning the balance or a stable client identifier
<xref target="MOLE-PROTOCOLS"/>.</t>
        <t>For example, a policy can issue <tt>B</tt> credits, charge one per accepted
presentation, and return zero with no top-ups. Each issuance then allows at most
<tt>B</tt> fresh accepted operation attempts. Further issuances increase that
allocation. Bounding consumption over time also requires policy for replenishment
and expiry. Refunding every spend in full leaves successful consumption
unbounded.</t>
      </section>
      <section anchor="composition">
        <name>Composition</name>
        <t>Attribution, authorization, and capacity evidence can compose in requests.
For instance, an origin might require that requests carry a platform signature
(WBA(P)) and a MOLE presentation corresponding to the user (MOLE(U)) on each
request. An origin might also require both a workload authorization token granting
permission to use a resource and a MOLE presentation for the user behind that workload.
Origins decide what composition is essential based on their needs.</t>
        <t>Carriage of a MOLE presentation alongside another authorization scheme or
additional budget presentations, and signaling cumulative requirements, remain
undefined here (see <eref target="https://github.com/thibmeu/draft-capacity-management/issues/9">issue #9</eref>).</t>
      </section>
    </section>
    <section anchor="processing-steps">
      <name>Moderating consumption</name>
      <t>Given mechanisms for establishing attribution, permission, and capacity,
origins can moderate composition accordingly.
The origin's moderator configures its accepted Anchor Set, credential contexts,
allocation and spending policy, and any permission requirements. A client may
send an initial request without evidence. The origin then verifies what is present
and serves the level of access its policy permits. For example, a public response
can offer more capacity through a MOLE challenge. A permissioned resource can
refuse the operation and challenge for the missing authorization.</t>
      <t>A MOLE redemption challenge lets a client obtain a credential. A presentation
challenge specifies its spending terms. Clients use the exchanges and
verification rules of <xref target="MOLE-PROTOCOLS"/> and <xref target="MOLE-HTTP"/>, including replay
checks and request binding. A failed proof supplies no evidence of budget. A
refund replaces the spent Credential with an updated one. Returning zero credits
leaves <tt>c - s</tt> credits. Recording <tt>NoUpdate</tt> ends the chain, as illustrated in
<xref target="scenarios"/>. The Moderator records this decision before attempting the
operation, using the evidence available at that point.</t>
      <t>After a lost response, the Client can recover the recorded credential result
within the retention period. Recovery does not repeat the protected operation
or establish whether it completed. Application recovery remains separate.
One credential permits one presentation in flight at a time.</t>
      <section anchor="flow1">
        <name>Public resource</name>
        <t>The origin offers a budget above its public baseline. Client C already holds
an Endorsement from an accepted Anchor. Intermediary P adds its own signature.
The Moderator is shown at the origin.</t>
        <figure anchor="sequence">
          <name>Public access with carrier attribution and an anonymous budget</name>
          <artset>
            <artwork type="svg"><svg xmlns="http://www.w3.org/2000/svg" version="1.1" height="592" width="576" viewBox="0 0 576 592" class="diagram" text-anchor="middle" font-family="monospace" font-size="13px" stroke-linecap="round">
                <path d="M 8,32 L 8,80" fill="none" stroke="black"/>
                <path d="M 64,80 L 64,544" fill="none" stroke="black"/>
                <path d="M 120,32 L 120,80" fill="none" stroke="black"/>
                <path d="M 144,32 L 144,80" fill="none" stroke="black"/>
                <path d="M 200,80 L 200,544" fill="none" stroke="black"/>
                <path d="M 256,32 L 256,80" fill="none" stroke="black"/>
                <path d="M 288,32 L 288,80" fill="none" stroke="black"/>
                <path d="M 352,80 L 352,224" fill="none" stroke="black"/>
                <path d="M 352,272 L 352,288" fill="none" stroke="black"/>
                <path d="M 352,320 L 352,336" fill="none" stroke="black"/>
                <path d="M 352,368 L 352,480" fill="none" stroke="black"/>
                <path d="M 352,528 L 352,544" fill="none" stroke="black"/>
                <path d="M 416,32 L 416,80" fill="none" stroke="black"/>
                <path d="M 440,32 L 440,80" fill="none" stroke="black"/>
                <path d="M 504,80 L 504,544" fill="none" stroke="black"/>
                <path d="M 568,32 L 568,80" fill="none" stroke="black"/>
                <path d="M 8,32 L 120,32" fill="none" stroke="black"/>
                <path d="M 144,32 L 256,32" fill="none" stroke="black"/>
                <path d="M 288,32 L 416,32" fill="none" stroke="black"/>
                <path d="M 440,32 L 568,32" fill="none" stroke="black"/>
                <path d="M 8,80 L 120,80" fill="none" stroke="black"/>
                <path d="M 144,80 L 256,80" fill="none" stroke="black"/>
                <path d="M 288,80 L 416,80" fill="none" stroke="black"/>
                <path d="M 440,80 L 568,80" fill="none" stroke="black"/>
                <path d="M 64,112 L 192,112" fill="none" stroke="black"/>
                <path d="M 200,160 L 344,160" fill="none" stroke="black"/>
                <path d="M 352,208 L 496,208" fill="none" stroke="black"/>
                <path d="M 208,272 L 504,272" fill="none" stroke="black"/>
                <path d="M 200,320 L 496,320" fill="none" stroke="black"/>
                <path d="M 208,368 L 504,368" fill="none" stroke="black"/>
                <path d="M 200,416 L 344,416" fill="none" stroke="black"/>
                <path d="M 352,464 L 496,464" fill="none" stroke="black"/>
                <path d="M 208,528 L 504,528" fill="none" stroke="black"/>
                <polygon class="arrowhead" points="504,464 492,458.4 492,469.6" fill="black" transform="rotate(0,496,464)"/>
                <polygon class="arrowhead" points="504,320 492,314.4 492,325.6" fill="black" transform="rotate(0,496,320)"/>
                <polygon class="arrowhead" points="504,208 492,202.4 492,213.6" fill="black" transform="rotate(0,496,208)"/>
                <polygon class="arrowhead" points="352,416 340,410.4 340,421.6" fill="black" transform="rotate(0,344,416)"/>
                <polygon class="arrowhead" points="352,160 340,154.4 340,165.6" fill="black" transform="rotate(0,344,160)"/>
                <polygon class="arrowhead" points="216,528 204,522.4 204,533.6" fill="black" transform="rotate(180,208,528)"/>
                <polygon class="arrowhead" points="216,368 204,362.4 204,373.6" fill="black" transform="rotate(180,208,368)"/>
                <polygon class="arrowhead" points="216,272 204,266.4 204,277.6" fill="black" transform="rotate(180,208,272)"/>
                <polygon class="arrowhead" points="200,112 188,106.4 188,117.6" fill="black" transform="rotate(0,192,112)"/>
                <g class="text">
                  <text x="60" y="52">Anchor</text>
                  <text x="188" y="52">Client</text>
                  <text x="224" y="52">C</text>
                  <text x="348" y="52">Intermediary</text>
                  <text x="500" y="52">Origin</text>
                  <text x="352" y="68">P</text>
                  <text x="504" y="68">Moderator</text>
                  <text x="128" y="100">Endorsement</text>
                  <text x="272" y="148">request</text>
                  <text x="400" y="196">+</text>
                  <text x="436" y="196">WBA(P)</text>
                  <text x="284" y="244">baseline</text>
                  <text x="328" y="244">+</text>
                  <text x="380" y="244">redemption</text>
                  <text x="440" y="244">and</text>
                  <text x="308" y="260">presentation</text>
                  <text x="404" y="260">challenges</text>
                  <text x="316" y="308">Redeem</text>
                  <text x="352" y="308">&amp;</text>
                  <text x="384" y="308">Issue</text>
                  <text x="292" y="356">Credential</text>
                  <text x="356" y="356">with</text>
                  <text x="408" y="356">balance</text>
                  <text x="448" y="356">c</text>
                  <text x="232" y="404">request</text>
                  <text x="272" y="404">+</text>
                  <text x="312" y="404">MOLE(C)</text>
                  <text x="400" y="452">+</text>
                  <text x="436" y="452">WBA(P)</text>
                  <text x="252" y="500">record</text>
                  <text x="308" y="500">spend,</text>
                  <text x="368" y="500">attempt</text>
                  <text x="440" y="500">operation</text>
                  <text x="300" y="516">response</text>
                  <text x="356" y="516">with</text>
                  <text x="404" y="516">Update</text>
                  <text x="180" y="580">Client</text>
                  <text x="244" y="580">balance:</text>
                  <text x="288" y="580">c</text>
                  <text x="304" y="580">-</text>
                  <text x="320" y="580">s</text>
                  <text x="336" y="580">+</text>
                  <text x="352" y="580">t</text>
                </g>
              </svg>
            </artwork>
            <artwork type="ascii-art"><![CDATA[
+-------------+  +-------------+   +---------------+  +---------------+
|   Anchor    |  |  Client C   |   | Intermediary  |  |    Origin     |
|             |  |             |   |       P       |  |   Moderator   |
+------+------+  +------+------+   +-------+-------+  +-------+-------+
       |  Endorsement   |                  |                  |
       +--------------->+                  |                  |
       |                |                  |                  |
       |                |     request      |                  |
       |                +----------------->+                  |
       |                |                  |                  |
       |                |                  |     + WBA(P)     |
       |                |                  +----------------->+
       |                |                  |                  |
       |                |      baseline + redemption and      |
       |                |       presentation challenges       |
       |                +<-----------------+------------------+
       |                |                  |                  |
       |                |           Redeem & Issue            |
       |                +------------------+----------------->+
       |                |                  |                  |
       |                |      Credential with balance c      |
       |                +<-----------------+------------------+
       |                |                  |                  |
       |                |request + MOLE(C) |                  |
       |                +----------------->+                  |
       |                |                  |                  |
       |                |                  |     + WBA(P)     |
       |                |                  +----------------->+
       |                |                  |                  |
       |                |   record spend, attempt operation   |
       |                |        response with Update         |
       |                +<-----------------+------------------+
       |                |                  |                  |

                   Client balance: c - s + t
]]></artwork>
          </artset>
        </figure>
        <ol spacing="normal" type="1"><li>
            <t>The Client sends a request. The intermediary adds WBA(P), which the
origin verifies before applying its aggregate limit.</t>
          </li>
          <li>
            <t>The origin serves the public baseline and offers redemption and presentation
challenges on the 200 response. A Client with a suitable Credential skips
redemption. Otherwise, it fetches the Moderator configuration and redeems
its Endorsement at the Redeem &amp; Issue endpoint.</t>
          </li>
          <li>
            <t>The Client presents proof that its balance covers <tt>s</tt> and requests a top-up
of at most <tt>a</tt>. The Moderator verifies and records the spend result atomically
before attempting the operation.</t>
          </li>
          <li>
            <t>The response returns an Update with <tt>0 &lt;= t &lt;= s + a</tt>, continuing the chain
with balance <tt>c - s + t</tt>. Each subsequent spend answers an applicable
challenge and uses a fresh nullifier, without a credential link to the
previous presentation.</t>
          </li>
        </ol>
        <t>A direct Client uses the same exchange without WBA(P). A Client unable or
unwilling to present continues at the public baseline, subject to origin policy.
A failed presentation does not authorize the budgeted level of access.</t>
      </section>
    </section>
    <section anchor="examples">
      <name>Deploying PACE</name>
      <t>A deployment of PACE admits two kinds of capacity control: (1) through
an intermediary which identify itself for an aggregate, platform-specific limit,
and (2) through smaller, anonymous budgets that are allocated to indivdual
clients (tenants, workloads, agents, or users) based on the chosen anchor
set. <xref target="fig-intermediary"/> shows one example where an Anchor endorses Client C
while intermediary platform P carries its requests. The same provider can operate
both the role of anchor and intermediary, but need not.</t>
      <figure anchor="fig-intermediary">
        <name>Separate endorsement and forwarding relationships</name>
        <artset>
          <artwork type="svg"><svg xmlns="http://www.w3.org/2000/svg" version="1.1" height="192" width="512" viewBox="0 0 512 192" class="diagram" text-anchor="middle" font-family="monospace" font-size="13px" stroke-linecap="round">
              <path d="M 8,32 L 8,64" fill="none" stroke="black"/>
              <path d="M 8,112 L 8,176" fill="none" stroke="black"/>
              <path d="M 48,64 L 48,104" fill="none" stroke="black"/>
              <path d="M 96,32 L 96,64" fill="none" stroke="black"/>
              <path d="M 96,112 L 96,176" fill="none" stroke="black"/>
              <path d="M 192,112 L 192,176" fill="none" stroke="black"/>
              <path d="M 312,112 L 312,176" fill="none" stroke="black"/>
              <path d="M 408,112 L 408,176" fill="none" stroke="black"/>
              <path d="M 504,112 L 504,176" fill="none" stroke="black"/>
              <path d="M 8,32 L 96,32" fill="none" stroke="black"/>
              <path d="M 8,64 L 96,64" fill="none" stroke="black"/>
              <path d="M 8,112 L 96,112" fill="none" stroke="black"/>
              <path d="M 192,112 L 312,112" fill="none" stroke="black"/>
              <path d="M 408,112 L 504,112" fill="none" stroke="black"/>
              <path d="M 96,144 L 184,144" fill="none" stroke="black"/>
              <path d="M 312,144 L 400,144" fill="none" stroke="black"/>
              <path d="M 8,176 L 96,176" fill="none" stroke="black"/>
              <path d="M 192,176 L 312,176" fill="none" stroke="black"/>
              <path d="M 408,176 L 504,176" fill="none" stroke="black"/>
              <polygon class="arrowhead" points="408,144 396,138.4 396,149.6" fill="black" transform="rotate(0,400,144)"/>
              <polygon class="arrowhead" points="192,144 180,138.4 180,149.6" fill="black" transform="rotate(0,184,144)"/>
              <polygon class="arrowhead" points="56,104 44,98.4 44,109.6" fill="black" transform="rotate(90,48,104)"/>
              <g class="text">
                <text x="52" y="52">Anchor</text>
                <text x="104" y="84">Endorsement</text>
                <text x="44" y="132">Client</text>
                <text x="80" y="132">C</text>
                <text x="144" y="132">MOLE(C)</text>
                <text x="252" y="132">Intermediary</text>
                <text x="360" y="132">MOLE(C)</text>
                <text x="460" y="132">Origin</text>
                <text x="248" y="148">P</text>
                <text x="456" y="148">Moderator</text>
                <text x="360" y="164">+WBA(P)</text>
              </g>
            </svg>
          </artwork>
          <artwork type="ascii-art"><![CDATA[
+----------+
|  Anchor  |
+----+-----+
     | Endorsement
     v
+----------+           +--------------+           +-----------+
| Client C |  MOLE(C)  | Intermediary |  MOLE(C)  |   Origin  |
|          +---------->|      P       +---------->| Moderator |
|          |           |              |  +WBA(P)  |           |
+----------+           +--------------+           +-----------+
]]></artwork>
        </artset>
      </figure>
      <t>The origin can attribute the forwarded request to P through WBA(P) and
charge C's separate budget through MOLE(C), without learning C's identity
from that presentation. The allocation relationship, rather than P's
signature, establishes which smaller unit bears the spend. The following
subsection applies this separation to product, project, and session
capacity limits.</t>
      <section anchor="capacity-scopes">
        <name>Capacity scopes</name>
        <t>Using the terms in <xref target="parties"/>, a platform P hosts tenants T with workloads W.
An account U may authorize several agents. <xref target="endorsement-examples"/> describes
allocations at these scopes. The noisy-neighbour
boundary depends on which of these shares a budget. A tenant budget separates
tenants but leaves workloads within a tenant sharing its allocation. A
workload budget separates workloads but still groups their Clients. Account
budgets can isolate those allocations further, while several Clients sharing
one account budget remain subject to the same limit.</t>
        <t>In <xref target="fig-stacked"/>, Anchor W endorses workload-scoped allocations and Anchor U
endorses account-scoped allocations. Client components obtain and spend the
corresponding Credentials. The request carries both budget proofs alongside
the platform's signature.
The Anchors need not operate the forwarding intermediary. These are intended
scopes, pending the multiple-presentation carriage in <xref target="composition"/>.</t>
        <figure anchor="fig-stacked">
          <name>Workload and account budget scopes in a proposed stacked deployment</name>
          <artset>
            <artwork type="svg"><svg xmlns="http://www.w3.org/2000/svg" version="1.1" height="416" width="520" viewBox="0 0 520 416" class="diagram" text-anchor="middle" font-family="monospace" font-size="13px" stroke-linecap="round">
                <path d="M 8,32 L 8,64" fill="none" stroke="black"/>
                <path d="M 8,112 L 8,272" fill="none" stroke="black"/>
                <path d="M 32,144 L 32,176" fill="none" stroke="black"/>
                <path d="M 96,64 L 96,104" fill="none" stroke="black"/>
                <path d="M 152,224 L 152,256" fill="none" stroke="black"/>
                <path d="M 184,32 L 184,64" fill="none" stroke="black"/>
                <path d="M 216,176 L 216,216" fill="none" stroke="black"/>
                <path d="M 248,32 L 248,64" fill="none" stroke="black"/>
                <path d="M 272,224 L 272,256" fill="none" stroke="black"/>
                <path d="M 336,64 L 336,104" fill="none" stroke="black"/>
                <path d="M 392,144 L 392,176" fill="none" stroke="black"/>
                <path d="M 400,352 L 400,400" fill="none" stroke="black"/>
                <path d="M 416,112 L 416,272" fill="none" stroke="black"/>
                <path d="M 424,32 L 424,64" fill="none" stroke="black"/>
                <path d="M 448,240 L 448,344" fill="none" stroke="black"/>
                <path d="M 496,352 L 496,400" fill="none" stroke="black"/>
                <path d="M 8,32 L 184,32" fill="none" stroke="black"/>
                <path d="M 248,32 L 424,32" fill="none" stroke="black"/>
                <path d="M 8,64 L 184,64" fill="none" stroke="black"/>
                <path d="M 248,64 L 424,64" fill="none" stroke="black"/>
                <path d="M 8,112 L 416,112" fill="none" stroke="black"/>
                <path d="M 32,144 L 392,144" fill="none" stroke="black"/>
                <path d="M 32,176 L 392,176" fill="none" stroke="black"/>
                <path d="M 152,224 L 272,224" fill="none" stroke="black"/>
                <path d="M 272,240 L 448,240" fill="none" stroke="black"/>
                <path d="M 152,256 L 272,256" fill="none" stroke="black"/>
                <path d="M 8,272 L 416,272" fill="none" stroke="black"/>
                <path d="M 400,352 L 496,352" fill="none" stroke="black"/>
                <path d="M 400,400 L 496,400" fill="none" stroke="black"/>
                <polygon class="arrowhead" points="456,344 444,338.4 444,349.6" fill="black" transform="rotate(90,448,344)"/>
                <polygon class="arrowhead" points="344,104 332,98.4 332,109.6" fill="black" transform="rotate(90,336,104)"/>
                <polygon class="arrowhead" points="224,216 212,210.4 212,221.6" fill="black" transform="rotate(90,216,216)"/>
                <polygon class="arrowhead" points="104,104 92,98.4 92,109.6" fill="black" transform="rotate(90,96,104)"/>
                <g class="text">
                  <text x="84" y="52">Anchor</text>
                  <text x="120" y="52">W</text>
                  <text x="324" y="52">Anchor</text>
                  <text x="360" y="52">U</text>
                  <text x="152" y="84">Endorsement</text>
                  <text x="392" y="84">Endorsement</text>
                  <text x="52" y="132">Platform</text>
                  <text x="96" y="132">P</text>
                  <text x="68" y="164">Tenant</text>
                  <text x="132" y="164">workload</text>
                  <text x="188" y="164">with</text>
                  <text x="240" y="164">Clients</text>
                  <text x="212" y="244">Intermediary</text>
                  <text x="488" y="292">MOLE(W)</text>
                  <text x="488" y="308">MOLE(U)</text>
                  <text x="488" y="324">+WBA(P)</text>
                  <text x="452" y="372">Origin</text>
                  <text x="448" y="388">Moderator</text>
                </g>
              </svg>
            </artwork>
            <artwork type="ascii-art"><![CDATA[
+---------------------+       +---------------------+
|      Anchor W       |       |      Anchor U       |
+----------+----------+       +----------+----------+
           | Endorsement                 | Endorsement
           v                             v
+--------------------------------------------------+
| Platform P                                       |
|  +--------------------------------------------+  |
|  | Tenant workload with Clients               |  |
|  +----------------------+---------------------+  |
|                         |                        |
|                         v                        |
|                 +--------------+                 |
|                 | Intermediary +-----------------+---+
|                 +--------------+                 |   |
+--------------------------------------------------+   |
                                                       | MOLE(W)
                                                       | MOLE(U)
                                                       | +WBA(P)
                                                       v
                                                 +-----------+
                                                 |   Origin  |
                                                 | Moderator |
                                                 +-----------+
]]></artwork>
          </artset>
        </figure>
      </section>
      <section anchor="composing-scopes">
        <name>Composing scopes</name>
        <t>A client can receive several independent budgets, for example one per
tenant, workload, agent, or user. Authorization determines what each client
may do, while issuance policy constrains their combined allocation. Adding
Clients must not yield unbounded replacement budgets. Aggregate platform
limits should still apply when individual budgets remain available.</t>
        <t>With MOLE, concurrent requests need separate credentials, since each request
effectively "locks" a credential while in use. Several Credentials provide
separate balances, so assigning more Credentials can increase both
concurrency and total consumption. Shared-budget splitting and simultaneous
checks of aggregate and individual anonymous budgets remain open.</t>
        <t>A CDN can operate a Moderator for several origins. All servers accepting the
same credential key and context must share atomic replay state. Independent
stores need disjoint keys or contexts. Sharing a credential context also
shares its budget scope and can enlarge the anonymity set, but activity at
one origin can then reduce the balance available at another.</t>
        <t>Anchors constrain new allocations, but the Moderator cannot identify the
issuing Anchor from a redemption. <xref target="MOLE-ARCH"/> leaves privacy-preserving
Anchor feedback unresolved. Removing an Anchor requires evidence from outside
that anonymous redemption, and changing sets to probe Clients can reduce
privacy. Endorsements from several accepted Anchors can yield several budgets.</t>
        <t>Deployment also needs bounded verification work and replay storage. Invalid
proofs can consume verification capacity, and replay markers must remain while
their Credentials are accepted. Aggregate controls remain necessary for
availability, even though they can reintroduce shared fate.</t>
      </section>
    </section>
    <section anchor="security">
      <name>Security considerations</name>
      <t>The security goals and limitations of MOLE are described in <xref target="MOLE-ARCH"/>,
<xref target="MOLE-PROTOCOLS"/>, and <xref target="EC-ACT"/>.</t>
      <section anchor="request-integrity">
        <name>Request integrity</name>
        <t>An intermediary that sees a presentation can race the Client to spend it.
The request-bound profile of <xref target="MOLE-PROTOCOLS"/> must cover every component
that affects authorization, including the body where relevant. Single-use
verification alone determines only which spend wins. Binding does not prevent
an intermediary from withholding or racing an identical request.</t>
        <t>A carrier signature should cover the MOLE presentation and the relevant
request components. It attributes that pairing to the carrier. Protection
against the carrier itself requires binding the Client's evidence to the
operation. This document does not define the intermediary signing profile.</t>
        <t>Sender-constrained access tokens limit reuse of stolen tokens, but DPoP does
not protect the request body <xref target="DPOP"/>. Mutual TLS authenticates the connection
at its TLS endpoint <xref target="MTLS"/>, which may be an intermediary. Application-layer
signatures can preserve integrity across TLS termination <xref target="WIMSE-AIMS"/>.
For signed HTTP content, the verifier checks <tt>Content-Digest</tt> against the
received content and verifies a signature covering it, together with the
other components needed to bind the operation <xref target="HTTPSIG"/> <xref target="DIGEST"/>.</t>
      </section>
      <section anchor="correlation-and-policy">
        <name>Correlation and policy</name>
        <t>Unlinkable Credentials do not hide addresses, cookies, content, or timing.
A forwarding provider that knows the Client can disclose its requests to the
origin. This deployment does not prevent that collusion.</t>
        <t>Credential keys, contexts, and public amounts partition the anonymity set.
They should follow coarse policies shared by many Clients. Per-Client values
or repeated balance probes can make presentations distinguishable. The
privacy requirements of <xref target="MOLE-PROTOCOLS"/> apply to allocation, spending,
and refund policy. A platform signature further narrows the observable set.</t>
        <t>A Moderator can end a credential chain without identifying its holder. Clients
therefore rely on it to apply the advertised policy. Withholding refunds can
consume scarce Endorsements, while recovery distinguishes a lost response from
a recorded termination only while the result remains available.</t>
      </section>
    </section>
    <section anchor="iana">
      <name>IANA considerations</name>
      <t>This document has no IANA actions.</t>
    </section>
  </middle>
  <back>
    <references anchor="sec-informative-references">
      <name>Informative References</name>
      <reference anchor="CIVIL">
        <front>
          <title>Principles for Web Traffic Management by Intermediaries</title>
          <author fullname="Thibault Meunier" initials="T." surname="Meunier">
            <organization>Cloudflare</organization>
          </author>
          <author fullname="Kevin Kelley" initials="K." surname="Kelley">
            <organization>Anthropic</organization>
          </author>
          <date day="29" month="October" year="2026"/>
          <abstract>
            <t>   This document describes principles for web intermediaries that
   balance security needs with open web accessibility.

About This Document

   This note is to be removed before publishing as an RFC.

   The latest revision of this draft can be found at
   https://thibmeu.github.io/draft-bcp-civil/draft-meunier-bcp-
   civil.html.  Status information for this document may be found at
   https://datatracker.ietf.org/doc/draft-meunier-bcp-civil/.

   Source for this draft and an issue tracker can be found at
   https://github.com/thibmeu/draft-bcp-civil.

            </t>
          </abstract>
        </front>
        <seriesInfo name="Internet-Draft" value="draft-meunier-bcp-civil-00"/>
      </reference>
      <reference anchor="AAUTH">
        <front>
          <title>AAuth Protocol</title>
          <author fullname="Dick Hardt" initials="D." surname="Hardt">
            <organization>Hellō</organization>
          </author>
          <date day="25" month="September" year="2026"/>
          <abstract>
            <t>   This document defines the AAuth authorization protocol for agent-to-
   resource authorization and identity claim retrieval.  The protocol
   supports five resource access modes — agent identity, resource-
   managed (two-party), person identity, PS authorization (three-party),
   and federated authorization (four-party) — with agent governance as
   an orthogonal layer.  It builds on the HTTP Signature Keys
   specification ([I-D.hardt-httpbis-signature-key]) for HTTP Message
   Signatures and key discovery.

            </t>
          </abstract>
        </front>
        <seriesInfo name="Internet-Draft" value="draft-hardt-oauth-aauth-protocol-11"/>
      </reference>
      <reference anchor="CONNECT-UDP">
        <front>
          <title>Proxying UDP in HTTP</title>
          <author fullname="D. Schinazi" initials="D." surname="Schinazi"/>
          <date month="August" year="2022"/>
          <abstract>
            <t>This document describes how to proxy UDP in HTTP, similar to how the HTTP CONNECT method allows proxying TCP in HTTP. More specifically, this document defines a protocol that allows an HTTP client to create a tunnel for UDP communications through an HTTP server that acts as a proxy.</t>
          </abstract>
        </front>
        <seriesInfo name="RFC" value="9298"/>
        <seriesInfo name="DOI" value="10.17487/RFC9298"/>
      </reference>
      <reference anchor="CONNECT-IP">
        <front>
          <title>Proxying IP in HTTP</title>
          <author fullname="T. Pauly" initials="T." role="editor" surname="Pauly"/>
          <author fullname="D. Schinazi" initials="D." surname="Schinazi"/>
          <author fullname="A. Chernyakhovsky" initials="A." surname="Chernyakhovsky"/>
          <author fullname="M. Kühlewind" initials="M." surname="Kühlewind"/>
          <author fullname="M. Westerlund" initials="M." surname="Westerlund"/>
          <date month="October" year="2023"/>
          <abstract>
            <t>This document describes how to proxy IP packets in HTTP. This protocol is similar to UDP proxying in HTTP but allows transmitting arbitrary IP packets. More specifically, this document defines a protocol that allows an HTTP client to create an IP tunnel through an HTTP server that acts as an IP proxy. This document updates RFC 9298.</t>
          </abstract>
        </front>
        <seriesInfo name="RFC" value="9484"/>
        <seriesInfo name="DOI" value="10.17487/RFC9484"/>
      </reference>
      <reference anchor="OHTTP">
        <front>
          <title>Oblivious HTTP</title>
          <author fullname="M. Thomson" initials="M." surname="Thomson"/>
          <author fullname="C. A. Wood" initials="C. A." surname="Wood"/>
          <date month="January" year="2024"/>
          <abstract>
            <t>This document describes Oblivious HTTP, a protocol for forwarding encrypted HTTP messages. Oblivious HTTP allows a client to make multiple requests to an origin server without that server being able to link those requests to the client or to identify the requests as having come from the same client, while placing only limited trust in the nodes used to forward the messages.</t>
          </abstract>
        </front>
        <seriesInfo name="RFC" value="9458"/>
        <seriesInfo name="DOI" value="10.17487/RFC9458"/>
      </reference>
      <reference anchor="PRIVACY-PARTITIONING">
        <front>
          <title>Partitioning as an Architecture for Privacy</title>
          <author fullname="M. Kühlewind" initials="M." surname="Kühlewind"/>
          <author fullname="T. Pauly" initials="T." surname="Pauly"/>
          <author fullname="C. A. Wood" initials="C. A." surname="Wood"/>
          <date month="July" year="2024"/>
          <abstract>
            <t>This document describes the principle of privacy partitioning, which selectively spreads data and communication across multiple parties as a means to improve privacy by separating user identity from user data. This document describes emerging patterns in protocols to partition what data and metadata is revealed through protocol interactions, provides common terminology, and discusses how to analyze such models.</t>
          </abstract>
        </front>
        <seriesInfo name="RFC" value="9614"/>
        <seriesInfo name="DOI" value="10.17487/RFC9614"/>
      </reference>
      <reference anchor="HTTP">
        <front>
          <title>HTTP Semantics</title>
          <author fullname="R. Fielding" initials="R." role="editor" surname="Fielding"/>
          <author fullname="M. Nottingham" initials="M." role="editor" surname="Nottingham"/>
          <author fullname="J. Reschke" initials="J." role="editor" surname="Reschke"/>
          <date month="June" year="2022"/>
          <abstract>
            <t>The Hypertext Transfer Protocol (HTTP) is a stateless application-level protocol for distributed, collaborative, hypertext information systems. This document describes the overall architecture of HTTP, establishes common terminology, and defines aspects of the protocol that are shared by all versions. In this definition are core protocol elements, extensibility mechanisms, and the "http" and "https" Uniform Resource Identifier (URI) schemes.</t>
            <t>This document updates RFC 3864 and obsoletes RFCs 2818, 7231, 7232, 7233, 7235, 7538, 7615, 7694, and portions of 7230.</t>
          </abstract>
        </front>
        <seriesInfo name="STD" value="97"/>
        <seriesInfo name="RFC" value="9110"/>
        <seriesInfo name="DOI" value="10.17487/RFC9110"/>
      </reference>
      <reference anchor="HTTPSIG">
        <front>
          <title>HTTP Message Signatures</title>
          <author fullname="A. Backman" initials="A." role="editor" surname="Backman"/>
          <author fullname="J. Richer" initials="J." role="editor" surname="Richer"/>
          <author fullname="M. Sporny" initials="M." surname="Sporny"/>
          <date month="February" year="2024"/>
          <abstract>
            <t>This document describes a mechanism for creating, encoding, and verifying digital signatures or message authentication codes over components of an HTTP message. This mechanism supports use cases where the full HTTP message may not be known to the signer and where the message may be transformed (e.g., by intermediaries) before reaching the verifier. This document also describes a means for requesting that a signature be applied to a subsequent HTTP message in an ongoing HTTP exchange.</t>
          </abstract>
        </front>
        <seriesInfo name="RFC" value="9421"/>
        <seriesInfo name="DOI" value="10.17487/RFC9421"/>
      </reference>
      <reference anchor="OAUTH">
        <front>
          <title>The OAuth 2.0 Authorization Framework</title>
          <author fullname="D. Hardt" initials="D." role="editor" surname="Hardt"/>
          <date month="October" year="2012"/>
          <abstract>
            <t>The OAuth 2.0 authorization framework enables a third-party application to obtain limited access to an HTTP service, either on behalf of a resource owner by orchestrating an approval interaction between the resource owner and the HTTP service, or by allowing the third-party application to obtain access on its own behalf. This specification replaces and obsoletes the OAuth 1.0 protocol described in RFC 5849. [STANDARDS-TRACK]</t>
          </abstract>
        </front>
        <seriesInfo name="RFC" value="6749"/>
        <seriesInfo name="DOI" value="10.17487/RFC6749"/>
      </reference>
      <reference anchor="BEARER">
        <front>
          <title>The OAuth 2.0 Authorization Framework: Bearer Token Usage</title>
          <author fullname="M. Jones" initials="M." surname="Jones"/>
          <author fullname="D. Hardt" initials="D." surname="Hardt"/>
          <date month="October" year="2012"/>
          <abstract>
            <t>This specification describes how to use bearer tokens in HTTP requests to access OAuth 2.0 protected resources. Any party in possession of a bearer token (a "bearer") can use it to get access to the associated resources (without demonstrating possession of a cryptographic key). To prevent misuse, bearer tokens need to be protected from disclosure in storage and in transport. [STANDARDS-TRACK]</t>
          </abstract>
        </front>
        <seriesInfo name="RFC" value="6750"/>
        <seriesInfo name="DOI" value="10.17487/RFC6750"/>
      </reference>
      <reference anchor="DPOP">
        <front>
          <title>OAuth 2.0 Demonstrating Proof of Possession (DPoP)</title>
          <author fullname="D. Fett" initials="D." surname="Fett"/>
          <author fullname="B. Campbell" initials="B." surname="Campbell"/>
          <author fullname="J. Bradley" initials="J." surname="Bradley"/>
          <author fullname="T. Lodderstedt" initials="T." surname="Lodderstedt"/>
          <author fullname="M. Jones" initials="M." surname="Jones"/>
          <author fullname="D. Waite" initials="D." surname="Waite"/>
          <date month="September" year="2023"/>
          <abstract>
            <t>This document describes a mechanism for sender-constraining OAuth 2.0 tokens via a proof-of-possession mechanism on the application level. This mechanism allows for the detection of replay attacks with access and refresh tokens.</t>
          </abstract>
        </front>
        <seriesInfo name="RFC" value="9449"/>
        <seriesInfo name="DOI" value="10.17487/RFC9449"/>
      </reference>
      <reference anchor="MTLS">
        <front>
          <title>OAuth 2.0 Mutual-TLS Client Authentication and Certificate-Bound Access Tokens</title>
          <author fullname="B. Campbell" initials="B." surname="Campbell"/>
          <author fullname="J. Bradley" initials="J." surname="Bradley"/>
          <author fullname="N. Sakimura" initials="N." surname="Sakimura"/>
          <author fullname="T. Lodderstedt" initials="T." surname="Lodderstedt"/>
          <date month="February" year="2020"/>
          <abstract>
            <t>This document describes OAuth client authentication and certificate-bound access and refresh tokens using mutual Transport Layer Security (TLS) authentication with X.509 certificates. OAuth clients are provided a mechanism for authentication to the authorization server using mutual TLS, based on either self-signed certificates or public key infrastructure (PKI). OAuth authorization servers are provided a mechanism for binding access tokens to a client's mutual-TLS certificate, and OAuth protected resources are provided a method for ensuring that such an access token presented to it was issued to the client presenting the token.</t>
          </abstract>
        </front>
        <seriesInfo name="RFC" value="8705"/>
        <seriesInfo name="DOI" value="10.17487/RFC8705"/>
      </reference>
      <reference anchor="DIGEST">
        <front>
          <title>Digest Fields</title>
          <author fullname="R. Polli" initials="R." surname="Polli"/>
          <author fullname="L. Pardue" initials="L." surname="Pardue"/>
          <date month="February" year="2024"/>
          <abstract>
            <t>This document defines HTTP fields that support integrity digests. The Content-Digest field can be used for the integrity of HTTP message content. The Repr-Digest field can be used for the integrity of HTTP representations. Want-Content-Digest and Want-Repr-Digest can be used to indicate a sender's interest and preferences for receiving the respective Integrity fields.</t>
            <t>This document obsoletes RFC 3230 and the Digest and Want-Digest HTTP fields.</t>
          </abstract>
        </front>
        <seriesInfo name="RFC" value="9530"/>
        <seriesInfo name="DOI" value="10.17487/RFC9530"/>
      </reference>
      <reference anchor="EC-ACT">
        <front>
          <title>EC-ACT: Anonymous Credit Tokens</title>
          <author fullname="Samuel Schlesinger" initials="S." surname="Schlesinger">
            <organization>Google LLC</organization>
          </author>
          <author fullname="Jonathan Katz" initials="J." surname="Katz">
            <organization>Google LLC</organization>
          </author>
          <author fullname="Armando Faz-Hernandez" initials="A. F." surname="Faz-Hernandez">
            <organization>Cloudflare, Inc.</organization>
          </author>
          <date day="2" month="November" year="2026"/>
          <abstract>
            <t>   This document specifies EC-ACT, a keyed-verification anonymous
   credential scheme with a hidden credit balance.  A Client spends a
   public amount and receives a fresh Credential for the remaining
   balance and a return amount chosen by the Moderator.  The Moderator
   verifies each spend without learning the balance or linking it to
   issuance or to other spends.

   EC-ACT uses a pairing-free BBS-style signature and proofs over linear
   relations.  This document defines its cryptographic operations,
   encodings, and security requirements.  The MOLE protocols use EC-ACT
   as a Credential scheme.

            </t>
          </abstract>
        </front>
        <seriesInfo name="Internet-Draft" value="draft-schlesinger-mole-ec-act-00"/>
      </reference>
      <reference anchor="PRIVACYPASS">
        <front>
          <title>The Privacy Pass Architecture</title>
          <author fullname="A. Davidson" initials="A." surname="Davidson"/>
          <author fullname="J. Iyengar" initials="J." surname="Iyengar"/>
          <author fullname="C. A. Wood" initials="C. A." surname="Wood"/>
          <date month="June" year="2024"/>
          <abstract>
            <t>This document specifies the Privacy Pass architecture and requirements for its constituent protocols used for authorization based on privacy-preserving authentication mechanisms. It describes the conceptual model of Privacy Pass and its protocols, its security and privacy goals, practical deployment models, and recommendations for each deployment model, to help ensure that the desired security and privacy goals are fulfilled.</t>
          </abstract>
        </front>
        <seriesInfo name="RFC" value="9576"/>
        <seriesInfo name="DOI" value="10.17487/RFC9576"/>
      </reference>
      <reference anchor="PRIVATETOKEN">
        <front>
          <title>The Privacy Pass HTTP Authentication Scheme</title>
          <author fullname="T. Pauly" initials="T." surname="Pauly"/>
          <author fullname="S. Valdez" initials="S." surname="Valdez"/>
          <author fullname="C. A. Wood" initials="C. A." surname="Wood"/>
          <date month="June" year="2024"/>
          <abstract>
            <t>This document defines an HTTP authentication scheme for Privacy Pass, a privacy-preserving authentication mechanism used for authorization. The authentication scheme specified in this document can be used by Clients to redeem Privacy Pass tokens with an Origin. It can also be used by Origins to challenge Clients to present Privacy Pass tokens.</t>
          </abstract>
        </front>
        <seriesInfo name="RFC" value="9577"/>
        <seriesInfo name="DOI" value="10.17487/RFC9577"/>
      </reference>
      <reference anchor="OIDC" target="https://openid.net/specs/openid-connect-core-1_0.html">
        <front>
          <title>OpenID Connect Core 1.0</title>
          <author>
            <organization/>
          </author>
          <date>n.d.</date>
        </front>
      </reference>
      <reference anchor="INTERNET-FOR-END-USER">
        <front>
          <title>The Internet is for End Users</title>
          <author fullname="M. Nottingham" initials="M." surname="Nottingham"/>
          <date month="August" year="2020"/>
          <abstract>
            <t>This document explains why the IAB believes that, when there is a conflict between the interests of end users of the Internet and other parties, IETF decisions should favor end users. It also explores how the IETF can more effectively achieve this.</t>
          </abstract>
        </front>
        <seriesInfo name="RFC" value="8890"/>
        <seriesInfo name="DOI" value="10.17487/RFC8890"/>
      </reference>
      <reference anchor="IPv6">
        <front>
          <title>Internet Protocol, Version 6 (IPv6) Specification</title>
          <author fullname="S. Deering" initials="S." surname="Deering"/>
          <author fullname="R. Hinden" initials="R." surname="Hinden"/>
          <date month="July" year="2017"/>
          <abstract>
            <t>This document specifies version 6 of the Internet Protocol (IPv6). It obsoletes RFC 2460.</t>
          </abstract>
        </front>
        <seriesInfo name="STD" value="86"/>
        <seriesInfo name="RFC" value="8200"/>
        <seriesInfo name="DOI" value="10.17487/RFC8200"/>
      </reference>
      <reference anchor="MOLE-ARCH">
        <front>
          <title>Moderation of unLinkable Endorsements (MOLE) Architecture</title>
          <author fullname="Samuel Schlesinger" initials="S." surname="Schlesinger">
            <organization>Google LLC</organization>
          </author>
          <author fullname="Dennis Jackson" initials="D." surname="Jackson">
            <organization>Mozilla</organization>
          </author>
          <author fullname="Thibault Meunier" initials="T." surname="Meunier">
            <organization>Cloudflare</organization>
          </author>
          <date day="2" month="November" year="2026"/>
          <abstract>
            <t>   Moderation of unLinkable Endorsements (MOLE) is an architecture that
   lets a party performing access control (a Moderator) bootstrap trust
   in a client from a third party (an Anchor) that already has a trust
   relationship with that client, and then adjust that trust over time
   in response to the client's behaviour, for example by dynamically
   rate-limiting access.

   MOLE targets open deployments, in which independent parties may be
   responsible for access control and for vouching for clients, whilst
   maintaining strong privacy protections for clients.  These
   protections are designed to hold even if participants in the
   ecosystem collude or otherwise misbehave.

   This document specifies the roles, the information flows between
   them, the privacy and security requirements, and deployment
   considerations.

            </t>
          </abstract>
        </front>
        <seriesInfo name="Internet-Draft" value="draft-jms-mole-architecture-01"/>
      </reference>
      <reference anchor="MOLE-PROTOCOLS">
        <front>
          <title>MOLE Protocols</title>
          <author fullname="Samuel Schlesinger" initials="S." surname="Schlesinger">
            <organization>Google LLC</organization>
          </author>
          <author fullname="Dennis Jackson" initials="D." surname="Jackson">
            <organization>Mozilla</organization>
          </author>
          <author fullname="Thibault Meunier" initials="T." surname="Meunier">
            <organization>Cloudflare</organization>
          </author>
          <date day="2" month="November" year="2026"/>
          <abstract>
            <t>   This document defines protocols that instantiate the MOLE
   architecture: two endorsement protocols, by which a Client proves to
   a Moderator that it holds an Endorsement from a trusted Anchor
   without revealing which one, and two credential protocols, by which a
   Moderator issues, verifies, and updates per-Client state without
   being able to link presentations.  It also establishes the registries
   that identify these protocols.

            </t>
          </abstract>
        </front>
        <seriesInfo name="Internet-Draft" value="draft-jms-mole-protocols-01"/>
      </reference>
      <reference anchor="MOLE-HTTP">
        <front>
          <title>MOLE HTTP Transport</title>
          <author fullname="Samuel Schlesinger" initials="S." surname="Schlesinger">
            <organization>Google LLC</organization>
          </author>
          <author fullname="Dennis Jackson" initials="D." surname="Jackson">
            <organization>Mozilla</organization>
          </author>
          <author fullname="Thibault Meunier" initials="T." surname="Meunier">
            <organization>Cloudflare</organization>
          </author>
          <date day="2" month="November" year="2026"/>
          <abstract>
            <t>   MOLE targets browser deployments, so Clients, Anchors, and Moderators
   need an HTTP transport for the protocol flows defined by the
   architecture.

   This document defines the Mole HTTP authentication scheme, which
   carries challenges and presentations for the endorsement and
   credential flows, and the headers used to return credential material.
   The grant exchanges with the Anchor are defined per protocol in
   [PROTOCOLS].

            </t>
          </abstract>
        </front>
        <seriesInfo name="Internet-Draft" value="draft-jms-mole-http-transport-01"/>
      </reference>
      <reference anchor="OAUTH-XAA">
        <front>
          <title>Identity Assertion JWT Authorization Grant</title>
          <author fullname="Aaron Parecki" initials="A." surname="Parecki">
            <organization>Okta</organization>
          </author>
          <author fullname="Karl McGuinness" initials="K." surname="McGuinness">
            <organization>Independent</organization>
          </author>
          <author fullname="Brian Campbell" initials="B." surname="Campbell">
            <organization>Ping Identity</organization>
          </author>
          <date day="30" month="October" year="2026"/>
          <abstract>
            <t>   This specification provides a mechanism for an application to use an
   identity assertion to obtain an access token for a third-party API by
   coordinating through an identity provider that the downstream
   Resource Authorization Server already trusts for single sign-on
   (SSO), using OAuth 2.0 Token Exchange (RFC 8693) and the JWT Profile
   for OAuth 2.0 Authorization Grants (RFC 7523).  This pattern is
   informally referred to as Cross-App Access (XAA).

            </t>
          </abstract>
        </front>
        <seriesInfo name="Internet-Draft" value="draft-ietf-oauth-identity-assertion-authz-grant-06"/>
      </reference>
      <reference anchor="WAG">
        <front>
          <title>Workload Authorization Grant</title>
          <author fullname="Paul Carleton" initials="P." surname="Carleton">
            <organization>Anthropic</organization>
          </author>
          <author fullname="Nick Steele" initials="N." surname="Steele">
            <organization>OpenAI</organization>
          </author>
          <author fullname="Aaron Parecki" initials="A." surname="Parecki">
            <organization>Okta</organization>
          </author>
          <author fullname="Arndt Schwenkschuster" initials="A." surname="Schwenkschuster">
            <organization>Defakto Security</organization>
          </author>
          <author fullname="Brian Campbell" initials="B." surname="Campbell">
            <organization>Ping Identity</organization>
          </author>
          <date day="2" month="November" year="2026"/>
          <abstract>
            <t>   This document defines the Workload Authorization Grant (WAG), by
   which a workload hosted on a platform -- an AI agent is the
   motivating case -- obtains access tokens from a third party's OAuth
   authorization server without requiring an administrator to perform a
   per-workload provisioning step.  Each workload is identified by an
   opaque identifier that is never reassigned.  A Workload Identity
   Provider operated by the platform signs a JWT authorization grant
   ([RFC7523]) that names one workload, and the workload presents it at
   the token endpoint of an authorization server that has been
   configured, once, to trust that Workload Identity Provider.  The
   authorization server does not reject a workload because it has not
   seen it before.  The workload's access is determined by the
   authorization server's own policy, which may consult claims the
   Workload Identity Provider asserts about the workload.  This document
   covers workloads acting on their own behalf.  Access on behalf of a
   user or other principal is out of scope, though the grant is intended
   to compose with delegation mechanisms in which the workload is the
   actor.

            </t>
          </abstract>
        </front>
        <seriesInfo name="Internet-Draft" value="draft-carleton-workload-authz-grant-02"/>
      </reference>
      <reference anchor="POPPY" target="https://github.com/personal-agent-protocol/personal-agent-protocol/blob/1de1990e4978b42179fe0e3e0ae554271808c308/content/versions/0.3/spec.mdx">
        <front>
          <title>Personal Agent Protocol, draft 0.3</title>
          <author>
            <organization/>
          </author>
          <date>n.d.</date>
        </front>
      </reference>
      <reference anchor="AAP" target="https://github.com/agentic-protocols/agentic-attestation-protocol/blob/1328a6bbfda24ea18bcd8996a5417a6702c8f274/draft-agent-attestation-protocol.md">
        <front>
          <title>Agent Attestation Protocol</title>
          <author>
            <organization/>
          </author>
          <date>n.d.</date>
        </front>
      </reference>
      <reference anchor="USER-AGENT-HISTORY" target="https://webaim.org/blog/user-agent-string-history/">
        <front>
          <title>WebAIM: History of the browser user-agent string</title>
          <author>
            <organization/>
          </author>
          <date>n.d.</date>
        </front>
      </reference>
      <reference anchor="WBA-ABA">
        <front>
          <title>Anonymous Bot Authentication: Authorization and Rate Limiting for Web Agents</title>
          <author fullname="Eric Rescorla" initials="E." surname="Rescorla">
            <organization>Independent</organization>
          </author>
          <author fullname="Richard Barnes" initials="R." surname="Barnes">
            <organization>Cisco</organization>
          </author>
          <date day="19" month="July" year="2026"/>
          <abstract>
            <t>   Automated agents ("bots") represent a large fraction of the traffic
   to many Web sites.  In some cases, this traffic is desired, in others
   undesired, and in yet others, desired as long as it remains within
   certain rate limits.  This memo describes Anonymous Bot
   Authentication (ABA), a system that allows Web site operators to
   distinguish wanted from unwanted traffic, while not tying a given
   request to a specific sender.

            </t>
          </abstract>
        </front>
        <seriesInfo name="Internet-Draft" value="draft-rescorla-anonymous-webbotauth-01"/>
      </reference>
      <reference anchor="WBA-HTTPSIG">
        <front>
          <title>HTTP Message Signatures for automated traffic</title>
          <author fullname="Thibault Meunier" initials="T." surname="Meunier">
            <organization>Cloudflare</organization>
          </author>
          <author fullname="Sandor Major" initials="S." surname="Major">
            <organization>Google</organization>
          </author>
          <date day="29" month="October" year="2026"/>
          <abstract>
            <t>   This document describes a protocol for identifying automated traffic
   using [HTTP-MESSAGE-SIGNATURES].  The goal is to allow automated HTTP
   clients to cryptographically sign outbound requests, allowing HTTP
   servers to verify their identity with confidence.

   It defines the Signature-Agent header field for in-band key
   discovery, a key directory format based on JWKS, and a well-known URI
   at which that directory is served.

            </t>
          </abstract>
        </front>
        <seriesInfo name="Internet-Draft" value="draft-ietf-webbotauth-httpsig-protocol-01"/>
      </reference>
      <reference anchor="SIGNED-AGENT-METADATA">
        <front>
          <title>The Signed-Agent-Metadata HTTP Header Field for Web Bot Auth</title>
          <author fullname="Rim Vilgalys" initials="R." surname="Vilgalys">
            <organization>Google</organization>
          </author>
          <author fullname="Pablo González" initials="P." surname="González">
            <organization>Google</organization>
          </author>
          <date day="2" month="November" year="2026"/>
          <abstract>
            <t>   Web Bot Auth uses HTTP Message Signatures to let automated HTTP
   clients authenticate their identity to origins via the Signature-
   Agent header field.  When a multi-tenant agent platform signs
   outbound requests with a shared key, all customer workloads on that
   platform share a single cryptographic identity.  Abusive traffic from
   one tenant or workload can degrade the reputation of the entire
   platform and cause origins to block legitimate tenants.

   This document defines the Signed-Agent-Metadata HTTP header field, an
   RFC 9651 Dictionary that carries optional, pseudonymous per-request
   metadata (product, project, and session).  By signing Signed-Agent-
   Metadata inside the Web Bot Auth signature, a platform enables
   verifiers to evaluate traffic, apply rate limits, and isolate
   reputation at a narrower scope than the signing origin.

            </t>
          </abstract>
        </front>
        <seriesInfo name="Internet-Draft" value="draft-vilgalys-webbotauth-signed-agent-metadata-00"/>
      </reference>
      <reference anchor="WIMSE-AIMS">
        <front>
          <title>AI Identity Management System</title>
          <author fullname="Pieter Kasselman" initials="P." surname="Kasselman">
            <organization>Defakto Security</organization>
          </author>
          <author fullname="Jeff Lombardo" initials="J." surname="Lombardo">
            <organization>AWS</organization>
          </author>
          <author fullname="Yaroslav Rosomakho" initials="Y." surname="Rosomakho">
            <organization>Zscaler</organization>
          </author>
          <author fullname="Brian Campbell" initials="B." surname="Campbell">
            <organization>Ping Identity</organization>
          </author>
          <author fullname="Nick Steele" initials="N." surname="Steele">
            <organization>OpenAI</organization>
          </author>
          <author fullname="Aaron Parecki" initials="A." surname="Parecki">
            <organization>Okta</organization>
          </author>
          <date day="2" month="November" year="2026"/>
          <abstract>
            <t>   This document proposes best practices for authentication and
   authorization of AI agent interactions.  It leverages existing
   standards such as the Workload Identity in Multi-System Environments
   (WIMSE) architecture and OAuth 2.0 family of specifications.  Rather
   than defining new protocols, this document describes how existing and
   widely deployed standards can be applied or extended to establish
   agent authentication and authorization.  By doing so, it aims to
   provide a framework within which to use existing standards, identify
   gaps and guide future standardization efforts for agent
   authentication and authorization.

            </t>
          </abstract>
        </front>
        <seriesInfo name="Internet-Draft" value="draft-ietf-wimse-aims-01"/>
      </reference>
      <reference anchor="WIMSE-ARCH">
        <front>
          <title>Workload Identity in a Multi System Environment (WIMSE) Architecture</title>
          <author fullname="Joseph A. Salowey" initials="J. A." surname="Salowey">
            <organization>Palo Alto Networks</organization>
          </author>
          <author fullname="Yaroslav Rosomakho" initials="Y." surname="Rosomakho">
            <organization>Zscaler</organization>
          </author>
          <author fullname="Hannes Tschofenig" initials="H." surname="Tschofenig">
            <organization>University of the Bundeswehr Munich</organization>
          </author>
          <date day="6" month="July" year="2026"/>
          <abstract>
            <t>   The increasing prevalence of cloud computing and micro service
   architectures has led to the rise of complex software functions being
   built and deployed as workloads, where a workload is defined as
   software executing for a specific purpose, potentially comprising one
   or more running instances.  This document discusses an architecture
   for designing and standardizing protocols and payloads for conveying
   workload identity and security context information.

            </t>
          </abstract>
        </front>
        <seriesInfo name="Internet-Draft" value="draft-ietf-wimse-arch-08"/>
      </reference>
    </references>
    <?line 739?>

<section anchor="acknowledgments">
      <name>Acknowledgments</name>
      <t>The authors would like to thank Bryan Becker, Marwan Fayed, and Mari Galicer for
discussions that helped shape this document.</t>
    </section>
    <section anchor="capacity-without-identifiers">
      <name>Capacity without identifiers</name>
      <t>Signed product, project, and session IDs give an origin visible keys with which
to keep separate counters <xref target="SIGNED-AGENT-METADATA"/>. For capacity alone, the
Anchor can constrain Endorsements per product, project, or task allocation
only if it knows and enforces those boundaries. The platform can play this
role. An independent Anchor needs a way to account for the same allocations,
or its Endorsements cannot enforce the platform's project and product totals.</t>
      <t>Clients redeem those endorsements for independent MOLE credentials and
spend the credentials on requests. The Moderator checks
each spend and carries the remaining balance into an unlinkable
replacement Credential. The presentation itself does not reveal which
product, project, task, or Client holds a particular budget.</t>
      <t>The platform can still sign WBA(P), so the origin can also limit its traffic in aggregate.
Credential contexts and spending terms should be shared across many
endorsements. Unique contexts or amounts would distinguish those units and
shrink their anonymity sets.</t>
      <t>The allocation boundary can be chosen without adding a corresponding ID to
requests:</t>
      <ul spacing="normal">
        <li>
          <t>Product: A platform or other suitable Anchor can reserve a finite pool of
Endorsements for a product and distribute them among its projects or
Clients. Spending is limited by the budgets derived from that pool. The
origin does not learn which product a presentation came from, nor can it
impose a different per-product policy from presentations alone.</t>
        </li>
        <li>
          <t>Project: The platform can cap Endorsements for a tenant workspace and
allocate them among its Clients. Exhausting one Client's independent
Credential does not exhaust another's. The platform, not the origin,
accounts for how many endorsements belong to that project. A single
project-wide balance shared by many concurrent Clients is a different
requirement, discussed below.</t>
        </li>
        <li>
          <t>Session: An Anchor can allocate Endorsements to a task (like an ephemeral
workload). The Anchor must count task allocations against a scarce parent
allowance so starting new tasks does not reset the limit. The allocating
party and Client must keep the resulting Credentials assigned to those
tasks: a MOLE presentation does not prove which task is using a Credential.
Alternatively, a Client can spend the same Credential across tasks when
task-specific budgets are unnecessary. Neither choice lets the origin link
requests as one task. Such correlation requires an additional mechanism.</t>
        </li>
      </ul>
      <t>The example in <xref target="scenarios"/> bounds a tenant's total allocation while the
origin checks each Credential separately. Applying that policy to a project
requires the Anchor to account for all allocations to that project.</t>
      <t>The Anchor can also cap the sum of Endorsements across projects belonging
to one product. The origin enforces each Credential's ten-credit limit
without seeing the project or product. The project and product totals depend
on the Anchor's issuance policy, not on an origin-side counter keyed by an ID.
Additional accepted Anchors or Endorsements increase the available budget
unless their policies constrain the combined allocation. Public-baseline
requests and the platform's aggregate limit remain separate from this bound.</t>
      <t>These budgets replace identifiers only for the capacity decisions their
issuance and spending policies implement. An origin cannot use an anonymous
presentation to identify and block a particular project, audit its activity,
or correlate all requests in a session. If it needs those functions, signed
metadata can accompany MOLE, making requests linkable at the metadata's scope.</t>
    </section>
  </back>
  <!-- ##markdown-source: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-->

</rfc>
