<?xml version='1.0' encoding='utf-8'?>
<!DOCTYPE rfc [
  <!ENTITY nbsp    "&#160;">
  <!ENTITY zwsp   "&#8203;">
  <!ENTITY nbhy   "&#8209;">
  <!ENTITY wj     "&#8288;">
]>
<!-- name="GENERATOR" content="github.com/mmarkdown/mmark Mmark Markdown Processor - mmark.miek.nl" -->
<rfc xmlns:xi="http://www.w3.org/2001/XInclude" version="3" ipr="trust200902" docName="draft-bormann-jwp-modular-bbs-02" submissionType="IETF" category="std" xml:lang="en" indexInclude="true">

<front>
<title abbrev="JWP-BBS">BBS and Modular Sub-proofs with JSON Web Proofs</title><seriesInfo value="draft-bormann-jwp-modular-bbs-02" stream="IETF" status="standard" name="Internet-Draft"/>
<author initials="C." surname="Bormann" fullname="Christian Bormann"><organization>SPRIND GmbH</organization><address><postal><street/>
</postal><email>chris.bormann@gmx.de</email>
</address></author><date/>
<area>Security</area>
<workgroup>JOSE</workgroup>
<keyword>zero-knowledge proofs</keyword>
<keyword>multi-message signatures</keyword>
<keyword>BBS</keyword>
<keyword>JWP</keyword>

<abstract>
<t>This document defines a digital credential format that uses JSON Web Proofs (JWP) as its container format and Blind BBS Signatures as its signature scheme combined with a modular framework for attaching zero-knowledge sub-proofs. This allows a Holder to reveal some attributes directly while proving predicates such as range or equality over the ones they keep hidden. A credential can additionally be bound to an ECDSA P-256 device key, with possession of the key proven in every presentation without revealing the public key. The credential type definition and data model follow SD-JWT VC <xref target="I-D.ietf-oauth-sd-jwt-vc"/>.</t>
</abstract>

<note title="Discussion Venues" removeInRFC="true">
<t>Source for this draft and an issue tracker can be found at
    <eref target="https://github.com/c2bo/draft-bormann-jwp-modular-bbs"/>.</t>
</note>
</front>

<middle>

<section anchor="introduction"><name>Introduction</name>
<t>The BBS signature scheme <xref target="I-D.irtf-cfrg-bbs-signatures"/> is a multi-message signature (MMS) scheme where the signer produces a single signature over a vector of messages m<em>0 through m</em>(n-1), and the Holder can prove knowledge of the signature in zero knowledge while disclosing only a chosen subset of those messages.</t>
<t>The Blind BBS Signatures extension <xref target="I-D.irtf-cfrg-bbs-blind-signatures"/> adds Pedersen commitments to the scheme that allow the Holder to mark each message as disclosed, hidden, or committed at proof time, and the resulting proof carries a fresh Pedersen commitment for every committed message. Those commitments become public inputs to further proofs over the values they hide.</t>
<t>Building on those core building blocks, this document defines a digital credential format that:</t>

<ul spacing="compact">
<li>Uses JSON Web Proofs <xref target="I-D.ietf-jose-json-web-proof"/> as the serialization/container format for both issuance and presentation, and defines a JSON Proof Algorithm <xref target="I-D.ietf-jose-json-proof-algorithms"/> profile based on Blind BBS Signatures.</li>
<li>Builds its core proof on <tt>CoreProofGen</tt> of <xref target="I-D.irtf-cfrg-bbs-blind-signatures"/>, exposing fresh Pedersen commitments to selected messages as public inputs for sub-proofs.</li>
<li>Defines a sub-proof container carrying optional sub-proofs, each bound to the core proof via a Pedersen commitment.</li>
<li>Optionally binds a credential to an ECDSA P-256 device key by encoding that key as messages in the BBS signature vector.</li>
</ul>
<t>This modular architecture builds on prior work <xref target="TS14"/> and <xref target="LSZ25"/>, and the credential type model is reused from SD-JWT VC <xref target="I-D.ietf-oauth-sd-jwt-vc"/>.</t>

<sourcecode type="ascii-art"><![CDATA[ +----------------+
 |                | ---------> +----------------+
 |                |            |   Revealed     |
 |                |            |   Attributes   |
 |                |            +----------------+
 |                |                    |
 |                |                    |
 |                | ---------> +--------------+      +--------------+
 |      MMS       |            |  Commitment  | ---> |  Sub-Proof   |
 |   Signature    |            +--------------+      +--------------+
 |                |                    |
 |                |                    |
 |                | ---------> +--------------+      +--------------+
 |                |            |  Commitment  | ---> |  Sub-Proof   |
 |                |            +--------------+      +--------------+
 |                |                    |
 +----------------+                    |
        |                              |
        |         +-----+------+-------+
        |         |     |      |
        v         v     v      v
        +----------------------------------------> +----------------+
        (revealed + commitment openings feed down) |    Core Proof  |
                                                   +----------------+
]]>
</sourcecode>

<section anchor="requirements-notation-and-conventions"><name>Requirements Notation and Conventions</name>
<t>The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD",
"SHOULD NOT", "RECOMMENDED", "NOT RECOMMENDED", "MAY", and "OPTIONAL" in this
document are to be interpreted as described in BCP 14 <xref target="RFC2119"/> <xref target="RFC8174"/>
when, and only when, they appear in all capitals, as shown here.</t>
</section>

<section anchor="notational-conventions"><name>Notational Conventions</name>
<t>All examples in this document are non-normative.</t>
<t>Indexing into vectors is 0-based. The notation <tt>m_i</tt> denotes the i-th element of the message vector: <tt>m_0</tt> is the first element. Ranges are written <tt>[a, b]</tt> for inclusive endpoints and <tt>[a, b)</tt> for a half-open interval.</t>
</section>

<section anchor="terms-and-definitions"><name>Terms and Definitions</name>
<t>This document uses the Issuer-Holder-Verifier model and terminology of <xref target="I-D.ietf-oauth-sd-jwt-vc"/>.</t>
<t>Additional terms used are:</t>

<dl spacing="compact">
<dt>Core proof:</dt>
<dd>A zero-knowledge proof of knowledge of a BBS signature on a message vector, where some messages are disclosed and others are exposed only as commitments.</dd>
<dt>Sub-proof:</dt>
<dd>A zero-knowledge proof attached to a core proof, asserting a predicate over a message whose Pedersen commitment that core proof exposes.</dd>
<dt>Committed disclosure:</dt>
<dd>Exposing a Pedersen commitment to a signed message in place of the value itself that is used as an input for sub-proofs.</dd>
<dt>Device binding:</dt>
<dd>Tying a credential presentation to control of a Holder-held private key, by carrying a fresh proof of possession in every presentation.</dd>
</dl>
</section>
</section>

<section anchor="data-model"><name>Data Model</name>
<t>A credential exists in two forms: the Issued Form an Issuer transmits to a Holder, and the Presented Form a Holder derives from it for a Verifier (see <xref target="presentation"/>).</t>

<section anchor="issued-credential"><name>Issued Credential</name>
<t>A credential is issued in the Issued Form (see <xref target="I-D.ietf-jose-json-web-proof" sectionFormat="of" section="6.1"/>) consisting of:</t>

<ul spacing="compact">
<li>An Issuer Header (<xref target="I-D.ietf-jose-json-web-proof" sectionFormat="of" section="6.1.1"/>) with the contents specified in <xref target="issuer-header"/>.</li>
<li><tt>n</tt> Issuer Payloads (<xref target="I-D.ietf-jose-json-web-proof" sectionFormat="of" section="6.1.2"/>), where <tt>n</tt> is the length of the BBS message vector (see <xref target="claims-mapping"/>). The Issuer Payload at position <tt>i</tt> is the octet string from which the scalar message <tt>m_i</tt> is derived per <xref target="message-derivation"/>.</li>
<li>An Issuer Proof (<xref target="I-D.ietf-jose-json-web-proof" sectionFormat="of" section="6.1.3"/>) carrying the BBS signature over <tt>header_octets</tt> and the message vector <tt>(m_0, ..., m_(n-1))</tt>.</li>
</ul>
<t><tt>header_octets</tt> is the Issuer Header as transmitted, i.e., the octets obtained by base64url-decoding the Issuer Header component of the Compact Serialization. All parties MUST use those octets as received and MUST NOT alter the header (e.g., re-encode).</t>
<t>The Issued Form is serialized using the Compact Serialization (see <xref target="I-D.ietf-jose-json-web-proof" sectionFormat="of" section="7.1"/>). CBOR Serialization is (currently) out of scope for this document.</t>
</section>

<section anchor="issuer-header"><name>Issuer Header</name>
<t>The Issuer Header is a JSON object with the following Header Parameters.</t>

<dl spacing="compact">
<dt><tt>alg</tt> (REQUIRED):</dt>
<dd>The Algorithm Header Parameter (<xref target="I-D.ietf-jose-json-web-proof" sectionFormat="of" section="5.2.1"/>). This profile defines the JPA value <tt>BBS-MOD</tt> (see <xref target="cipher-suite"/>).</dd>
<dt><tt>vct</tt> (string, REQUIRED):</dt>
<dd>The credential type identifier as defined in <xref target="I-D.ietf-oauth-sd-jwt-vc" sectionFormat="of" section="2.2.2.1"/>.</dd>
<dt><tt>cmap</tt> (JSON object, REQUIRED):</dt>
<dd>The mapping from claim names to message-vector positions and per-message encoding - see <xref target="claims-mapping"/> for more details.</dd>
<dt><tt>kb</tt> (string, OPTIONAL):</dt>
<dd>The device-binding identifier - see <xref target="device-binding-header"/>. When absent, the credential is not device-bound, and a presentation MUST NOT include a device-binding sub-proof.</dd>
</dl>
<t>Temporal claims (<tt>exp</tt>, <tt>nbf</tt>, <tt>iat</tt>) MUST NOT appear as Issuer Header values - see <xref target="temporal-claims"/> for more details.</t>
<t>The JWP <tt>iek</tt>, <tt>hpk</tt>, and <tt>hpa</tt> Header Parameters (Sections 5.2.5, 5.2.6, and 5.2.7 of <xref target="I-D.ietf-jose-json-web-proof"/>) MUST NOT appear in the Issuer Header.</t>
</section>

<section anchor="claims-mapping"><name>Claims Mapping</name>
<t><tt>cmap</tt> mirrors the credential's JSON tree structurally. Each leaf is replaced by an index annotation: a two-element JSON array <tt>[i, scalar]</tt>, where:</t>

<ul spacing="compact">
<li><tt>i</tt> is the 0-based index of the leaf value in the message vector.</li>
<li><t><tt>scalar</tt> is a boolean selecting how the leaf becomes the BBS message m_i:</t>

<ul spacing="compact">
<li><tt>false</tt>: the leaf is encoded as octets and mapped to a scalar via the cipher suite's hash-to-scalar primitive (see <xref target="message-derivation"/>).</li>
<li><tt>true</tt>: the leaf MUST be a JSON integer in <tt>[0, r - 1]</tt> (where <tt>r</tt> is the order of the BBS scalar field) and is used directly as m_i (see <xref target="scalar-encoding"/>).</li>
</ul></li>
</ul>
<t>Let <tt>n</tt> be the length of the message vector, and <tt>N</tt> the number of payload slots reserved for the device-key encoding (see <xref target="device-binding-header"/>), with <tt>N = 0</tt> when <tt>kb</tt> is absent. Every index in <tt>[N, n-1]</tt> MUST appear in exactly one annotation in <tt>cmap</tt>. Indices <tt>[0, N-1]</tt> MUST NOT appear in <tt>cmap</tt>.</t>
<t>The top-level member names of <tt>cmap</tt> MUST NOT be <tt>vct</tt>, <tt>alg</tt>, <tt>cmap</tt>, or <tt>kb</tt>. This keeps the reconstructed payload (see <xref target="reconstructed-payload"/>) free of collisions with the <tt>vct</tt> member taken from the Issuer Header and prevents claim values from masquerading as header-derived members.</t>
<t>Receivers MUST validate <tt>cmap</tt> before use: every leaf is a two-element annotation of a non-negative integer index and a boolean, every index in <tt>[N, n-1]</tt> appears in exactly one annotation, no other index appears, and the name restrictions above hold. Holders MUST reject an issued credential and Verifiers MUST reject a presentation that violates any of these constraints.</t>
<t>Payload slots defined by the credential type's structural layout (see <xref target="layout"/>) but not populated by a given credential MUST carry the decoy value defined in <xref target="decoys"/>.</t>
</section>

<section anchor="example-issuer-header"><name>Example: Issuance</name>
<t>Starting from an SD-JWT VC-style claim set <xref target="I-D.ietf-oauth-sd-jwt-vc"/>:</t>

<sourcecode type="json"><![CDATA[{
  "vct": "https://credentials.example.com/identity_credential",
  "given_name": "Erika",
  "family_name": "Mustermann",
  "email": "erika@example.com",
  "phone_number": "+49 123456789",
  "address": {
    "street_address": "Heidestraße 17",
    "locality": "Köln",
    "region": "Nordrhein-Westfalen",
    "country": "DE"
  },
  "birthdate": 19630812,
  "iat": 1683000000,
  "exp": 1786000000
}
]]>
</sourcecode>
<t>The <tt>vct</tt> claim becomes a Header Parameter and the other 11 attributes become leaves in <tt>cmap</tt>, with <tt>address</tt> mirrored as a nested object. No device binding is used, so <tt>N = 0</tt> and the leaves occupy indices 0 through 10. The temporal claims <tt>iat</tt> and <tt>exp</tt> are carried as <tt>scalar = true</tt> leaves (see <xref target="temporal-claims"/>) to allow range sub-proofs over them. The resulting Issuer Header is:</t>

<sourcecode type="json"><![CDATA[{
  "alg": "BBS-MOD",
  "vct": "https://credentials.example.com/identity_credential",
  "cmap": {
    "given_name": [0, false],
    "family_name": [1, false],
    "email": [2, false],
    "phone_number": [3, false],
    "address": {
      "street_address": [4, false],
      "locality": [5, false],
      "region": [6, false],
      "country": [7, false]
    },
    "birthdate": [8, true],
    "iat": [9, true],
    "exp": [10, true]
  }
}
]]>
</sourcecode>
<t>Indices 0–7 use hash-to-scalar and indices 8–10 carry their integer values directly as scalars, with <tt>iat</tt> and <tt>exp</tt> as NumericDate integers (<xref target="RFC7519"/>). A presentation can then mark <tt>iat</tt>/<tt>exp</tt> as <tt>COMMIT</tt> (see <xref target="core-proof"/>) and attach <tt>sigma-range</tt> sub-proofs (see <xref target="range-proof"/>) to prove validity without disclosing the timestamps.</t>
<t>A real deployment would define a structural layout covering all optional attributes and array slots up to their maximum length, with absent slots filled by decoys (see <xref target="decoys"/>).</t>
</section>

<section anchor="message-derivation"><name>Message Derivation</name>
<t>For an annotation <tt>[i, false]</tt> with leaf value <tt>v</tt>:</t>

<ol spacing="compact">
<li><tt>o</tt> is a JSON serialization of <tt>v</tt> - a single JSON text <xref target="RFC8259"/> encoded in UTF-8 (e.g., <tt>"Erika"</tt> for a string, <tt>true</tt> for a boolean) - carried as Issuer Payload <tt>i</tt>. The Issuer MAY produce any serialization of <tt>v</tt>, as the payload octets rather than the abstract value are what is mapped to the message scalar. Holders and Verifiers MUST use the received payload octets as-is and MUST NOT re-serialize them.</li>
<li>m_i = <tt>hash_to_scalar(o, map_dst)</tt>, with <tt>map_dst = api_id || "MAP_MSG_TO_SCALAR_AS_HASH_"</tt> and <tt>api_id</tt> the Interface identifier of <xref target="cipher-suite"/>. This is the per-message derivation of <tt>BBS.messages_to_scalars</tt> (<xref target="I-D.irtf-cfrg-bbs-signatures" sectionFormat="of" section="4.1.2"/>).</li>
</ol>
<t>Numeric leaves recovered via JSON parsing are subject to JSON number-precision interoperability limits - Issuers SHOULD keep <tt>scalar = false</tt> number values within the I-JSON <xref target="RFC7493"/> range.</t>
<t>For an annotation <tt>[i, true]</tt> with leaf value <tt>v</tt>:</t>

<ol spacing="compact">
<li><tt>o</tt> is the canonical decimal octet encoding of <tt>v</tt> (see <xref target="scalar-encoding"/>), carried as Issuer Payload <tt>i</tt>.</li>
<li>m_i is the integer denoted by <tt>o</tt>, interpreted as an element of the BBS scalar field.</li>
</ol>
</section>

<section anchor="scalar-encoding"><name>Scalar Encoding</name>
<t>A leaf with <tt>scalar = true</tt> MUST be a JSON integer in <tt>[0, r - 1]</tt>, where <tt>r</tt> is the order of the BBS scalar field. Implementations MUST reject any other value.</t>
<t>The Issuer Payload for such a leaf is the canonical decimal octet encoding of the integer: ASCII digits without sign or leading zeros, with <tt>0</tt> represented as the single digit <tt>0</tt>. Future extensions MAY define additional scalar encodings provided they deterministically map a JSON value to an element of <tt>[0, r - 1]</tt>.</t>
</section>

<section anchor="temporal-claims"><name>Temporal Claims</name>
<t>The JWT temporal claims <tt>exp</tt>, <tt>nbf</tt>, and <tt>iat</tt> (<xref target="RFC7519" sectionFormat="of" section="4.1"/>), when present in a credential, MUST be declared as <tt>scalar = true</tt> leaves in <tt>cmap</tt> carrying their NumericDate values. They MUST NOT appear as Issuer Header values.</t>
</section>

<section anchor="device-binding-header"><name>Device Binding Header</name>
<t>When present, the <tt>kb</tt> Header Parameter is a string identifier selecting both the device public key type and its encoding into the BBS message vector. The reserved slots are always indices <tt>[0, N-1]</tt>, where <tt>N</tt> depends on the <tt>kb</tt> value. If <tt>kb</tt> is not present, no slots are reserved. This document defines a single value for <tt>kb</tt>: <tt>ecdsa-p256-db</tt>.</t>
<t>A <tt>kb</tt> value and its matching device-binding sub-proof algorithm (see <xref target="sub-proofs"/>) share the same algorithm identifier string. Valid <tt>kb</tt> values are the entries of the Sub-Proof Algorithms registry whose Device Binding field is <tt>yes</tt> - see <xref target="iana"/>. The specification defining such an entry MUST define the number of reserved slots <tt>N</tt>, the encoding of the device public key into indices <tt>[0, N-1]</tt>, and the matching device-binding sub-proof.</t>
<t>For <tt>kb = "ecdsa-p256-db"</tt>, <tt>N = 4</tt> and:</t>

<ul spacing="compact">
<li>m_0..m_1 encode the x-coordinate of the device public key as two 128-bit little-endian limbs (m_0 least significant).</li>
<li>m_2..m_3 encode the y-coordinate the same way.</li>
</ul>
<t>Each limb is encoded as if <tt>scalar = true</tt>: the Issuer Payload is its canonical decimal octet encoding (see <xref target="scalar-encoding"/>).</t>
<t>The Issuer MUST make sure that <tt>(x, y)</tt> is a valid non-identity P-256 point <xref target="FIPS186-5"/> before computing the message vector.</t>
</section>

<section anchor="layout"><name>Structural Layout</name>
<t>For claims containing objects, the Issuer either mirrors the object structure within <tt>cmap</tt> or treats the JSON-encoded object as a single leaf. This is a policy decision by the Issuer and allows some objects to be discloseable only as one object containing all values or not at all.</t>
<t>For bounded-length array claims, <tt>cmap</tt> contains a JSON array of index annotations sized to the credential type's maximum array length. All entries in such an array SHOULD share the same <tt>scalar</tt> flag to guarantee a single decoy encoding (see <xref target="decoys"/>).</t>
<t>For optional claims, <tt>cmap</tt> MUST contain the index entry regardless of whether the attribute is present in a given credential.</t>
</section>

<section anchor="decoys"><name>Decoys</name>
<t>Decoys fill payload slots that the credential type's layout defines, but a specific credential does not populate. They keep the message-vector length and <tt>cmap</tt> identical across all credentials of a given <tt>vct</tt> to avoid correlation.</t>
<t>Every decoy slot carries the same fixed scalar:</t>

<artwork><![CDATA[m_decoy = hash_to_scalar("JWP-BBS-DECOY", map_dst)
]]>
</artwork>
<t>with <tt>hash_to_scalar</tt> and <tt>map_dst</tt> as defined in <xref target="message-derivation"/>.</t>
<t>The Issuer Payload for a decoy slot depends on the slot's <tt>scalar</tt> flag:</t>

<ul spacing="compact">
<li><tt>scalar = false</tt>: the ASCII octets of <tt>"JWP-BBS-DECOY"</tt>.</li>
<li><tt>scalar = true</tt>: the canonical decimal octet encoding of <tt>m_decoy</tt> (see <xref target="scalar-encoding"/>).</li>
</ul>
<t>A Verifier detects a disclosed decoy by comparing the disclosed Presentation Payload octets to the fixed decoy octets defined above. The <tt>scalar = false</tt> decoy octets are deliberately not a valid JSON text, so no payload produced per <xref target="message-derivation"/> can collide with them. Decoys SHOULD NOT be disclosed unless required by the use case (for example, a proof over all members of a bounded-length array).</t>
</section>
</section>

<section anchor="issuance"><name>Issuance</name>

<section anchor="issuer-key-generation"><name>Issuer Key Generation</name>
<t>The Issuer key pair is a BBS key pair (<xref target="I-D.irtf-cfrg-bbs-signatures" sectionFormat="of" section="3.4"/>) using the cipher suite of <xref target="cipher-suite"/>.</t>
</section>

<section anchor="credential-issuance"><name>Credential Issuance</name>
<t>To issue a credential, the Issuer performs the following steps:</t>

<ol spacing="compact">
<li>Construct the Issuer Header per <xref target="issuer-header"/> and <xref target="claims-mapping"/>.</li>
<li>Derive the message vector <tt>(m_0, ..., m_(n-1))</tt> per <xref target="message-derivation"/> and <xref target="device-binding-header"/>, filling decoys per <xref target="decoys"/>.</li>
<li>Compute the signature with <tt>CoreSign</tt> (<xref target="I-D.irtf-cfrg-bbs-signatures" sectionFormat="of" section="3.6.1"/>) over <tt>generators = create_generators(n + 1, api_id)</tt>, <tt>header_octets</tt>, and the message vector, with <tt>api_id</tt> as in <xref target="cipher-suite"/>. No messages are Holder-committed at issuance, so the <tt>Commit</tt>/<tt>BlindSign</tt> flow of <xref target="I-D.irtf-cfrg-bbs-blind-signatures"/> is not used.</li>
<li>Assemble and serialize the Issued Form per <xref target="issued-credential"/>.</li>
</ol>
<t>A non-normative example of the Compact Serialization:</t>

<artwork><![CDATA[<base64url(Issuer Header)>
.
<m_0>~<m_1>~ ... ~<m_10>
.
<base64url(BBS signature)>
]]>
</artwork>
<t>Each <tt>&lt;m_i&gt;</tt> is the base64url-encoded Issuer Payload for index <tt>i</tt> (e.g., m_1 is <tt>"Mustermann"</tt> including the quotes, m_10 is <tt>1786000000</tt>). For <tt>scalar = true</tt> leaves the canonical decimal encoding coincides with the JSON serialization of the integer.</t>
</section>

<section anchor="holder-verification"><name>Holder Verification</name>
<t>The Holder verifies an issued credential by:</t>

<ol spacing="compact">
<li>Parsing the Issued Form.</li>
<li>Validating the <tt>cmap</tt> object per <xref target="claims-mapping"/>. Reject on violation.</li>
<li>Verifying the signature with <tt>CoreVerify</tt> (<xref target="I-D.irtf-cfrg-bbs-signatures" sectionFormat="of" section="3.6.2"/>) over the same generators, <tt>header_octets</tt>, and message vector as issuance. Reject on failure.</li>
<li>For every <tt>scalar = true</tt> leaf, confirming the corresponding Issuer Payload decodes to an integer in <tt>[0, r - 1]</tt>.</li>
<li>For every <tt>scalar = false</tt> leaf, confirming the corresponding Issuer Payload either is byte-equal to the decoy octets (see <xref target="decoys"/>) or parses as a single JSON text <xref target="RFC8259"/>.</li>
<li>If <tt>kb</tt> is present, confirming that the point reconstructed from the limb messages matches the Holder's device public key. How the Holder obtains the corresponding device key pair is out of scope.</li>
</ol>
</section>
</section>

<section anchor="presentation"><name>Presentation</name>

<section anchor="presented-form"><name>Presented Form</name>
<t>A presentation is a Presented Form (<xref target="I-D.ietf-jose-json-web-proof" sectionFormat="of" section="6.2"/>) consisting of:</t>

<ol spacing="compact">
<li>A Presentation Header as defined in <xref target="presentation-header"/>.</li>
<li>The unmodified Issuer Header.</li>
<li><tt>n</tt> Presentation Payloads (<xref target="I-D.ietf-jose-json-web-proof" sectionFormat="of" section="6.2.2"/>): disclosed positions carry the corresponding Issuer Payload and undisclosed positions are omitted (see <xref target="I-D.ietf-jose-json-web-proof" sectionFormat="of" section="7.1"/>).</li>
<li>A Presentation Proof (<xref target="I-D.ietf-jose-json-web-proof" sectionFormat="of" section="6.2.4"/>) consisting of one or more octet strings. The first octet string is the encoded core proof (see <xref target="core-proof"/>). Subsequent optional octet strings are UTF-8 JSON-serialized sub-proof objects (see <xref target="sub-proofs"/>) and MAY appear in any order. The Compact Serialization base64url-encodes each octet string.</li>
</ol>
</section>

<section anchor="presentation-header"><name>Presentation Header</name>
<t>The Presentation Header is a JSON object with the following Header Parameters.</t>

<dl spacing="compact">
<dt><tt>alg</tt> (REQUIRED):</dt>
<dd>The Algorithm Header Parameter (<xref target="I-D.ietf-jose-json-web-proof" sectionFormat="of" section="5.2.1"/>). MUST be identical to the <tt>alg</tt> value of the Issuer Header.</dd>
<dt><tt>nonce</tt> (string, REQUIRED):</dt>
<dd>The Nonce Header Parameter (<xref target="I-D.ietf-jose-json-web-proof" sectionFormat="of" section="5.2.10"/>).</dd>
<dt><tt>aud</tt> (string, REQUIRED):</dt>
<dd>The Audience Header Parameter (<xref target="I-D.ietf-jose-json-web-proof" sectionFormat="of" section="5.2.9"/>).</dd>
</dl>
<t>Additional Header Parameters MAY be present, but their use is out of scope for this document.</t>
<t><tt>presentation_header_octets</tt> is the Presentation Header as transmitted, i.e., the octets obtained by base64url-decoding the Presentation Header component of the Compact Serialization. It is bound into the core proof challenge (see <xref target="core-proof"/>). Verifiers MUST use those octets as received.</t>
</section>

<section anchor="core-proof"><name>Core Proof</name>
<t>The Holder builds a per-message disclosure map assigning each index in <tt>[0, n-1]</tt> one of <tt>DISCLOSE</tt>, <tt>HIDE</tt>, or <tt>COMMIT</tt>:</t>

<ul spacing="compact">
<li><tt>DISCLOSE</tt>: the message is revealed and its value MUST match the corresponding disclosed Presentation Payload.</li>
<li><tt>COMMIT</tt>: a fresh Pedersen commitment to the message is carried in the proof. Every index referenced by a sub-proof (see <xref target="sub-proofs"/>) MUST be marked <tt>COMMIT</tt>.</li>
<li><tt>HIDE</tt>: all other indices in <tt>[0, n-1]</tt></li>
</ul>
<t>The Holder generates the core proof by invoking <tt>CoreProofGen</tt> of <xref target="I-D.irtf-cfrg-bbs-blind-signatures"/> with:</t>

<ul spacing="compact">
<li><tt>PK</tt>: Issuer public key.</li>
<li><tt>signature</tt>: BBS signature from the Issuer Proof.</li>
<li><tt>generators</tt>: <tt>create_generators(n + 1, api_id)</tt> (see <xref target="I-D.irtf-cfrg-bbs-signatures" sectionFormat="of" section="4.1.1"/>).</li>
<li><tt>header</tt>: <tt>header_octets</tt>.</li>
<li><tt>ph</tt>: <tt>presentation_header_octets</tt> (binds <tt>nonce</tt> and <tt>aud</tt> into the challenge).</li>
<li><tt>messages</tt>: <tt>(m_0, ..., m_(n-1))</tt>.</li>
<li><tt>disclosed_indexes</tt>: indices marked <tt>DISCLOSE</tt>.</li>
<li><tt>commits_indexes</tt>: indices marked <tt>COMMIT</tt>.</li>
<li><tt>api_id</tt>: the cipher suite identifier of <xref target="cipher-suite"/>.</li>
</ul>
<t><tt>CoreProofGen</tt> returns <tt>(proof, add_zkp_info)</tt>. <tt>add_zkp_info</tt> contains, per committed index, the Pedersen commitment <tt>C_i</tt> and the blinding scalar <tt>s_i</tt>. The Holder retains it locally to build sub-proofs and MUST NOT transmit it. Only <tt>proof</tt> is carried as the first octet string of the Presentation Proof.</t>
<t>The core proof establishes that the Holder knows a BBS signature under the Issuer's public key on a message vector whose disclosed-index values match the disclosed Presentation Payloads, and that each carried <tt>C_i</tt> commits to the message at index <tt>i</tt> of that vector.</t>
<t>The Verifier verifies the core proof with <tt>CoreProofVerify</tt>, passing <tt>PK</tt>, the core proof, the generators, <tt>header_octets</tt>, <tt>presentation_header_octets</tt>, the disclosed scalar messages, and <tt>api_id</tt>. The disclosed and committed indices are recovered from the proof octets, not passed separately. On success, the Verifier recovers the committed indices and the corresponding <tt>C_i</tt> from the proof octets which are used in the sub-proof verification (see <xref target="sub-proofs"/>).</t>
</section>

<section anchor="sub-proofs"><name>Sub-Proofs</name>
<t>A sub-proof is a JSON object carried as an additional octet string of the Presentation Proof (see <xref target="presented-form"/>) with the following members:</t>

<dl spacing="compact">
<dt><tt>alg</tt> (string, REQUIRED):</dt>
<dd>The sub-proof algorithm identifier from the Sub-Proof Algorithms registry (see <xref target="iana"/>).</dd>
<dt><tt>input</tt> (JSON object, REQUIRED):</dt>
<dd>Public inputs to the sub-proof. MUST contain <tt>i</tt> and MAY contain algorithm-specific members.</dd>
</dl>
<t><tt>i</tt> is a non-empty JSON array of message-vector indices, each of which MUST be a <tt>COMMIT</tt>-marked index of the core proof. Each algorithm fixes the length of <tt>i</tt> and the role of its entries.</t>

<dl spacing="compact">
<dt><tt>proof</tt> (string, REQUIRED):</dt>
<dd>The base64url <xref target="RFC4648"/> encoding of the sub-proof bytes specified by <tt>alg</tt>.</dd>
</dl>
<t>For each sub-proof, the Verifier MUST confirm that every value in <tt>i</tt> is among the committed indices recovered from the core proof, and MUST then run the algorithm-specific verification routine against the corresponding <tt>C_i</tt>, <tt>input</tt>, and <tt>proof</tt>.</t>
<t>Sub-proof freshness is inherited from the core proof: every <tt>C_i</tt> is randomized per presentation, and the core proof's challenge binds to <tt>presentation_header_octets</tt>. Sub-proof algorithms that include public material not derived from <tt>C_i</tt> (for example, the device ECDSA signature in <tt>ecdsa-p256-db</tt>) MUST bind that material to the current presentation by other means (<tt>ecdsa-p256-db</tt> does so via <tt>db_msg</tt> - see <xref target="ecdsa-db"/>).</t>
<t>Sub-proof transcripts use the BBS encoding primitives of <xref target="I-D.irtf-cfrg-bbs-signatures" sectionFormat="of" section="4.2.4.1"/>:</t>

<ul spacing="compact">
<li>BLS12-381 G1 points are serialized in their compressed form (48 octets)</li>
<li>scalars as 32-octet big-endian integers</li>
<li>integer lengths are encoded as <tt>I2OSP(int, 8)</tt></li>
</ul>
<t>A Verifier MUST reject a sub-proof carrying an encoded group element (in <tt>input</tt> or <tt>proof</tt>) that does not decode to a valid non-identity point of the G1 subgroup.</t>
<t>[Editor's Note: Decision needed: Need to define a serialization scheme for the Sigma proofs - Re-use the existing one from the <xref target="I-D.irtf-cfrg-sigma-protocols"/> (although it uses different encodings etc.), or define an optimized one for BLS12-381? Some of the following sub-proofs already propose very concrete choices to make the construction more concrete - all of these are open for discussion and will very likely see significant changes.]</t>

<section anchor="equality-proof"><name>Equality Proof Sub-Proof</name>

<dl spacing="compact">
<dt>Algorithm identifier:</dt>
<dd><tt>schnorr-eq</tt></dd>
</dl>
<t>The <tt>i</tt> field MUST be a single-element array <tt>[idx]</tt>.</t>
<t>Inputs (beyond the base sub-proof fields):</t>

<ul spacing="compact">
<li><tt>c_ext</tt>: a base64url-encoded BLS12-381 G1 point.</li>
</ul>
<t>The sub-proof attests that <tt>C_idx</tt> (from the core proof) and <tt>c_ext</tt> open to the same scalar under the generators <tt>(G, H)</tt> of <xref target="cipher-suite"/>. Cross-group equality is out of scope.</t>
<t>The construction is a 3-DL Schnorr discrete-logarithm-equality (DLEQ) proof over BLS12-381 G1 with <tt>(G, H)</tt>, with witness <tt>(m, s_1, s_2)</tt> such that:</t>

<artwork><![CDATA[C_idx = m * G + s_1 * H
c_ext = m * G + s_2 * H
]]>
</artwork>
<t>The Holder samples fresh random scalars <tt>(r_m, r_s1, r_s2)</tt> and computes Schnorr commitments <tt>T_1 = r_m * G + r_s1 * H</tt> and <tt>T_2 = r_m * G + r_s2 * H</tt>. The challenge is <tt>c = hash_to_scalar(transcript, challenge_dst)</tt> with <tt>challenge_dst = api_id || "SCHNORR_EQ_CHAL_"</tt> and <tt>hash_to_scalar</tt> the base BBS primitive of <xref target="cipher-suite"/>.</t>
<t>[Editor's Note: describe wire format of proof]</t>
</section>

<section anchor="ecdsa-db"><name>ECDSA Device-Binding Sub-Proof</name>
<t>This sub-proof MUST be present whenever <tt>kb = "ecdsa-p256-db"</tt> and MUST NOT be present otherwise. The algorithm identifier deliberately matches the <tt>kb</tt> value it verifies (see <xref target="device-binding-header"/>).</t>

<dl spacing="compact">
<dt>Algorithm identifier:</dt>
<dd><tt>ecdsa-p256-db</tt></dd>
</dl>
<t>The <tt>i</tt> field MUST be <tt>[0, 1, 2, 3]</tt>, naming the four indices that carry the device public-key limbs (see <xref target="device-binding-header"/>).</t>
<t>Inputs (beyond the base sub-proof fields): none.</t>
<t>The device-signed message is not transmitted, it is recomputed as:</t>

<artwork><![CDATA[db_msg = "JWP-BBS-DB-CHAL" || presentation_header_octets
]]>
</artwork>
<t>where <tt>"JWP-BBS-DB-CHAL"</tt> is the literal ASCII string. Binding <tt>db_msg</tt> to <tt>presentation_header_octets</tt> carries <tt>nonce</tt> and <tt>aud</tt> and is therefore sufficient for freshness.</t>
<t>The proof bytes encode a non-interactive zero-knowledge proof of knowledge of <tt>(dpk, (r, s))</tt> such that:</t>

<ol spacing="compact">
<li>The 4 commitments at the indices in <tt>i</tt> open to the 128-bit limbs of <tt>dpk</tt> (in the layout of <tt>kb</tt>) under <tt>(G, H)</tt> (see <xref target="cipher-suite"/>).</li>
<li><tt>(r, s)</tt> is a valid ECDSA P-256 signature on <tt>db_msg</tt> under <tt>dpk</tt>.</li>
</ol>
<t>[Editor's Note: TODO - This is currently a placeholder until we can reference a construction; expectation is that this will be described in another IETF draft]</t>
</section>

<section anchor="range-proof"><name>Range Proof Sub-Proof</name>

<dl spacing="compact">
<dt>Algorithm identifier:</dt>
<dd><tt>sigma-range</tt></dd>
</dl>
<t>The <tt>i</tt> field MUST be a single-element array <tt>[idx]</tt>.</t>
<t>Inputs (beyond the base sub-proof fields): bounds <tt>l</tt> and <tt>u</tt> as JSON integers. The sub-proof attests that m_idx, the message committed in the core proof at index <tt>idx</tt>, satisfies <tt>l &lt;= m_idx &lt; u</tt>.</t>
<t>[Editor's Note: describe/reference algorithm]</t>
</section>
</section>

<section anchor="presentation-verification"><name>Presentation Verification</name>
<t>The Verifier verifies a presentation by:</t>

<ol spacing="compact">
<li>Parsing the Presented Form and validating the <tt>cmap</tt> object of the Issuer Header per <xref target="claims-mapping"/>. Reject on violation.</li>
<li>Confirming that the Presentation Header <tt>alg</tt> equals the Issuer Header <tt>alg</tt>, that <tt>nonce</tt> matches the value the Verifier supplied for this presentation, and that <tt>aud</tt> identifies this Verifier.</li>
<li>Deriving the disclosed message scalars from the disclosed Presentation Payloads per <xref target="message-derivation"/>. For a <tt>scalar = true</tt> leaf, the payload MUST be the canonical decimal encoding of an integer in <tt>[0, r - 1]</tt> (see <xref target="scalar-encoding"/>). Reject otherwise.</li>
<li>Verifying the core proof with <tt>CoreProofVerify</tt> - see <xref target="core-proof"/>. Reject on failure. Confirming that the disclosed indices recovered from the proof are exactly the positions of the non-empty Presentation Payloads.</li>
<li>If <tt>kb</tt> is present in the Issuer Header, confirming that exactly one sub-proof with <tt>alg</tt> equal to the <tt>kb</tt> value is present. If <tt>kb</tt> is absent, confirming that no device-binding sub-proof is present.</li>
<li>Verifying every sub-proof per <xref target="sub-proofs"/>. Reject if any sub-proof fails to verify or carries an <tt>alg</tt> the Verifier does not support.</li>
</ol>
<t>Whether the disclosed claims and the predicates established by sub-proofs satisfy the Verifier's requirements is an application-level decision and out of scope for this document. After successful verification, the Verifier reconstructs the JSON payload per <xref target="reconstructed-payload"/>.</t>
</section>

<section anchor="example-presentation"><name>Example Presentation</name>
<t>Continuing the example of <xref target="example-issuer-header"/>, a Verifier requests <tt>family_name</tt> and asks the Holder to prove <tt>exp</tt> is in the future without disclosing it. The Presentation Header:</t>

<sourcecode type="json"><![CDATA[{
  "alg": "BBS-MOD",
  "nonce": "f4Oa3wT0r8m2Vn1pQ7sKdA",
  "aud": "https://verifier.example.com"
}
]]>
</sourcecode>
<t>The Holder marks index 1 (<tt>family_name</tt>) as <tt>DISCLOSE</tt>, index 10 (<tt>exp</tt>) as <tt>COMMIT</tt>, and the rest as <tt>HIDE</tt>. The core proof then carries a fresh Pedersen commitment to m_10. The Holder attaches a <tt>sigma-range</tt> sub-proof over index 10 proving <tt>now &lt;= exp &lt; 2^63</tt> (with <tt>now = 1779926400</tt>):</t>

<sourcecode type="json"><![CDATA[{
  "alg": "sigma-range",
  "input": { "i": [10], "l": 1779926400, "u": 9223372036854775808 },
  "proof": "..."
}
]]>
</sourcecode>
<t>The Compact Serialization concatenates with <tt>.</tt>: Presentation Header, Issuer Header, Presentation Payloads, Presentation Proof. The disclosed <tt>family_name</tt> at index 1 is the only populated payload and the other ten slots are empty:</t>

<artwork><![CDATA[<base64url(Presentation Header)>
.
<base64url(Issuer Header)>
.
~Ik11c3Rlcm1hbm4i~~~~~~~~~
.
<core proof>~<sigma-range sub-proof>
]]>
</artwork>
<t>The Verifier verifies the core proof, recovers <tt>C_10</tt>, and checks the sub-proof against it. It learns <tt>family_name</tt> and that the credential has not expired.</t>
</section>

<section anchor="reconstructed-payload"><name>Reconstructed JSON Payload</name>
<t>After verifying the core proof and any sub-proofs, the Verifier SHOULD convey to the application a JSON object reconstructed from the disclosed information, analogous to the Processed SD-JWT Payload of <xref target="RFC9901"/>. Reconstruction presupposes that <tt>cmap</tt> passed the validation of <xref target="claims-mapping"/> - a presentation whose <tt>cmap</tt> object fails it MUST be rejected, not reconstructed. The procedure:</t>

<ol spacing="compact">
<li>Start from <tt>{ "vct": &lt;vct from Issuer Header&gt; }</tt>.</li>
<li>Walk <tt>cmap</tt>. For each leaf at a disclosed index <tt>i</tt>, first compare the Presentation Payload octets to the decoy octets for that leaf's <tt>scalar</tt> flag (see <xref target="decoys"/>) - on a byte-equal match, omit the leaf. Otherwise set the leaf's value by parsing the payload octets as a single JSON text <xref target="RFC8259"/> when <tt>scalar</tt> is <tt>false</tt>, or as the integer they denote (see <xref target="scalar-encoding"/>) when <tt>scalar</tt> is <tt>true</tt>. A presentation containing a disclosed payload that fails to parse MUST be rejected. Hidden and committed-but-not-disclosed leaves are omitted.</li>
<li>Preserve the object and array structure of <tt>cmap</tt> for surviving leaves. Array entries that were omitted do not appear, so reconstructed array indices may differ from those in the <tt>cmap</tt> annotations.</li>
</ol>
<t>Predicates established by sub-proofs are not represented as leaf values. The reconstruction procedure MUST NOT populate values for hidden or committed-but-not-disclosed leaves.</t>
<t>For <xref target="example-presentation"/>, the reconstructed payload is:</t>

<sourcecode type="json"><![CDATA[{
  "vct": "https://credentials.example.com/identity_credential",
  "family_name": "Mustermann"
}
]]>
</sourcecode>
</section>
</section>

<section anchor="cipher-suite"><name>Cipher Suite</name>
<t>This profile fixes exactly one cipher suite, so that <tt>alg</tt> does not vary across a credential population and split its anonymity set (see <xref target="anonymity"/>).</t>

<section anchor="identifier"><name>Identifier</name>
<t>JPA Algorithm JSON Label: <tt>BBS-MOD</tt>.</t>
<t>Cipher suite identifier (also used as <tt>api_id</tt> for hash-to-scalar, generator derivation, and sub-proof domain separation):</t>

<artwork><![CDATA[BBS-MOD_BLS12381G1_XMD:SHA-256_SSWU_RO_BLIND_H2G_HM2S_
]]>
</artwork>
<t>The <tt>BBS-MOD_</tt> prefix separates this profile from both the base BBS JPA (<tt>BBS</tt> of <xref target="I-D.ietf-jose-json-proof-algorithms" sectionFormat="of" section="9.1.2.4"/>) and the base blind BBS Interface (<tt>BBS_BLS12381G1_XMD:SHA-256_SSWU_RO_BLIND_H2G_HM2S_</tt>). This profile invokes the core proof operations of that Interface directly to expose committed-message proofs - see <xref target="core-proof"/>. It also bypasses hash-to-scalar on a per-message basis under the <tt>scalar</tt> flag and attaches sub-proofs as described in <xref target="sub-proofs"/>.</t>
</section>

<section anchor="parameters"><name>Parameters</name>

<ul spacing="compact">
<li><strong>Curve / group</strong>: BLS12-381, G1 subgroup.</li>
<li><strong>BBS ciphersuite</strong>: <tt>BBS-MOD_BLS12381G1_XMD:SHA-256_SSWU_RO_</tt> - identical to <tt>BLS12-381-SHA-256</tt> (<xref target="I-D.irtf-cfrg-bbs-signatures" sectionFormat="of" section="7.2.2"/>, with hash-to-curve SHA-256 SSWU random oracle <xref target="RFC9380"/>) in all parameters except the ciphersuite identifier.</li>
<li><strong>Hash-to-scalar</strong>: as in the underlying BBS ciphersuite, with domain separation derived from <tt>api_id</tt>.</li>
<li><strong>Core proof operations</strong>: <tt>CoreProofGen</tt> / <tt>CoreProofVerify</tt> of <xref target="I-D.irtf-cfrg-bbs-blind-signatures"/> invoked directly (not via <tt>BlindProofGen</tt>), so implementations MUST apply the <tt>commits_indexes</tt> and <tt>disclosed_indexes</tt> checks of <tt>CoreProofGen</tt>.</li>
<li><strong>Pedersen commitment generators</strong>: <tt>(G, H) = (Y_1, Y_0)</tt> where <tt>(Y_0, Y_1) = BBS.create_generators(2, "COM_DIS_" || api_id)</tt>. Every committed-index commitment has the form <tt>C_i = m_i * G + s_i * H</tt> with <tt>s_i</tt> sampled per presentation by <tt>CoreProofGen</tt>.</li>
<li><strong>Per-message hash-to-scalar bypass</strong>: governed by each leaf's <tt>scalar</tt> flag (see <xref target="claims-mapping"/>).</li>
</ul>
<t>The <tt>api_id</tt> above follows the Interface identifier rule of <xref target="I-D.irtf-cfrg-bbs-blind-signatures" sectionFormat="of" section="4.2"/> - <tt>ciphersuite_id || "BLIND_H2G_HM2S_"</tt> - applied to the ciphersuite identifier <tt>BBS-MOD_BLS12381G1_XMD:SHA-256_SSWU_RO_</tt>. All BBS operations used by this document are the Blind BBS Interface operations, or the core operations they wrap, which that document parameterizes with this <tt>api_id</tt>.</t>
</section>
</section>

<section anchor="security-considerations"><name>Security Considerations</name>

<section anchor="random"><name>Random Number Generation</name>
<t>All randomness used by this document MUST be generated using a cryptographically secure random number generator. Reuse or predictability of a blinding scalar or proof nonce can break unlinkability or soundness, or even leak the signing key.</t>
</section>

<section anchor="hash-to-scalar-bypass"><name>Hash-to-Scalar Bypass</name>
<t>[Editor's Note: TODO - Check what exactly the attack scenarios are / if there are some]</t>
</section>

<section anchor="replay-and-presentation-freshness"><name>Replay and Presentation Freshness</name>
<t>Freshness relies on the Verifier-supplied <tt>nonce</tt>. Verifiers MUST generate nonces with enough entropy to make them unpredictable and MUST NOT accept a presentation carrying a <tt>nonce</tt> they did not supply for that transaction - see <xref target="presentation-verification"/>. The <tt>aud</tt> binding limits a captured presentation to its intended Verifier.</t>
</section>

<section anchor="holder-binding"><name>Holder Binding</name>
<t>Without device binding (<tt>kb</tt> absent), possession of the Issued Form is sufficient to derive presentations, so anyone who obtains the credential in its Issued Form can present it. Deployments that need resistance against credential theft or pooling SHOULD use device binding - see <xref target="device-binding-header"/>.</t>
</section>

<section anchor="unlinkability-scope"><name>Unlinkability Scope</name>
<t>The core proof hides everything except the disclosed messages, the carried commitments, and the sub-proof predicates. Parties observing multiple presentations (including colluding Verifiers, or an Issuer colluding with a Verifier) can still correlate them through disclosed attribute values, sub-proof predicate parameters, or transport-level metadata.</t>
</section>
</section>

<section anchor="privacy-considerations"><name>Privacy Considerations</name>

<section anchor="anonymity"><name>Issuer Header Correlation</name>
<t>The Issuer Header is sent in clear to the Verifier. Any variation in it across Holders of the same <tt>vct</tt> narrows the anonymity set.</t>
<t>Implementations SHOULD make the Issuer Header byte-identical across the entire population of a <tt>vct</tt>, by:</t>

<ul spacing="compact">
<li>Fixing the <tt>cmap</tt> layout (including all optional attributes and maximum-length array slots) with a constant serialization.</li>
<li>Filling unused slots with decoys per <xref target="decoys"/>.</li>
<li>Carrying per-credential metadata (issuance time, expiry, identifiers) as messages in the message vector</li>
</ul>
</section>

<section anchor="cipher-suite-and-algorithm-identifiers"><name>Cipher Suite and Algorithm Identifiers</name>
<t><tt>alg</tt> and <tt>kb</tt> likewise split the anonymity set when they vary across the population of a <tt>vct</tt>. Implementations SHOULD use a single <tt>alg</tt> and a single <tt>kb</tt> value (or omit <tt>kb</tt> entirely) across all credentials of a <tt>vct</tt>, and SHOULD NOT mix device-bound and non-device-bound credentials under the same <tt>vct</tt>.</t>
</section>
</section>

<section anchor="iana"><name>IANA Considerations</name>
<t>This document requests the following registrations and registry creations.</t>

<section anchor="jpa-alg-value"><name>JPA <tt>alg</tt> Value</name>
<t>IANA is requested to register the following JSON Proof Algorithm in the "JSON Web Proof Algorithms" registry established by <xref target="I-D.ietf-jose-json-proof-algorithms"/>:</t>

<ul spacing="compact">
<li>Algorithm Name: BBS-MOD using SHA-256</li>
<li>Algorithm JSON Label: <tt>BBS-MOD</tt></li>
<li>Algorithm CBOR Label: TBD (requested assignment 11)</li>
<li>Algorithm Description: Blind BBS over BLS12-381 with <tt>CoreProofGen</tt>-based committed-message proofs, the per-message <tt>scalar</tt> flag, and the sub-proof attachment mechanism of <xref target="sub-proofs"/>. Cipher suite identifier <tt>BBS-MOD_BLS12381G1_XMD:SHA-256_SSWU_RO_BLIND_H2G_HM2S_</tt>.</li>
<li>Algorithm Usage Location(s): Issued, Presented</li>
<li>JWP Implementation Requirements: Optional</li>
<li>Change Controller: IETF</li>
<li>Specification Document(s): <xref target="cipher-suite"/> of this document.</li>
<li>Algorithm Analysis Document(s): <xref target="LSZ25"/>, <xref target="CT25"/></li>
</ul>
</section>

<section anchor="header-parameter-registrations"><name>Header Parameter Registrations</name>
<t>IANA is requested to register the following Header Parameters in the "JSON Web Proof Header Parameters" registry established by <xref target="I-D.ietf-jose-json-web-proof"/>:</t>

<ul>
<li>Header Parameter Name: Claims Mapping</li>
<li>Header Parameter JSON Label: <tt>cmap</tt></li>
<li>Header Parameter CBOR Label: TBD (requested assignment 11)</li>
<li>Header Parameter Usage Location(s): Issued</li>
<li>Change Controller: IETF</li>
<li><t>Specification Document(s): <xref target="claims-mapping"/> of this document.</t>
</li>
<li><t>Header Parameter Name: Device Key Binding</t>
</li>
<li><t>Header Parameter JSON Label: <tt>kb</tt></t>
</li>
<li><t>Header Parameter CBOR Label: TBD (requested assignment 12)</t>
</li>
<li><t>Header Parameter Usage Location(s): Issued</t>
</li>
<li><t>Change Controller: IETF</t>
</li>
<li><t>Specification Document(s): <xref target="device-binding-header"/> of this document.</t>
</li>
</ul>
</section>

<section anchor="sub-proof-algorithms-registry"><name>Sub-Proof Algorithms Registry</name>
<t>IANA is requested to create a new "Sub-Proof Algorithms" registry.</t>
<t>Allocation policy: Specification Required (<xref target="RFC8126"/>). Designated experts SHOULD verify that each entry pins its underlying group, generators, transcript hash, and Fiat-Shamir domain separation, and that the sub-proof is bound to a commitment attested by the core proof per <xref target="sub-proofs"/>. For entries with Device Binding set to <tt>yes</tt>, they SHOULD additionally verify that the reference defines the reserved slot count <tt>N</tt> and the device-key encoding required by <xref target="device-binding-header"/>.</t>
<t>Registry fields: Identifier (the <tt>alg</tt> value of a sub-proof object), Description, Device Binding (whether the identifier is also a valid <tt>kb</tt> value - see <xref target="device-binding-header"/>), Reference, Change Controller.</t>
<t>Initial entries:</t>

<ul>
<li>Identifier: <tt>ecdsa-p256-db</tt></li>
<li>Description: ECDSA P-256 device-binding sub-proof.</li>
<li>Device Binding: yes.</li>
<li>Reference: This document, <xref target="ecdsa-db"/>.</li>
<li><t>Change Controller: IETF.</t>
</li>
<li><t>Identifier: <tt>sigma-range</tt></t>
</li>
<li><t>Description: Sigma-protocol range proof over a committed scalar message.</t>
</li>
<li><t>Device Binding: no.</t>
</li>
<li><t>Reference: This document, <xref target="range-proof"/>.</t>
</li>
<li><t>Change Controller: IETF.</t>
</li>
<li><t>Identifier: <tt>schnorr-eq</tt></t>
</li>
<li><t>Description: Schnorr proof of equality between a committed message and an external commitment.</t>
</li>
<li><t>Device Binding: no.</t>
</li>
<li><t>Reference: This document, <xref target="equality-proof"/>.</t>
</li>
<li><t>Change Controller: IETF.</t>
</li>
</ul>
</section>
</section>

</middle>

<back>
<references><name>References</name>
<references><name>Normative References</name>
<reference anchor="FIPS186-5" target="https://doi.org/10.6028/NIST.FIPS.186-5">
  <front>
    <title>Digital Signature Standard (DSS)</title>
    <author>
      <organization>National Institute of Standards and Technology</organization>
    </author>
    <date year="2023" month="February"/>
  </front>
  <seriesInfo name="FIPS PUB" value="186-5"/>
  <seriesInfo name="DOI" value="10.6028/NIST.FIPS.186-5"/>
</reference>
<xi:include href="https://bib.ietf.org/public/rfc/bibxml3/reference.I-D.ietf-jose-json-proof-algorithms.xml"/>
<xi:include href="https://bib.ietf.org/public/rfc/bibxml3/reference.I-D.ietf-jose-json-web-proof.xml"/>
<xi:include href="https://bib.ietf.org/public/rfc/bibxml3/reference.I-D.ietf-oauth-sd-jwt-vc.xml"/>
<xi:include href="https://bib.ietf.org/public/rfc/bibxml3/reference.I-D.irtf-cfrg-bbs-blind-signatures.xml"/>
<xi:include href="https://bib.ietf.org/public/rfc/bibxml3/reference.I-D.irtf-cfrg-bbs-signatures.xml"/>
<xi:include href="https://bib.ietf.org/public/rfc/bibxml/reference.RFC.2119.xml"/>
<xi:include href="https://bib.ietf.org/public/rfc/bibxml/reference.RFC.4648.xml"/>
<xi:include href="https://bib.ietf.org/public/rfc/bibxml/reference.RFC.7519.xml"/>
<xi:include href="https://bib.ietf.org/public/rfc/bibxml/reference.RFC.8126.xml"/>
<xi:include href="https://bib.ietf.org/public/rfc/bibxml/reference.RFC.8174.xml"/>
<xi:include href="https://bib.ietf.org/public/rfc/bibxml/reference.RFC.8259.xml"/>
<xi:include href="https://bib.ietf.org/public/rfc/bibxml/reference.RFC.9380.xml"/>
</references>
<references><name>Informative References</name>
<reference anchor="CT25" target="https://eprint.iacr.org/2025/1093">
  <front>
    <title>On the Concrete Security of BBS/BBS+ Signatures</title>
    <author initials="R." surname="Chairattana-Apirom"/>
    <author initials="S." surname="Tessaro"/>
    <date year="2025"/>
  </front>
  <seriesInfo name="IACR ePrint" value="2025/1093"/>
</reference>
<xi:include href="https://bib.ietf.org/public/rfc/bibxml3/reference.I-D.irtf-cfrg-sigma-protocols.xml"/>
<reference anchor="LSZ25" target="https://eprint.iacr.org/2025/1981">
  <front>
    <title>Vision: A Modular Framework for Anonymous Credential Systems</title>
    <author initials="A." surname="Lehmann"/>
    <author initials="A." surname="Sidorenko"/>
    <author initials="A." surname="Zacharakis"/>
    <date year="2025"/>
  </front>
  <seriesInfo name="IACR ePrint" value="2025/1981"/>
</reference>
<xi:include href="https://bib.ietf.org/public/rfc/bibxml/reference.RFC.7493.xml"/>
<xi:include href="https://bib.ietf.org/public/rfc/bibxml/reference.RFC.9901.xml"/>
<reference anchor="TS14" target="https://github.com/eu-digital-identity-wallet/eudi-doc-standards-and-technical-specifications/blob/main/docs/technical-specifications/ts14-zkps-from-mms.md">
  <front>
    <title>Specification for the implementation of Zero-Knowledge Proofs based on multi-message signatures in the EUDI Wallet (TS-14)</title>
    <author>
      <organization>European Commission, EUDI Wallet Expert Group</organization>
    </author>
    <date year="2025"/>
  </front>
  <seriesInfo name="EUDI" value="TS-14"/>
  <refcontent>Work in Progress.</refcontent>
</reference>
</references>
</references>

<section anchor="acknowledgments"><name>Acknowledgments</name>
<t>This document rests on the work captured in <xref target="TS14"/> by the EUDI Wallet expert group. The committed-message core proof builds on <xref target="I-D.irtf-cfrg-bbs-blind-signatures"/>, and the modular committed-disclosure framework draws on <xref target="LSZ25"/>.</t>
</section>

<section anchor="document-history"><name>Document History</name>
<t>[[ pre Working Group Adoption: ]]</t>
<t>-02</t>

<ul spacing="compact">
<li>rename the <tt>claims</tt> Header Parameter to <tt>cmap</tt> to avoid the JPT <tt>claims</tt> parameter</li>
<li>add <tt>alg</tt> to the Presentation Header</li>
<li>add more security considerations</li>
<li>align IANA registrations with the registry templates (usage locations, requested CBOR labels)</li>
<li>mandate JSON encoding for <tt>scalar = false</tt> payloads</li>
<li>require <tt>claims</tt> validation</li>
<li>add Presentation Verification section</li>
<li>pin issuance signing to <tt>CoreSign</tt></li>
<li>define <tt>kb</tt> registration via a Device Binding registry field</li>
<li>define canonical decimal encoding</li>
<li>remove sigma-range construction, trim and reorder sub-proofs</li>
<li>align cipher suite text with latest blind BBS draft</li>
</ul>
<t>-01</t>

<ul spacing="compact">
<li>Fix venue note (was showing JOSE, should've been empty)</li>
<li>Add ASCII Art overview</li>
<li>some tweaks for sub-proof text</li>
<li>update range proof reference to adopted draft-ietf-privacypass-arc-crypto (section moved to 5.4)</li>
<li>add missing generators input to the CoreProofGen and CoreProofVerify descriptions</li>
<li>align header parameter registrations with the JWP registry template</li>
</ul>
<t>-00</t>

<ul spacing="compact">
<li>Initial Version</li>
</ul>
</section>

</back>

</rfc>
