<?xml version="1.0" encoding="UTF-8"?>
  <?xml-stylesheet type="text/xsl" href="rfc2629.xslt" ?>
  <!-- generated by https://github.com/cabo/kramdown-rfc version 1.7.43 (Ruby 2.6.10) -->


<!DOCTYPE rfc  [
  <!ENTITY nbsp    "&#160;">
  <!ENTITY zwsp   "&#8203;">
  <!ENTITY nbhy   "&#8209;">
  <!ENTITY wj     "&#8288;">

]>


<rfc ipr="trust200902" docName="draft-oga-musubi-settlement-03" category="info" submissionType="IETF" tocInclude="true" sortRefs="true" symRefs="true">
  <front>
    <title abbrev="MUSUBI Settlement">MUSUBI Settlement: Recomputing Whether an Agent Acted Within a Signed Grant</title>

    <author fullname="Toshikatsu Oga">
      <organization>The HORIZONs Co., Ltd.</organization>
      <address>
        <postal>
          <country>Japan</country>
        </postal>
        <email>thehorizon.nnovation@icloud.com</email>
      </address>
    </author>

    <date year="2026" month="November" day="16"/>

    
    
    <keyword>settlement</keyword> <keyword>agents</keyword> <keyword>Ed25519</keyword> <keyword>Bitcoin anchoring</keyword> <keyword>canonical JSON</keyword>

    <abstract>


<?line 233?>

<t>MUSUBI settlement is a deterministic function that takes a contract
signed by two parties, a set of signed records of what an agent did
under that contract, and a view of Bitcoin block headers, and
computes whether the recorded acts stayed within the grant the
contract gives. Anyone holding the same bytes can recompute the
result without trusting the party that produced it. This document
specifies one well defined part of that function, settle core v0,
precisely enough that a second implementation can be written from
the text and the published test vectors alone. Outside settle core
v0, it also describes the admission record of version 0.4, in which a
relying party states the check it made of a request's content by
digests and marks, and how settle v1.19 reads it. Settlement is not
runtime enforcement and is not a judgement of liability.</t>



    </abstract>



  </front>

  <middle>


<?line 249?>

<section anchor="intro"><name>Introduction</name>

<section anchor="what"><name>What Settlement Is</name>

<t>A principal and a contractor sign a contract. The contract carries a
grant: the actions the contractor is authorized to take, the actions
it is prohibited from taking, the actions that need a separate
approval from the principal, the parties to whom work may be
delegated, an expiry height and a revocation policy. While the
contractor (or a witness named in the grant) works, it signs records
of what it did, and each record is committed to the Bitcoin block
chain <xref target="BITCOIN"/> through an anchor. Later, anyone who holds the
contract, the records and a view of the relevant block headers can
settle: walk the records in the order the block chain gives them,
compare each act with the grant, and produce a verdict, a status and
a bond outcome.</t>

<t>In the terms of the agent auditing architecture
<xref target="I-D.kuehlewind-audit-architecture"/>, settlement is an Auditor's
evaluation of whether recorded actions stayed within the authorization
in force at the time; <xref target="related-audit"/> maps the two.</t>

<t>The record schema names carry the prefix "a2a" because the format was
designed for agents that interact over the Agent2Agent (A2A) protocol
<xref target="A2A"/>. Settlement reads no A2A message and does not depend on A2A.</t>

</section>
<section anchor="why"><name>Why a Second Implementation Needs a Text</name>

<t>Until this document, the MUSUBI settle function was defined only by
Python code (settle_v1.py to settle_v1_16.py, each layer importing the
ones below it) <xref target="MUSUBI-REF"/>. A function that only one program can
compute is a tool, not a protocol. This document states one well
defined part of it, settle core v0, in plain language. A stranger
should be able to write a second implementation from this text and
the vectors alone, without reading the reference.</t>

<t>The text was written from the reference code, not from intent.
Writing it against settle_v1_10.py found places where the code
accepted input a careful second implementation would read
differently, and one place where a record nobody signed could change
the verdict (items N1 to N11 of <xref target="MUSUBI-NOTES"/>). settle_v1_16.py was
written to close them, and this document describes it.
<xref target="written-behaviour"/> lists what changed. Where the reference still
does something surprising, this document says so and marks it
WRITTEN BEHAVIOUR. Written behaviour is still normative for core v0:
a second implementation that differs on it will disagree with the
vectors.</t>

<t>This document is an adaptation of <xref target="MUSUBI-SETTLE-CORE"/> into the
Internet-Draft format. Where the two differ in meaning, that is an
error in this document. This version follows revision 0.13 of that
text. Revisions 0.1 to 0.3 answered questions and implementation notes
from the first separately written implementation
(<xref target="impl-others"/>); revision 0.4 checked which underspecified
reasons the vectors actually produce (<xref target="exercise"/>); revisions 0.5 to
0.8 corrected statements about the reference, stated two sentences a
third implementation asked for, stated a consequence of the unsigned
anchor, and named the commit of the reference's file list. No rule of
the reference and no vector changed in any of them. <xref target="clarifications"/>
lists each place. Revision 0.9 fixed one place where the reference did
not follow the text, and revision 0.11 stated two readings the text
had left open; no vector changed (<xref target="revisions-09-11"/>). Revisions
0.10, 0.12 and 0.13 added rules outside core v0 that read admission
records; <xref target="admission"/> describes them.</t>

</section>
<section anchor="not"><name>What Settlement Is Not</name>

<t>Settlement is not runtime enforcement. It never stops an agent from
acting; it reads records after the fact and says whether the recorded
acts were within the grant.</t>

<t>Settlement is not a judgement of liability or fault, and it does not
make the contract a legal contract. The verdict is a function anyone
can recompute from the same bytes. <xref target="not-established"/> lists what the
verdict does not establish.</t>

</section>
<section anchor="scope"><name>Scope of Core v0</name>

<t>Settle core v0 is exactly what the reference computes when the
contract carries none of the following:</t>

<t><list style="symbols">
  <t><spanx style="verb">requirements.spine</spanx> (the spine gate, settle v1.8),</t>
  <t><spanx style="verb">requirements.convergence</spanx> (drawn measurers, settle v1.9),</t>
  <t><spanx style="verb">grant.approval_policy.approvers</spanx> (pinned approvers, settle v1.10),</t>
  <t><spanx style="verb">grant.limits</spanx> (which forces <spanx style="verb">requirements.admission</spanx>),</t>
  <t><spanx style="verb">requirements.admission</spanx> (settle v1.11 to v1.14 read it; the
admissions input is passed through unchanged).</t>
</list></t>

<t>It is computed in binding mode "strict" over exactly one header view.</t>

<t>Out of scope, and not defined here:</t>

<t><list style="symbols">
  <t>choosing between several header views (fork choice by work above
the checkpoint),</t>
  <t>binding mode "legacy" (records bound by contract_id only),</t>
  <t>fetching anything. Settle core v0 never touches the network. Every
key it uses is inside the contract.</t>
</list></t>

<t>A key is carried when it is present with a value other than null,
with two exceptions. <spanx style="verb">requirements.spine</spanx> and
<spanx style="verb">requirements.convergence</spanx> are carried whenever the key is present,
null included: their gates then answer <spanx style="verb">spine_requirement_incomplete</spanx>
and <spanx style="verb">convergence_requirement_incomplete</spanx>. And
<spanx style="verb">grant.approval_policy.approvers</spanx> is carried only when it is a well
formed approver list (rule 10 of <xref target="contract-door"/>); any other value
that is present, null and <spanx style="verb">[]</spanx> included, refuses the contract in
<xref target="contract-door"/> (WRITTEN BEHAVIOUR). Vectors <spanx style="verb">synth_limits_null</spanx>,
<spanx style="verb">synth_admission_null</spanx>, <spanx style="verb">synth_approvers_object</spanx>,
<spanx style="verb">synth_approvers_string</spanx>, <spanx style="verb">synth_approvers_zero</spanx>.</t>

<t>If a contract carries any of the keys above, this document does not
define the result. An implementation MUST NOT claim core v0
conformance for such an input. <xref target="admission"/> describes, outside core
v0, the rules settle v1.17 to v1.19 add for a contract that requires
admission.</t>

</section>
</section>
<section anchor="conventions"><name>Conventions and Terminology</name>

<section anchor="bcp14"><name>Requirements Language</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>

<t>Most rules in this document are stated in the indicative ("is",
"gives", "adds"). They are normative all the same: they determine the
core fields (<xref target="core-fields"/>) that a second implementation MUST
reproduce exactly.</t>

</section>
<section anchor="terms"><name>Terminology</name>

<dl>
  <dt>contract:</dt>
  <dd>
    <t>A JSON object with schema <spanx style="verb">a2a-contract-v0</spanx>, signed by a principal
and a contractor, that carries the grant.</t>
  </dd>
  <dt>principal, contractor:</dt>
  <dd>
    <t>The two parties to a contract, identified by their <spanx style="verb">role</spanx> and holding
the Ed25519 <xref target="RFC8032"/> keys pinned in the contract.</t>
  </dd>
  <dt>witness:</dt>
  <dd>
    <t>A key named in <spanx style="verb">grant.witnesses</spanx> that may sign execution records.</t>
  </dd>
  <dt>record:</dt>
  <dd>
    <t>An execution (<spanx style="verb">a2a-execution-v0</spanx>), a revocation
(<spanx style="verb">a2a-revocation-v0</spanx>) or a revocation acknowledgement
(<spanx style="verb">a2a-revocation-ack-v0</spanx>).</t>
  </dd>
  <dt>header view:</dt>
  <dd>
    <t>A list of raw 80 byte Bitcoin block headers <xref target="BITCOIN-HEADERS"/> with
their heights.</t>
  </dd>
  <dt>walk:</dt>
  <dd>
    <t>A NENRIN walk record <xref target="NENRIN-VERIFIER"/> cited by an execution
through <spanx style="verb">nenrin_ref</spanx>.</t>
  </dd>
  <dt>the reference:</dt>
  <dd>
    <t>settle_v1_16.py and the files it imports, at the versions listed in
<xref target="impl-status"/> <xref target="MUSUBI-REF"/>.</t>
  </dd>
  <dt>WRITTEN BEHAVIOUR:</dt>
  <dd>
    <t>A behaviour of the reference that may surprise a reader, kept
because the vectors pin it, or would.</t>
  </dd>
  <dt>integer:</dt>
  <dd>
    <t>A JSON value written as an integer token: no fraction, no exponent.
<spanx style="verb">true</spanx> and <spanx style="verb">false</spanx> are never integers.</t>
  </dd>
</dl>

<t>The following notation is used for byte strings. A string in double
quotes, such as <spanx style="verb">"a2a-contract-v0\n"</spanx>, denotes the ASCII bytes between
the quotes, where <spanx style="verb">\n</spanx> stands for one line feed (0x0A). When A and B
are byte strings, <spanx style="verb">A + B</spanx> is their concatenation. <spanx style="verb">sha256(x)</spanx> is SHA-256
<xref target="FIPS180-4"/> of the bytes x. Byte ranges such as "bytes 4 to 36" are
zero-based and exclude the upper bound, so they name the 32 bytes at
offsets 4 through 35. "Lowercase hex" means the characters <spanx style="verb">0</spanx> to <spanx style="verb">9</spanx>
and <spanx style="verb">a</spanx> to <spanx style="verb">f</spanx>, two per byte.</t>

</section>
</section>
<section anchor="shape"><name>Inputs, Outputs and the Shape of the Computation</name>

<section anchor="inputs"><name>Inputs</name>

<t><list style="numbers" type="1">
  <t><spanx style="verb">contract</spanx>: one JSON object (schema <spanx style="verb">a2a-contract-v0</spanx>).</t>
  <t><spanx style="verb">events</spanx>: a list of JSON values. Each is meant to be a record: an
execution (<spanx style="verb">a2a-execution-v0</spanx>), a revocation (<spanx style="verb">a2a-revocation-v0</spanx>)
or a revocation acknowledgement (<spanx style="verb">a2a-revocation-ack-v0</spanx>). Anything
else is ignored. Order carries no meaning.</t>
  <t><spanx style="verb">view</spanx>: one JSON object
<spanx style="verb">{"headers": [{"height": H, "hex": "&lt;160 lowercase hex&gt;"}, ...]}</spanx>,
raw 80 byte Bitcoin block headers. Other keys are ignored.</t>
  <t><spanx style="verb">nenrin_records</spanx>: optional list of JSON values (walk records cited
by executions). Absent and empty mean the same thing. A value that
is not a list is treated as absent.</t>
  <t><spanx style="verb">parent</spanx>: optional, the parent contract, needed only when
<spanx style="verb">contract.parent_contract</spanx> is present (<xref target="strict-terms"/>).</t>
</list></t>

<t>If <spanx style="verb">contract</spanx> is not an object, the settlement is underspecified with
the single reason <spanx style="verb">contract_not_verified</spanx>. If <spanx style="verb">events</spanx> is not a list,
it is underspecified with <spanx style="verb">input_not_settleable</spanx>. Nothing else is
computed in either case.</t>

<t>Keys are not a separate input. The principal's and contractor's
Ed25519 public keys are <spanx style="verb">parties[i].public_key_ed25519_b64</spanx> inside the
contract, and witness keys are <spanx style="verb">grant.witnesses[i].public_key_ed25519_b64</spanx>.
<spanx style="verb">key_url</spanx> is never fetched. Checking that the key served at <spanx style="verb">key_url</spanx>
equals the pinned key is outside settle core.</t>

</section>
<section anchor="reading-json"><name>Reading JSON</name>

<t>Inputs are parsed as JSON <xref target="RFC8259"/> with one extra rule: an object
with a duplicate key is an error and no settlement is produced. Text
nested deeper than 512 levels of <spanx style="verb">[</spanx> or <spanx style="verb">{</spanx> (counted outside strings)
is also an error.</t>

<t>Numbers MUST be integers within plus or minus (2^53 - 1). The contract
is checked for this in <xref target="contract-door"/>, records in <xref target="triage"/>, the
view in <xref target="view-verify"/> and walks in <xref target="nenrin"/>. A JSON value with a
non-integer number is not "strict" in the sense of <xref target="triage"/>.</t>

<t>The tokens <spanx style="verb">NaN</spanx>, <spanx style="verb">Infinity</spanx> and <spanx style="verb">-Infinity</spanx> are read as numbers that
are not integers (WRITTEN BEHAVIOUR of the reference's reader,
Python's json module). A value holding one is therefore not strict. It
is not a parse error. Vectors <spanx style="verb">synth_nan_in_signed_body</spanx>,
<spanx style="verb">synth_infinity_in_signed_body</spanx>,
<spanx style="verb">synth_negative_infinity_in_signed_body</spanx>. Exactly these three
spellings are read, case sensitive. <spanx style="verb">nan</spanx>, <spanx style="verb">+Infinity</spanx> or <spanx style="verb">infinity</spanx>
is a parse error, and no settlement is produced. A number token whose
value is too large for a double (for example <spanx style="verb">1e400</spanx>) is read as an
infinity, so it too is a non-integer number.</t>

<t>An integer token longer than 4300 digits is an error, and no
settlement is produced: the reader raises (WRITTEN BEHAVIOUR). There
is no vector, because there is no settlement to compare.</t>

<t>A JSON value is an "integer" in this document only if it is written
as an integer token: no fraction, no exponent. <spanx style="verb">true</spanx> and <spanx style="verb">false</spanx> are
never integers.</t>

<t>These rules are close to, but not the same as, I-JSON <xref target="RFC7493"/>.</t>

</section>
<section anchor="computation"><name>The Computation in One Paragraph</name>

<t>Compute contract_sha256 (<xref target="digests"/>) and bind records by it
(<xref target="binding"/>). Triage every bound record (<xref target="triage"/>), whatever
<xref target="contract"/> and <xref target="view"/> find: records not strictly signed by a party
are rejected or set aside here with their dispositions
(<xref target="dispositions"/>), and no later step sees them. Check the contract's
terms (<xref target="contract"/>) and verify the header view (<xref target="view"/>). Every
problem found so far is collected; none of them stops the other
checks in this paragraph. If rules 1 to 8 of <xref target="view-verify"/> passed,
collapse copies of one signed body (<xref target="collapse"/>), whether or not
<xref target="contract"/> found a problem. If a problem was found in
<xref target="contract-door"/>, <xref target="settleable-terms"/> or rules 1 to 8 of
<xref target="view-verify"/>, no record is examined further: the dispositions given
by <xref target="triage"/> and <xref target="collapse"/> stand, the <spanx style="verb">revoked_by</spanx> test of
<xref target="authentication"/>, <xref target="conformance"/> and <xref target="anchors"/> do not run, the
event set is empty, and a bound record kept by <xref target="triage"/> and not
dropped by <xref target="collapse"/> has no disposition (revision 0.9). Otherwise:
check conformance (<xref target="conformance"/>), check anchors (<xref target="anchors"/>). Records that pass are
the event set, ordered by anchor height then record sha256
(<xref target="event-set"/>). Walk the event set against the grant (<xref target="grant"/>), add
the NENRIN checks (<xref target="nenrin"/>), and decide the verdict (<xref target="verdict"/>),
status (<xref target="status"/>) and bond outcome (<xref target="bond-outcome"/>). The checks of
<xref target="strict-terms"/>, rule 9 of <xref target="view-verify"/>, <xref target="triage"/> (a record that
fails closed), <xref target="anchor-block"/> (below the lower bound) and <xref target="nenrin"/>
(a cited walk) are collected in the same way and make the settlement
underspecified; they never stop the computation of the event set.</t>

</section>
<section anchor="output-overview"><name>Output</name>

<t>The output is one JSON object. <xref target="output"/> lists the core fields a
second implementation MUST reproduce exactly. Everything else the
reference writes is informative.</t>

</section>
</section>
<section anchor="crypto"><name>Canonical Form, Digests and Signatures</name>

<section anchor="canonical"><name>Canonical JSON (musubi-canonical-v0)</name>

<t><spanx style="verb">canonical(v)</spanx> is the UTF-8 <xref target="RFC3629"/> encoding of this text:</t>

<t><list style="numbers" type="1">
  <t>Objects: keys sorted by Unicode code point, at every level.
Written <spanx style="verb">{"k":v,...}</spanx> with no whitespace.</t>
  <t>Arrays: <spanx style="verb">[a,b,...]</spanx> with no whitespace.</t>
  <t>Strings: written between <spanx style="verb">"</spanx>. Escape only <spanx style="verb">"</spanx> as <spanx style="verb">\"</spanx>, <spanx style="verb">\</spanx> as <spanx style="verb">\\</spanx>,
and U+0000 to U+001F. Of those, U+0008 is <spanx style="verb">\b</spanx>, U+000C is <spanx style="verb">\f</spanx>,
U+000A is <spanx style="verb">\n</spanx>, U+000D is <spanx style="verb">\r</spanx>, U+0009 is <spanx style="verb">\t</spanx>, and the rest are
<spanx style="verb">\u00xx</spanx> with lowercase hex. Every other character is written as
itself, including U+007F, U+2028, U+2029 and all non-ASCII. A lone
surrogate code unit (U+D800 to U+DFFF not part of a valid pair) is
written <spanx style="verb">\udxxx</spanx> with lowercase hex.</t>
  <t>Integers: decimal, <spanx style="verb">-</spanx> only for negatives, no leading zeros, no
fraction, no exponent.</t>
  <t><spanx style="verb">true</spanx>, <spanx style="verb">false</spanx>, <spanx style="verb">null</spanx> as literals.</t>
  <t>A number that is not an integer can only occur in a value that is
not strict, but such a value can still be signed (<xref target="triage"/> needs
its signing bytes). It is written as Python's json module writes
it: <spanx style="verb">NaN</spanx>, <spanx style="verb">Infinity</spanx>, <spanx style="verb">-Infinity</spanx>, and a finite float as Python's
<spanx style="verb">repr</spanx> (<spanx style="verb">1.5</spanx>, <spanx style="verb">1e+16</spanx>, <spanx style="verb">1.0</spanx>). Not <spanx style="verb">nan</spanx>, not <spanx style="verb">null</spanx>. A signature
over a body holding one verifies only over these spellings (vectors
<spanx style="verb">synth_nan_in_signed_body</spanx>, <spanx style="verb">synth_infinity_in_signed_body</spanx>,
<spanx style="verb">synth_negative_infinity_in_signed_body</spanx>).</t>
</list></t>

<t>The reference ships fixed inputs for this rule
(musubi-v0/canonical_vectors.json) and a Node twin (canonical_v0.mjs)
<xref target="MUSUBI-REF"/>.</t>

<t>This form is not the JSON Canonicalization Scheme <xref target="RFC8785"/>. Among
other differences, RFC 8785 sorts keys by UTF-16 code units, which
gives a different order for keys with characters above U+FFFF, and
serializes non-integer numbers by the rules of ECMAScript, while this
form writes a non-integer number only inside a value that is not
strict, in the spellings of item 6, and does so only so that a
signature over such a value can be checked.</t>

</section>
<section anchor="base64-keys"><name>Base64 and Keys</name>

<t>Base64 is standard base64 (<xref section="4" sectionFormat="of" target="RFC4648"/>) with padding, and
MUST be canonical: decoding then re-encoding gives the same text. A
public key is 32 bytes after decoding, a signature 64 bytes. Any other
length, any non-canonical text, or any non-string is "unusable".</t>

<t>A public key is also unusable if it is not the canonical encoding of a
point on edwards25519, if it is the identity point, or if the point is
not in the prime order subgroup (small order or mixed order points).</t>

</section>
<section anchor="digests"><name>The Digests</name>

<texttable title="Digests">
      <ttcol align='left'>name</ttcol>
      <ttcol align='left'>definition</ttcol>
      <c>contract signing bytes</c>
      <c><spanx style="verb">"a2a-contract-v0\n" + canonical(contract without its "signatures" key)</spanx></c>
      <c>contract_sha256</c>
      <c>lowercase hex sha256 of the contract signing bytes</c>
      <c>record_sha256</c>
      <c>lowercase hex sha256 of <spanx style="verb">canonical(record)</spanx>, the whole record including <spanx style="verb">anchor</spanx> and <spanx style="verb">signatures</spanx></c>
      <c>body digest</c>
      <c>sha256 of <spanx style="verb">canonical(record without its "anchor" key)</spanx>. It includes <spanx style="verb">signatures</spanx>. Anchor proofs start from these 32 bytes</c>
      <c>record signing bytes</c>
      <c>context + <spanx style="verb">canonical(record without "signatures" and without "anchor")</spanx>. Context is <spanx style="verb">"a2a-execution-v0\n"</spanx>, <spanx style="verb">"a2a-revocation-v0\n"</spanx> or <spanx style="verb">"a2a-revocation-ack-v0\n"</spanx> by the record's schema</c>
</texttable>

<t>contract_sha256 is the identity of the terms. It does not change when
a signature is added. A record is bound to a contract only by it
(<xref target="binding"/>).</t>

</section>
<section anchor="ed25519"><name>Ed25519 Verification</name>

<t><spanx style="verb">verify(pub_b64, sig_b64, message)</spanx> returns one of: true, false,
unusable. It returns unusable when the key or signature fails
<xref target="base64-keys"/>. Otherwise it is Ed25519 verification (<xref target="RFC8032"/>) of
the 64 byte signature over the message bytes. The reference uses the
Python <spanx style="verb">cryptography</spanx> package (OpenSSL). See item N9 of
<xref target="MUSUBI-NOTES"/> on libraries that differ for adversarial signatures.</t>

<t>A 64 byte signature is "strict" when its last 32 bytes, read as a
little-endian integer S, are below the group order
L = 2^252 + 27742317777372353535851937790883648493, and its first 32
bytes R pass the public key test of <xref target="base64-keys"/> (a canonical
encoding of a point of prime order, not the identity). "Strictly
verifies" means the signature is strict and verify returned true.
Libraries that disagree on lax signatures all agree on strict ones.</t>

<t>Everywhere below, "verifies" means verify returned true. False and
unusable are both "does not verify", except in the contract door
(<xref target="contract-door"/>), which names them differently.</t>

</section>
</section>
<section anchor="contract"><name>The Contract</name>

<section anchor="contract-door"><name>The Contract Door (Contract Verification)</name>

<t>The contract is accepted only if every check below passes. Any failure
gives the single underspecified reason <spanx style="verb">contract_not_verified</spanx>
(<xref target="settleable-terms"/> is then skipped). The reference reports a list
of refusal codes alongside; those codes are informative. Only
<spanx style="verb">contract_not_verified</spanx> is core.</t>

<t><list style="numbers" type="1">
  <t><spanx style="verb">schema</spanx> is <spanx style="verb">"a2a-contract-v0"</spanx>.</t>
  <t><spanx style="verb">contract_id</spanx> and <spanx style="verb">nonce</spanx> are each 32 lowercase hex characters.</t>
  <t><spanx style="verb">parties</spanx> is a list of exactly two objects whose <spanx style="verb">role</spanx> values are
exactly one <spanx style="verb">"principal"</spanx> and one <spanx style="verb">"contractor"</spanx>.</t>
  <t>Each party's <spanx style="verb">domain</spanx> normalizes to a bare hostname: leading and
trailing whitespace removed (every character Python's <spanx style="verb">str.strip()</spanx>
removes, not only U+0020), then lowercased, then one or more
trailing dots removed. The result has no <spanx style="verb">/</spanx>, <spanx style="verb">@</spanx>, <spanx style="verb">:</spanx> or space, at
least two labels, each label matching
<spanx style="verb">^[a-z0-9]([a-z0-9-]*[a-z0-9])?$</spanx>. The two normalized domains
differ.</t>
  <t>Each party's <spanx style="verb">key_url</spanx> is an <spanx style="verb">https://</spanx> URL. (Whether its host is
under the domain is a finding, not a refusal.)</t>
  <t>Each party's <spanx style="verb">public_key_ed25519_b64</spanx> decodes as canonical base64
to 32 bytes. (The point check of <xref target="base64-keys"/> is applied later,
when signatures are verified.)</t>
  <t><spanx style="verb">task</spanx> is an object with a non-empty string <spanx style="verb">purpose</spanx>.
<spanx style="verb">task.payload_digest</spanx>, when present and not null, is 64 lowercase
hex.</t>
  <t><spanx style="verb">grant</spanx> is an object. Its keys are only: <spanx style="verb">authorized_actions</spanx>,
<spanx style="verb">prohibited_actions</spanx>, <spanx style="verb">conditional</spanx>, <spanx style="verb">delegation</spanx>, <spanx style="verb">data_access</spanx>,
<spanx style="verb">max_hops</spanx>, <spanx style="verb">privacy</spanx>, <spanx style="verb">revocation</spanx>, <spanx style="verb">finality</spanx>, <spanx style="verb">witnesses</spanx>,
<spanx style="verb">expiry_height</spanx>, <spanx style="verb">ordering</spanx>, <spanx style="verb">approval_policy</spanx>, <spanx style="verb">limits</spanx>. Any other
key refuses the contract.</t>
  <t><spanx style="verb">grant.authorized_actions</spanx> and <spanx style="verb">grant.prohibited_actions</spanx> are both
lists of non-empty strings, and no action is in both.</t>
  <t>Typed grant keys, each checked only when present and not null:
  <list style="symbols">
      <t><spanx style="verb">data_access</spanx>: list of non-empty strings.</t>
      <t><spanx style="verb">max_hops</spanx>, <spanx style="verb">expiry_height</spanx>: integer &gt;= 0.</t>
      <t><spanx style="verb">delegation</spanx>: object with exactly one key, <spanx style="verb">allowed</spanx>, a list of
non-empty strings. <spanx style="verb">{}</spanx> refuses the contract (vector
<spanx style="verb">synth_delegation_empty_object</spanx>).</t>
      <t><spanx style="verb">conditional</spanx>: list of objects, each with a non-empty string
<spanx style="verb">action</spanx>; actions unique; <spanx style="verb">requires</spanx>, when present, a string.</t>
      <t><spanx style="verb">finality</spanx>: object with keys only from <spanx style="verb">depth</spanx>,
<spanx style="verb">max_target_bits</spanx>. <spanx style="verb">depth</spanx>, when present, integer &gt;= 1.
<spanx style="verb">max_target_bits</spanx>, when present, 8 lowercase hex, the 32 bit
value written big-endian (<spanx style="verb">207fffff</spanx> is <spanx style="verb">0x207fffff</spanx>), that
decodes (<xref target="compact-target"/>) to a positive target.</t>
      <t><spanx style="verb">witnesses</spanx>: list of objects, each with a non-empty string
<spanx style="verb">name</spanx> and a <spanx style="verb">public_key_ed25519_b64</spanx> that decodes to 32 bytes.</t>
      <t><spanx style="verb">privacy</spanx>: non-empty string.</t>
      <t><spanx style="verb">ordering</spanx>: object with no key other than <spanx style="verb">same_height</spanx>, which
is optional (<spanx style="verb">{}</spanx> is accepted, vector
<spanx style="verb">synth_ordering_empty_object</spanx>); <spanx style="verb">same_height</spanx>, when present, a
string or null.</t>
      <t><spanx style="verb">approval_policy</spanx>: object with keys only from <spanx style="verb">allow_unscoped</spanx>
(boolean) and <spanx style="verb">approvers</spanx>. <spanx style="verb">approvers</spanx>, when present, must be a
list of 1 to 16 objects, each with keys only from <spanx style="verb">name</spanx>,
<spanx style="verb">public_key_ed25519_b64</spanx> and <spanx style="verb">actions</spanx>: a non-empty string
<spanx style="verb">name</spanx>, a canonical 32 byte <spanx style="verb">public_key_ed25519_b64</spanx>, names and
keys unique across the list, and <spanx style="verb">actions</spanx> a non-empty list of
distinct non-empty strings, each the <spanx style="verb">action</spanx> of an entry of
<spanx style="verb">conditional</spanx>. Anything else refuses. A list that passes is
carried and out of scope (<xref target="scope"/>).</t>
      <t><spanx style="verb">revocation</spanx>: an object whose <spanx style="verb">effective_at</spanx> is <spanx style="verb">"anchor"</spanx> or
<spanx style="verb">"delivery_ack"</spanx>, with keys only from <spanx style="verb">effective_at</spanx>,
<spanx style="verb">ack_window</spanx>; <spanx style="verb">ack_window</spanx>, when present, integer &gt;= 1.</t>
    </list></t>
  <t>No number anywhere in the contract (signatures included) is a
non-integer, NaN, an infinity, or an integer outside plus or minus
(2^53 - 1).</t>
  <t>Every object key anywhere in the contract is printable ASCII
(U+0020 to U+007E).</t>
  <t><spanx style="verb">establishes</spanx> is a non-empty list. No string in it matches,
case-insensitively, any of the patterns listed after this list.</t>
  <t><spanx style="verb">does_not_establish</spanx> is a non-empty list. Lowercased, its strings
together contain: one of <spanx style="verb">runtime</spanx> or <spanx style="verb">enforce</spanx>; one of <spanx style="verb">judge</spanx>,
<spanx style="verb">liability</spanx> or <spanx style="verb">fault</spanx>; <spanx style="verb">deviat</spanx>; and <spanx style="verb">legal</spanx>. Each of the four
groups must be hit by at least one string.</t>
  <t><spanx style="verb">signatures</spanx> is a list. For each object in it: its <spanx style="verb">domain</spanx>
normalizes (rule 4) to a party's normalized domain, or the contract
is refused. Its <spanx style="verb">signature</spanx> is verified (<xref target="ed25519"/>) with that
party's key over the contract signing bytes. False or unusable
refuses the contract, even if the same party also has a good
signature (WRITTEN BEHAVIOUR). <spanx style="verb">alg</spanx> is not read.</t>
  <t>Each party has at least one signature that verifies.</t>
</list></t>

<t>The patterns of rule 13, one per line, are:</t>

<figure title="Patterns refused in establishes"><artwork><![CDATA[
\bperformed\b
\bdeliver(ed|y)\b
\bmoney (moved|was sent)\b
\bfunds? (moved|were sent|were transferred)\b
\bpaid\b(?!\s+for\s+(this|the)\s+record)
\bpayment (was|has been) (made|completed|settled|received)\b
\bcontract\b
\bbinding\b
\bguarantee
\bescrow\b
\bcustody\b
\bsolvent\b
\blawful\b
\bfair\b
\bcertified\b
]]></artwork></figure>

<t>Not read by the door, and not read anywhere in core v0: <spanx style="verb">expiry</spanx> (a
wall clock), <spanx style="verb">agreed_at</spanx>, <spanx style="verb">requirements.recovery</spanx>,
<spanx style="verb">selection_provenance</spanx>, <spanx style="verb">liability_boundary</spanx>, <spanx style="verb">settlement</spanx>.
<spanx style="verb">requirements.furui</spanx> is not read in core v0 (<xref target="admission"/>, outside
it, reads it). <spanx style="verb">lower_bound</spanx> is not checked here; <xref target="view"/> checks it.
<spanx style="verb">requirements.evidence</spanx> is read in <xref target="nenrin"/>, <spanx style="verb">bond</spanx> and
<spanx style="verb">parent_contract</spanx> in <xref target="strict-terms"/>.</t>

</section>
<section anchor="settleable-terms"><name>Settleable Terms</name>

<t>Only when <xref target="contract-door"/> accepts, each of these adds the
underspecified reason shown:</t>

<texttable title="Settleable terms">
      <ttcol align='left'>condition</ttcol>
      <ttcol align='left'>reason</ttcol>
      <c><spanx style="verb">grant.authorized_actions</spanx> is not a list</c>
      <c><spanx style="verb">authorized_actions_missing</spanx></c>
      <c>an action name in <spanx style="verb">authorized_actions</spanx>, <spanx style="verb">prohibited_actions</spanx> or <spanx style="verb">conditional[].action</spanx> does not match <spanx style="verb">^[a-z][a-z0-9_.:-]{0,63}$</spanx> (see <xref target="written-behaviour"/> and item N1 of <xref target="MUSUBI-NOTES"/>)</c>
      <c><spanx style="verb">grant_action_name_invalid</spanx></c>
      <c>more than 32 witnesses</c>
      <c><spanx style="verb">too_many_witnesses</spanx></c>
      <c>a witness key does not decode to 32 bytes</c>
      <c><spanx style="verb">witness_key_invalid</spanx></c>
      <c>two witnesses share a name, or any two of {principal key, contractor key, each witness key} decode to the same 32 bytes</c>
      <c><spanx style="verb">key_conflict</spanx></c>
      <c><spanx style="verb">grant.ordering.same_height</spanx> is present and is neither null nor <spanx style="verb">"conservative"</spanx></c>
      <c><spanx style="verb">tie_policy_refused</spanx></c>
      <c><spanx style="verb">grant.revocation</spanx> is an object and <spanx style="verb">effective_at</spanx> is not <spanx style="verb">"anchor"</spanx> or <spanx style="verb">"delivery_ack"</spanx></c>
      <c><spanx style="verb">revocation_policy_not_recomputable</spanx></c>
      <c><spanx style="verb">effective_at</spanx> is <spanx style="verb">"delivery_ack"</spanx> and <spanx style="verb">ack_window</spanx> is not an integer &gt;= 1</c>
      <c><spanx style="verb">ack_window_missing</spanx></c>
      <c><spanx style="verb">grant.finality.depth</spanx> is not an integer &gt;= 1 (including a missing <spanx style="verb">finality</spanx> or <spanx style="verb">depth</spanx>)</c>
      <c><spanx style="verb">finality_policy_missing</spanx></c>
      <c><spanx style="verb">grant.expiry_height</spanx> is present, not null, and not an integer &gt;= 0</c>
      <c><spanx style="verb">bad_expiry_height</spanx></c>
      <c><spanx style="verb">allow_unscoped</spanx> in <spanx style="verb">grant.approval_policy</spanx> is truthy</c>
      <c><spanx style="verb">unscoped_approvals_refused</spanx></c>
</texttable>

<t>Any of these stops the computation: no record is examined further
(<xref target="computation"/>).</t>

<t>An empty <spanx style="verb">authorized_actions</spanx> list means nothing is authorized. It is
not an error.</t>

<t>Four rows of this table cannot add their reason in core v0:
<spanx style="verb">authorized_actions_missing</spanx>, <spanx style="verb">witness_key_invalid</spanx>,
<spanx style="verb">revocation_policy_not_recomputable</spanx> and <spanx style="verb">bad_expiry_height</spanx>. Rules 9
and 10 of <xref target="contract-door"/> refuse the same inputs first, and this
table applies only when <xref target="contract-door"/> accepts, so the input
settles as <spanx style="verb">contract_not_verified</spanx>. The rows are kept so that no code
changes meaning; an implementation that never emits these four agrees
with the reference (revision 0.4, <spanx style="verb">exercise_inputs.py</spanx> cases D1 to D6,
<xref target="exercise"/>).</t>

</section>
<section anchor="strict-terms"><name>Strict Terms</name>

<t>Whether or not <xref target="contract-door"/> accepts, each of these adds the
underspecified reason shown:</t>

<texttable title="Strict terms">
      <ttcol align='left'>condition</ttcol>
      <ttcol align='left'>reason</ttcol>
      <c>a <spanx style="verb">signatures[].signature</spanx> decodes to 64 bytes and is not strict (<xref target="ed25519"/>)</c>
      <c><spanx style="verb">contract_signature_not_strict</spanx></c>
      <c><spanx style="verb">bond</spanx> is present and not null, and is not an object whose <spanx style="verb">amount</spanx> is an integer from 1 to 2^53 - 1 (<spanx style="verb">true</spanx> is not an integer)</c>
      <c><spanx style="verb">bond_malformed</spanx></c>
      <c><spanx style="verb">parent_contract</spanx> is present and not null, and no parent was supplied</c>
      <c><spanx style="verb">parent_not_supplied</spanx></c>
      <c><spanx style="verb">parent_contract</spanx> is present and not null, and the parent is not an object, or does not pass <xref target="contract-door"/> on its own, or the contract does not pass <xref target="contract-door"/> with the parent supplied (the reference's door then also checks that <spanx style="verb">parent_contract.contract_id</spanx> equals the parent's <spanx style="verb">contract_id</spanx> and that the child's grant is a subset of the parent's: <spanx style="verb">contract_v0.grant_subset</spanx>, stated in <xref target="parent-narrowing"/>)</c>
      <c><spanx style="verb">parent_narrowing_failed</spanx></c>
</texttable>

<t>These checks do not stop the walk; the event set and dispositions are
still computed (<xref target="computation"/>). In the last row, "does not pass
<xref target="contract-door"/> with the parent supplied" is judged by
<xref target="parent-narrowing"/> and adds only <spanx style="verb">parent_narrowing_failed</spanx>; whether
the settlement is <spanx style="verb">contract_not_verified</spanx> is decided by
<xref target="contract-door"/> on the contract on its own (vectors
<spanx style="verb">synth_parent_full_*</spanx>).</t>

<section anchor="parent-narrowing"><name>Parent Narrowing (grant_subset)</name>

<t>This applies when <spanx style="verb">parent_contract</spanx> is present and not null. The
contract passes <xref target="contract-door"/> with the parent supplied if and only
if it passes <xref target="contract-door"/> on its own, <spanx style="verb">parent_contract</spanx> is an
object whose <spanx style="verb">contract_id</spanx> equals the parent's <spanx style="verb">contract_id</spanx> (a
missing <spanx style="verb">contract_id</spanx> fails), and the child grant C is within the
parent grant P. C is within P when none of the following holds. In
each rule, a list that is absent or null counts as empty, "has an
integer" means the key is present with an integer that is not a
boolean, and "the parent sets" a key means the key is present with a
value other than null. Both grants have already passed rule 10 of
<xref target="contract-door"/>, so every value has its declared type.</t>

<t><list style="numbers" type="1">
  <t>An entry of <spanx style="verb">C.authorized_actions</spanx> is not in
<spanx style="verb">P.authorized_actions</spanx>.</t>
  <t>An entry of <spanx style="verb">P.prohibited_actions</spanx> is not in
<spanx style="verb">C.prohibited_actions</spanx>.</t>
  <t>An entry of <spanx style="verb">C.data_access</spanx> is not in <spanx style="verb">P.data_access</spanx>. A child
without <spanx style="verb">data_access</spanx> is narrower, not wider.</t>
  <t>P has an integer <spanx style="verb">max_hops</spanx> p, and C has no integer <spanx style="verb">max_hops</spanx> or
one above p - 1. A child that copies the parent's <spanx style="verb">max_hops</spanx> is not
within (vector <spanx style="verb">synth_parent_full_identical</spanx>), and under a parent
with <spanx style="verb">max_hops</spanx> 0 no child is within.</t>
  <t>An entry of <spanx style="verb">C.delegation.allowed</spanx> is not in
<spanx style="verb">P.delegation.allowed</spanx>.</t>
  <t>For each entry of <spanx style="verb">P.conditional</spanx>, with <spanx style="verb">a</spanx> its <spanx style="verb">action</spanx> and r its
<spanx style="verb">requires</spanx> (a missing <spanx style="verb">requires</spanx> is null, on either side): skip it
when either (i) <spanx style="verb">a</spanx> is in <spanx style="verb">C.prohibited_actions</spanx>, or (ii) <spanx style="verb">a</spanx> is
not in <spanx style="verb">C.authorized_actions</spanx> and is not the <spanx style="verb">action</spanx> of any entry
of <spanx style="verb">C.conditional</spanx> (the child cannot do <spanx style="verb">a</spanx> at all). Otherwise it
is a violation if no entry of <spanx style="verb">C.conditional</spanx> has the action <spanx style="verb">a</spanx>,
or the one that has it (actions are unique, rule 10 of
<xref target="contract-door"/>) has a <spanx style="verb">requires</spanx> other than r. An action that C
both authorizes and makes conditional is judged by this rule like
any other.</t>
  <t>The <spanx style="verb">action</spanx> of an entry of <spanx style="verb">C.conditional</spanx> is in neither
<spanx style="verb">P.authorized_actions</spanx> nor the actions of <spanx style="verb">P.conditional</spanx>.</t>
  <t>P has an integer <spanx style="verb">expiry_height</spanx> e, and C has no integer
<spanx style="verb">expiry_height</spanx> or one above e.</t>
  <t>P has a <spanx style="verb">revocation</spanx> object; let m be its <spanx style="verb">effective_at</spanx>. Then C
must have a <spanx style="verb">revocation</spanx> object with an <spanx style="verb">effective_at</spanx>. If m is
<spanx style="verb">"anchor"</spanx>, C's must be <spanx style="verb">"anchor"</spanx>. If m is <spanx style="verb">"delivery_ack"</spanx>, C's
may be <spanx style="verb">"anchor"</spanx> (authority ends sooner) or <spanx style="verb">"delivery_ack"</spanx>; in
the second case, when P has an integer <spanx style="verb">ack_window</spanx> w, C must have
an integer <spanx style="verb">ack_window</spanx> of at most w (a child without one is a
violation), and when P has none, C's <spanx style="verb">ack_window</spanx> is not compared.
Anything else is a violation.</t>
  <t>P has a <spanx style="verb">finality</spanx> object. Then C must have one; when P has an
integer <spanx style="verb">depth</spanx> d, C must have an integer <spanx style="verb">depth</spanx> of at least d;
when P's <spanx style="verb">max_target_bits</spanx> decodes (<xref target="compact-target"/>), C's must
be present and decode to a target no larger (absent, or not
decoding, is a violation). Both grants passed rule 10 of
<xref target="contract-door"/>, so a target that is present is positive.</t>
  <t>P has a witness with a string <spanx style="verb">public_key_ed25519_b64</spanx>. Then C
must have at least one witness with such a key, and every such key
of P must equal, as a string, the key of one of C's witnesses.
Witnesses are compared by key; the names are labels (vector
<spanx style="verb">synth_parent_full_witness_renamed</spanx>).</t>
  <t>For each of <spanx style="verb">privacy</spanx> and <spanx style="verb">ordering</spanx> separately: when the parent
sets it, C must set it to an equal JSON value. After rule 10 of
<xref target="contract-door"/> these values hold only strings, objects and
null, so equality is exact: <spanx style="verb">{}</spanx> is not equal to
<spanx style="verb">{"same_height": null}</spanx>. No order is defined for these two, so
they are held equal. An <spanx style="verb">ordering</spanx> of <spanx style="verb">{"same_height": null}</spanx>
counts as set (vector <spanx style="verb">synth_parent_full_without_ordering</spanx>).</t>
  <t><spanx style="verb">C.approval_policy.allow_unscoped</spanx> is <spanx style="verb">true</spanx> and
<spanx style="verb">P.approval_policy.allow_unscoped</spanx> is not (absent counts as not
true). (In core v0 a true <spanx style="verb">allow_unscoped</spanx> already gives
<spanx style="verb">unscoped_approvals_refused</spanx>, <xref target="settleable-terms"/>.)</t>
  <t>The pinned approvers differ: the parent's
<spanx style="verb">approval_policy.approvers</spanx> and the child's, each taken as the
empty list when absent, compared as sets of canonical entries. The
order of the entries does not matter; each entry is compared as
its canonical bytes (<xref target="canonical"/>), so the order of the strings
in its <spanx style="verb">actions</spanx> does, as for any array. A child that carries
approvers is out of scope (<xref target="scope"/>), so in core v0 this applies
when the parent pins approvers and the child has none: a violation
(vector <spanx style="verb">synth_parent_approvers_child_drops</spanx>). No core v0 input
can reach the order of <spanx style="verb">actions</spanx>.</t>
</list></t>

<t>The parent must also pass <xref target="contract-door"/> on its own. Vectors
<spanx style="verb">synth_parent_full_*</spanx> take a parent that sets every axis and change
one axis per vector.</t>

</section>
</section>
</section>
<section anchor="view"><name>The Header View</name>

<section anchor="view-verify"><name>Verifying the View</name>

<t>The view is checked under two contract terms: the difficulty floor
<spanx style="verb">grant.finality.max_target_bits</spanx> and the checkpoint <spanx style="verb">lower_bound</spanx>
(<spanx style="verb">height</spanx>, <spanx style="verb">hash</spanx>). Any failure below gives the underspecified reason
<spanx style="verb">chain_view_rejected</spanx>. The view is checked even when <xref target="contract"/>
already failed.</t>

<t><list style="numbers" type="1">
  <t><spanx style="verb">grant.finality.max_target_bits</spanx> is 8 lowercase hex, the 32 bit
value written big-endian, and decodes (<xref target="compact-target"/>) to a
target, the floor.</t>
  <t><spanx style="verb">lower_bound</spanx> is an object with an integer <spanx style="verb">height</spanx> and a 64
lowercase hex <spanx style="verb">hash</spanx>.</t>
  <t><spanx style="verb">view.headers</spanx> is a non-empty list.</t>
  <t>Each header is an object with an integer <spanx style="verb">height</spanx> &gt;= 0 and a <spanx style="verb">hex</spanx>
of 160 lowercase hex characters, optionally followed by one line
feed (WRITTEN BEHAVIOUR of the lower layers; rule 9 below then
refuses the line feed under its own reason).</t>
  <t>Heights are consecutive in list order: each is the previous plus
one.</t>
  <t>From the second header on, bytes 4 to 36 of the header, reversed,
as hex, equal the previous header's hash.</t>
  <t>The header's own nBits (bytes 72 to 76, little-endian unsigned 32
bit) decode to a target; that target is not above the floor; the
header hash read as a big-endian 256 bit number is not above that
target.</t>
  <t>The view holds a header at <spanx style="verb">lower_bound.height</spanx> whose hash equals
<spanx style="verb">lower_bound.hash</spanx>.</t>
</list></t>

<t>The header hash is sha256(sha256(80 raw bytes)), byte-reversed, as
lowercase hex (the usual display order). The merkle root used by
anchors (<xref target="anchor-proof"/>) is bytes 36 to 68 of the raw header, not
reversed. The tip is the height of the last header. The view may
start below the checkpoint. The first header's previous-hash field is
not checked. Consensus rules beyond these (retargeting, timestamps,
transactions) are not checked.</t>

<t><list style="numbers" start="9" type="1">
  <t>Separately, when <spanx style="verb">view</spanx> is an object whose <spanx style="verb">headers</spanx> is a list,
every header is an object with an integer <spanx style="verb">height</spanx> and a <spanx style="verb">hex</spanx> of
exactly 160 lowercase hex characters, and every string, key and
number in the view is strict (first paragraph of <xref target="triage"/>). If
not, the reason <spanx style="verb">view_not_strict</spanx> is added. A header in upper case
therefore gives both <spanx style="verb">chain_view_rejected</spanx> and <spanx style="verb">view_not_strict</spanx>;
one that ends in a line feed gives only <spanx style="verb">view_not_strict</spanx>. A view
that is not an object, or whose <spanx style="verb">headers</spanx> is not a list, gives
<spanx style="verb">chain_view_rejected</spanx> only (vectors <spanx style="verb">synth_view_not_an_object</spanx>,
<spanx style="verb">synth_view_headers_not_a_list</spanx>). Rule 9 does not stop the walk.</t>
</list></t>

<t>If the view verifies but holds more than 200000 headers, the reason
is <spanx style="verb">input_too_large</spanx> instead, and no record is examined.</t>

</section>
<section anchor="compact-target"><name>Compact Target Decoding</name>

<t>For a 32 bit value <spanx style="verb">bits</spanx>: <spanx style="verb">exp = bits &gt;&gt; 24</spanx>,
<spanx style="verb">mant = bits &amp; 0x007fffff</spanx>. If <spanx style="verb">bits &amp; 0x00800000</spanx> is set, or <spanx style="verb">mant</spanx>
is 0, there is no target (invalid). Otherwise the target is
<spanx style="verb">mant &lt;&lt; (8 * (exp - 3))</spanx> when <spanx style="verb">exp &gt;= 3</spanx>, else
<spanx style="verb">mant &gt;&gt; (8 * (3 - exp))</spanx>.</t>

<t><spanx style="verb">max_target_bits</spanx> gives <spanx style="verb">bits</spanx> as 8 lowercase hex: the 32 bit value
written big-endian, so <spanx style="verb">207fffff</spanx> is <spanx style="verb">0x207fffff</spanx> and <spanx style="verb">17080000</spanx> is
<spanx style="verb">0x17080000</spanx>. A header's own nBits (rule 7 of <xref target="view-verify"/>) is the
same kind of value stored little-endian in bytes 72 to 76.</t>

</section>
</section>
<section anchor="records"><name>Records Before Their Anchors</name>

<section anchor="triage"><name>Strict Records and Triage</name>

<t>A JSON value is strict when, at every depth up to 512 levels: every
string and every object key is free of U+0000 to U+001F and U+007F;
every key is ASCII; every number is an integer within plus or minus
(2^53 - 1). A value nested deeper is not strict.</t>

<t>A record's parts:</t>

<t><list style="symbols">
  <t>the body: the record without <spanx style="verb">signatures</spanx> and <spanx style="verb">anchor</spanx>. This is what
its signers signed (<xref target="digests"/>, record signing bytes).</t>
  <t>the anchor: <spanx style="verb">anchor</spanx>. Nobody signed it.</t>
  <t>the signatures list: <spanx style="verb">signatures</spanx>. Nobody signed it either.</t>
</list></t>

<t>The body is not strict when the body as a value is not strict, or
(when present) <spanx style="verb">nenrin_ref</spanx> is neither null nor 64 lowercase hex, or
<spanx style="verb">contract_ref.contract_sha256</spanx> or <spanx style="verb">contract_ref.payload_digest</spanx> is
present and not 64 lowercase hex, or, for an execution:
<spanx style="verb">performed_actions</spanx> or <spanx style="verb">delegated_to</spanx> is present and not a list of
strings; <spanx style="verb">approvals</spanx> is present and not a list of objects; or an
approval has a <spanx style="verb">sig_b64</spanx> or a <spanx style="verb">signature</spanx> that decodes to 64 bytes and
is not strict (<xref target="ed25519"/>). Each of the two keys is checked on its
own, whichever are present: a strict <spanx style="verb">sig_b64</spanx> does not excuse a
<spanx style="verb">signature</spanx> that is not strict. This is the only place an approval's
<spanx style="verb">signature</spanx> is read; <xref target="approvals"/> never verifies it.</t>

<t>The anchor is not strict when it is present and: it is not strict as
a value; or its <spanx style="verb">block_hash</spanx> is present and not exactly 64 lowercase
hex; or a proof operation's <spanx style="verb">hex</spanx> is present and is not exactly pairs
of lowercase hex digits (the empty string is allowed).</t>

<t>A record is strictly signed when one entry of <spanx style="verb">signatures</spanx> has a role
its schema allows (<xref target="authentication"/>), a signature that strictly
verifies (<xref target="ed25519"/>) with the matching key over the record signing
bytes, and is itself a strict value.</t>

<t>First, a record of one of the three schemas whose record_sha256
cannot be computed at all (for example an integer too long to write
out) is set aside with reason <spanx style="verb">record_not_strict</spanx>. The others are
bound (<xref target="binding"/>). Each bound record is then triaged, in this order:</t>

<texttable title="Triage">
      <ttcol align='left'>condition</ttcol>
      <ttcol align='left'>result</ttcol>
      <c>not strictly signed, body not strict</c>
      <c>set aside, <spanx style="verb">record_not_strict</spanx></c>
      <c>not strictly signed, but a signature verifies under <xref target="ed25519"/> alone (lax)</c>
      <c>set aside, <spanx style="verb">signature_not_strict</spanx></c>
      <c>not strictly signed otherwise</c>
      <c>rejected (disposition <spanx style="verb">rejected</spanx>)</c>
      <c>strictly signed, body not strict</c>
      <c>set aside, <spanx style="verb">record_not_strict</spanx>, and the underspecified reason <spanx style="verb">record_not_strict</spanx> is added: the record is a party's own statement, so setting it aside does not make the input settleable</c>
      <c>strictly signed, anchor not strict, and another bound record with the same record signing bytes is kept by this table and anchors validly (<xref target="collapse"/>)</c>
      <c>set aside, <spanx style="verb">anchor_not_strict</spanx></c>
      <c>strictly signed, anchor not strict, otherwise</c>
      <c>underspecified reason <spanx style="verb">anchor_not_strict</spanx></c>
      <c>a revocation, and <spanx style="verb">grant.revocation</spanx> is not an object</c>
      <c>set aside, <spanx style="verb">revocation_not_provided_for</spanx> (the contract cannot be revoked)</c>
      <c>otherwise</c>
      <c>kept</c>
</texttable>

<t>Only the fourth row both sets a record aside and adds a reason; the
first three set aside or reject without one (vectors
<spanx style="verb">synth_nan_in_signed_body</spanx> and <spanx style="verb">synth_unsigned_float_beside_good</spanx>).</t>

<t>Set aside and rejected records are never examined again: no collapse
(<xref target="collapse"/>), size limit (<xref target="size-limit"/>), conformance check, anchor
check or NENRIN check sees them. They are listed with their
disposition (<xref target="dispositions"/>). Problems in the signatures list of a
strictly signed record (an extra entry that does not verify or is not
strict) are ignored.</t>

<t>Any party holding copies of genuine records can change their
anchors and signature lists but not their bodies. The table is
arranged so that such copies are set aside or rejected whenever the
genuine record is in the input.</t>

</section>
<section anchor="binding"><name>Binding</name>

<t>Let <spanx style="verb">cid = contract.contract_id</spanx>, <spanx style="verb">pdg = contract.task.payload_digest</spanx>
(it may be absent), <spanx style="verb">csha = contract_sha256</spanx>. Walk <spanx style="verb">events</spanx> in the
given order. Skip anything that is not an object, or whose <spanx style="verb">schema</spanx> is
not one of the three record schemas. Skip a record whose record_sha256
was already seen (identical bytes are one record). Then, with
<spanx style="verb">ref = contract_ref</spanx> if it is an object, else empty:</t>

<texttable title="Binding">
      <ttcol align='left'>condition, in this order</ttcol>
      <ttcol align='left'>class</ttcol>
      <c><spanx style="verb">ref.contract_sha256</spanx> absent or null, and <spanx style="verb">ref.contract_id == cid</spanx></c>
      <c>unbound (not settled in strict mode)</c>
      <c><spanx style="verb">ref.contract_sha256</spanx> absent or null otherwise</c>
      <c>not ours: dropped silently</c>
      <c><spanx style="verb">ref.contract_sha256</spanx> is not 64 lowercase hex</c>
      <c>inconsistent</c>
      <c><spanx style="verb">ref.contract_sha256 != csha</spanx></c>
      <c>foreign</c>
      <c><spanx style="verb">ref.contract_id != cid</spanx></c>
      <c>inconsistent</c>
      <c><spanx style="verb">pdg</spanx> is present and not null, and <spanx style="verb">ref.payload_digest != pdg</spanx></c>
      <c>inconsistent</c>
      <c>otherwise</c>
      <c>bound</c>
</texttable>

<t>Only bound records go further. Foreign, inconsistent and unbound
records are listed (<xref target="output"/>) and never settled. Binding runs even
when <xref target="contract"/> or <xref target="view"/> failed; its classes are then still
reported. Binding is by <spanx style="verb">contract_sha256</spanx> alone, so a record made
under earlier terms with the same <spanx style="verb">contract_id</spanx> is foreign and is
never triaged (<xref target="triage"/>).</t>

</section>
<section anchor="size-limit"><name>Size Limit</name>

<t>If more than 10000 records are kept by <xref target="triage"/>, the reason
<spanx style="verb">input_too_large</spanx> is added and no record is examined further. Records
are counted after identical bytes are merged (<xref target="identical-bytes"/>):
10001 copies of one record are one record (revision 0.4,
<spanx style="verb">exercise_inputs.py</spanx> cases E4 to E6, <xref target="exercise"/>).</t>

</section>
<section anchor="identical-bytes"><name>Identical Bytes</name>

<t>Two inputs with the same record_sha256 are one record. Supplying a
record twice changes nothing.</t>

</section>
<section anchor="collapse"><name>Several Copies of One Signed Body</name>

<t>The same signed body can reach the input more than once: anchored
twice (two calendars, two batches), or with another <spanx style="verb">signatures</spanx> list.
These copies differ only in their unsigned parts. Group the records
kept by <xref target="triage"/> by their record signing bytes (<xref target="digests"/>). In a
group with more than one distinct record_sha256:</t>

<t><list style="symbols">
  <t>A copy "anchors validly" if: its <spanx style="verb">anchor.height</spanx> is an integer &gt;= 0
that the view holds; <spanx style="verb">anchor.block_hash</spanx> equals the view's hash at
that height; the proof (<xref target="anchor-proof"/>) reaches that block's
merkle root; and the batch leg rule (<xref target="batch-leg"/>) accepts it.</t>
  <t>If at least one copy anchors validly, keep the one with the smallest
(height, record_sha256) and drop the others. Dropped copies are
listed (disposition <spanx style="verb">duplicate_dropped</spanx>).</t>
  <t>If none does, keep them all. Each is then examined on its own.</t>
</list></t>

<t>This is done when rules 1 to 8 of <xref target="view-verify"/> passed, whether or
not <xref target="contract"/> did. Note that "anchors validly" does not compare the
height with the checkpoint (see item N7 of <xref target="MUSUBI-NOTES"/>).</t>

</section>
<section anchor="authentication"><name>Authentication</name>

<t>Authentication is part of <xref target="triage"/>: only strictly signed records are
kept. The roles and keys are:</t>

<texttable title="Signers by schema">
      <ttcol align='left'>schema</ttcol>
      <ttcol align='left'>signer it needs</ttcol>
      <ttcol align='left'>key</ttcol>
      <c>a2a-execution-v0</c>
      <c>a signature with <spanx style="verb">role</spanx> <spanx style="verb">"contractor"</spanx>, or with <spanx style="verb">role</spanx> <spanx style="verb">"witness"</spanx> and a <spanx style="verb">witness_name</spanx></c>
      <c>the contractor's party key, or the key of the grant witness with that exact name</c>
      <c>a2a-revocation-v0</c>
      <c><spanx style="verb">role</spanx> <spanx style="verb">"principal"</spanx></c>
      <c>the principal's party key</c>
      <c>a2a-revocation-ack-v0</c>
      <c><spanx style="verb">role</spanx> <spanx style="verb">"contractor"</spanx></c>
      <c>the contractor's party key</c>
</texttable>

<t><spanx style="verb">signatures</spanx> must be a non-empty list. Each entry is an object with
<spanx style="verb">role</spanx>, <spanx style="verb">sig_b64</spanx> and, for a witness, <spanx style="verb">witness_name</spanx>. An entry with
another role, or one that does not strictly verify, is skipped.</t>

<t>Then, for a kept revocation: if its <spanx style="verb">revoked_by</spanx> field is not
<spanx style="verb">"principal"</spanx>, it is ignored (disposition <spanx style="verb">ignored</spanx>), even though the
principal signed it.</t>

</section>
<section anchor="conformance"><name>Conformance</name>

<t>Allowed top-level fields:</t>

<t><list style="symbols">
  <t>execution: <spanx style="verb">schema</spanx>, <spanx style="verb">contract_ref</spanx>, <spanx style="verb">performed_actions</spanx>,
<spanx style="verb">approvals</spanx>, <spanx style="verb">delegated_to</spanx>, <spanx style="verb">nenrin_ref</spanx>, <spanx style="verb">outcomes</spanx>, <spanx style="verb">anchor</spanx>,
<spanx style="verb">signatures</spanx>,</t>
  <t>revocation: <spanx style="verb">schema</spanx>, <spanx style="verb">contract_ref</spanx>, <spanx style="verb">revoked_by</spanx>, <spanx style="verb">anchor</spanx>,
<spanx style="verb">signatures</spanx>,</t>
  <t>acknowledgement: <spanx style="verb">schema</spanx>, <spanx style="verb">contract_ref</spanx>, <spanx style="verb">revocation_sha256</spanx>,
<spanx style="verb">acked_by</spanx>, <spanx style="verb">anchor</spanx>, <spanx style="verb">signatures</spanx>.</t>
</list></t>

<t>A record has a conformance problem if:</t>

<t><list style="symbols">
  <t>it has a field outside its list,</t>
  <t>(execution) <spanx style="verb">performed_actions</spanx> is not a list of strings,</t>
  <t>(execution) <spanx style="verb">outcomes</spanx> is present, not null, and not an object; or
one of its keys is not in <spanx style="verb">performed_actions</spanx>; or one of its values
is not <spanx style="verb">"ok"</spanx>, <spanx style="verb">"error"</spanx>, <spanx style="verb">"timeout"</spanx> or <spanx style="verb">"partial"</spanx>,</t>
  <t>(execution) it lists more than 256 performed actions,</t>
  <t>(execution) a performed action, or the <spanx style="verb">action</spanx> of an approval that
is an object, does not match <spanx style="verb">^[a-z][a-z0-9_.:-]{0,63}$</spanx>. An
approval object with no <spanx style="verb">action</spanx> counts as not matching. (A name
ending in a line feed never gets here: its body is not strict,
<xref target="triage"/>.)</t>
</list></t>

<t>If an execution with a problem has a contractor signature that
verifies, it is a charge: the deviation <spanx style="verb">nonconforming_record</spanx> is
added, and the record continues to <xref target="anchors"/>. Its acts are still
judged. Any other record with a problem is rejected (disposition
<spanx style="verb">rejected</spanx>). The computation is then repeated without the rejected
records, until no new record is rejected, so no check reads a
rejected record.</t>

<t>The types of <spanx style="verb">approvals</spanx>, <spanx style="verb">delegated_to</spanx>, <spanx style="verb">nenrin_ref</spanx> and
<spanx style="verb">contract_ref</spanx> are checked in <xref target="triage"/>.</t>

</section>
</section>
<section anchor="anchors"><name>Anchors</name>

<section anchor="anchor-what"><name>What an Anchor Is</name>

<t>An anchor says where a record's body was committed to Bitcoin:</t>

<figure><artwork><![CDATA[
"anchor": {"height": H, "block_hash": "<64 hex>", "proof": [op, ...]}
]]></artwork></figure>

<t>The anchor is not signed. Anyone holding a record can change it. The
checks below decide whether the claim holds against the verified
view.</t>

</section>
<section anchor="anchor-height"><name>Height</name>

<t>The height of a record is <spanx style="verb">anchor.height</spanx> when <spanx style="verb">anchor</spanx> is an object
and <spanx style="verb">height</spanx> is an integer &gt;= 0. Otherwise the record has no height.</t>

</section>
<section anchor="anchor-order"><name>Order of the Checks</name>

<t>For each record that passed <xref target="records"/>, apply <xref target="anchor-block"/>,
<xref target="anchor-proof"/> and <xref target="batch-leg"/> in this order. The first failure
decides what happens to the record. Every record that reaches this
section is strictly signed by a party (<xref target="triage"/>).</t>

</section>
<section anchor="anchor-block"><name>Height and Block</name>

<texttable title="Height and block">
      <ttcol align='left'>condition, in this order</ttcol>
      <ttcol align='left'>result</ttcol>
      <c>no height (<xref target="anchor-height"/>), or <spanx style="verb">anchor.block_hash</spanx> is not 64 lowercase hex</c>
      <c>underspecified reason <spanx style="verb">anchor_incomplete</spanx></c>
      <c>height &lt; <spanx style="verb">lower_bound.height</spanx></c>
      <c>underspecified reason <spanx style="verb">anchor_below_lower_bound</spanx>. The body names this contract, and the contract names the checkpoint's block hash, so no honest anchor can be below it</c>
      <c>height &gt; tip, or the view holds no header at that height</c>
      <c>underspecified reason <spanx style="verb">anchor_beyond_observed_chain</spanx></c>
      <c><spanx style="verb">anchor.block_hash</spanx> != the view's hash at that height</c>
      <c>disposition <spanx style="verb">orphaned</spanx>: excluded, no reason added, but its height counts for finality (<xref target="status"/>)</c>
</texttable>

</section>
<section anchor="anchor-proof"><name>The Proof</name>

<t><spanx style="verb">anchor.proof</spanx> must be a non-empty list of at most 512 operations.
Start with the 32 byte body digest (<xref target="digests"/>) as the current bytes,
and apply each operation:</t>

<texttable title="Proof operations">
      <ttcol align='left'>op</ttcol>
      <ttcol align='left'>effect</ttcol>
      <c><spanx style="verb">{"op": "sha256"}</spanx></c>
      <c>current = sha256(current)</c>
      <c><spanx style="verb">{"op": "hexlify"}</spanx></c>
      <c>current = the ASCII bytes of the lowercase hex of current</c>
      <c><spanx style="verb">{"op": "append", "hex": X}</spanx></c>
      <c>current = current + bytes(X)</c>
      <c><spanx style="verb">{"op": "prepend", "hex": X}</spanx></c>
      <c>current = bytes(X) + current</c>
</texttable>

<t><spanx style="verb">X</spanx> is a string of pairs of lowercase hex digits (the empty string is
allowed) of at most 4096 bytes, and bytes(X) is the byte string that X
encodes. (An operand with anything else, a line feed included, never
gets here: the anchor is not strict, <xref target="triage"/>.) Any other op, a
missing or malformed operand, a non-object entry, or current bytes
longer than 65536 at any step, fails the proof. The proof holds if the
final bytes equal the merkle root of the header at the record's
height (<xref target="view-verify"/>). If it fails: underspecified reason
<spanx style="verb">anchor_proof_invalid</spanx>.</t>

<t>The operation set resembles a subset of the operations used by
OpenTimestamps <xref target="OPENTIMESTAMPS"/>; only the operations above are
defined here.</t>

</section>
<section anchor="batch-leg"><name>The Batch Leg Rule</name>

<t>A ledger lists records by their hex sha inside a batch, and stamps
the batch. A proof can pass through such a batch with <spanx style="verb">hexlify</spanx>. A hex
digest that merely appears somewhere in stamped bytes is not a
listing. So:</t>

<t><list style="numbers" type="1">
  <t>If no op in the proof is <spanx style="verb">hexlify</spanx>, the rule accepts.</t>
  <t>If any <spanx style="verb">hexlify</spanx> is not the first op, or there is more than one,
refuse (<spanx style="verb">hexlify_not_leading</spanx>).</t>
  <t>Rebuild the batch: walk the ops after the first one, up to the
first <spanx style="verb">sha256</spanx>. Each must be <spanx style="verb">append</spanx> or <spanx style="verb">prepend</spanx>, else refuse
(<spanx style="verb">batch_leg_malformed</spanx>). If there is no <spanx style="verb">sha256</spanx>, refuse
(<spanx style="verb">batch_leg_malformed</spanx>). The batch is (all prepend operands, the
later one in front) + the ASCII lowercase hex of the body digest +
(all append operands, in order).</t>
  <t>The batch bytes must be UTF-8 text that parses as strict JSON
(<xref target="reading-json"/>; so <spanx style="verb">NaN</spanx> is read as a number, vector
<spanx style="verb">synth_batch_bytes_with_nan</spanx>) into an object whose <spanx style="verb">schema</spanx> is one of
<spanx style="verb">nenrin-agreement-batch-v1</spanx> (typed), <spanx style="verb">nenrin-witness-batch-v1</spanx>,
<spanx style="verb">nenrin-trace-pin-batch-v0</spanx>. Else refuse (<spanx style="verb">batch_not_json</spanx> or
<spanx style="verb">batch_schema_unknown</spanx>).</t>
  <t><spanx style="verb">records</spanx> is a list of objects (else <spanx style="verb">batch_schema_unknown</spanx>). If
<spanx style="verb">count</spanx> is present it equals the length of <spanx style="verb">records</spanx> (else
<spanx style="verb">batch_count_mismatch</spanx>).</t>
  <t>Exactly one <spanx style="verb">records[i]</spanx> has <spanx style="verb">sha</spanx> equal to the hex digest. None:
<spanx style="verb">record_not_listed_in_batch</spanx>. More: <spanx style="verb">record_listed_twice</spanx>.</t>
  <t>For a typed batch: the record's schema maps to a kind
(<spanx style="verb">a2a-execution-v0</spanx> to <spanx style="verb">execution</spanx>; revocations and
acknowledgements have no kind, so they are refused with
<spanx style="verb">record_kind_unknown</spanx>, WRITTEN BEHAVIOUR). <spanx style="verb">records[i].kind</spanx> equals
that kind (else <spanx style="verb">record_kind_mismatch</spanx>). If <spanx style="verb">records[i].schema</spanx> is
present it equals the record's schema (else
<spanx style="verb">record_schema_mismatch</spanx>).</t>
  <t>The string value of <spanx style="verb">records[i].sha</spanx> must be written in the raw
text as the 64 hex characters themselves, with no escapes (else
refuse, code <spanx style="verb">internal</spanx>), and its first character must sit at byte
offset equal to the total length of the prepend operands (else
refuse, <spanx style="verb">batch_leg_splice_not_at_listing</spanx>).</t>
</list></t>

<t>If the rule refuses: underspecified reason <spanx style="verb">anchor_batch_leg_refused</spanx>.
The refusal code is informative detail.</t>

<t>If all of <xref target="anchor-block"/> to <xref target="batch-leg"/> pass, the record is
accepted (disposition <spanx style="verb">accepted</spanx>).</t>

</section>
</section>
<section anchor="event-set"><name>The Event Set and Finality</name>

<section anchor="event-order"><name>Order</name>

<t>The event set is the accepted records sorted by (height, record_sha256
as a lowercase hex string). Arrival order plays no part.</t>

</section>
<section anchor="event-kinds"><name>Kinds</name>

<t>Each event has a kind: <spanx style="verb">execution</spanx>, <spanx style="verb">revocation</spanx> or <spanx style="verb">revocation_ack</spanx>,
by schema.</t>

</section>
<section anchor="same-block"><name>Same Block</name>

<t>Two records in the same block are not ordered by their bytes for any
rule that matters. <xref target="grant"/> says how same-block cases are decided.
The order of <xref target="event-order"/> only fixes the listing and the order in
which single use approvals are consumed (<xref target="approvals"/>).</t>

<t>That order is not a signed fact: the anchor is unsigned
(<xref target="anchor-what"/>) and an extra key in it passes <xref target="triage"/>, so whoever
holds the copies can change a record's record_sha256 without touching
its body or its proof. When two executions in one block compete for
one single use approval carried by an earlier record, the holder can
therefore choose which of the two is reported <spanx style="verb">conditional</spanx>. The
verdict, the deviation codes and the bond outcome do not change
(<xref target="sec-unsigned"/>).</t>

</section>
<section anchor="horizon"><name>Horizon</name>

<t>With tip T and <spanx style="verb">grant.finality.depth</spanx> k, the finality horizon is
<spanx style="verb">T - k + 1</spanx>.</t>

</section>
<section anchor="status"><name>Status</name>

<t><list style="symbols">
  <t>If the verdict is <spanx style="verb">underspecified</spanx>: <spanx style="verb">undetermined</spanx>.</t>
  <t>Otherwise <spanx style="verb">final</spanx> if every accepted record's height and every
orphaned record's height is at or below the horizon; else
<spanx style="verb">provisional</spanx>.</t>
</list></t>

<t>With no accepted and no orphaned record, the status is <spanx style="verb">final</spanx>.
Foreign, rejected, ignored and set aside records do not affect status.</t>

</section>
</section>
<section anchor="output"><name>Output</name>

<section anchor="core-fields"><name>Core Fields</name>

<t>A second implementation MUST reproduce these exactly. They are what
the vectors' expected.json files hold.</t>

<texttable title="Core fields">
      <ttcol align='left'>field</ttcol>
      <ttcol align='left'>value</ttcol>
      <c><spanx style="verb">contract_sha256</spanx></c>
      <c><xref target="digests"/>, computed even for a contract the door refuses; null when the contract is not a JSON object (vector <spanx style="verb">synth_contract_not_object</spanx>)</c>
      <c><spanx style="verb">verdict</spanx></c>
      <c><spanx style="verb">within_grant</spanx>, <spanx style="verb">deviation</spanx>, <spanx style="verb">nothing_to_settle</spanx> or <spanx style="verb">underspecified</spanx> (<xref target="verdict"/>)</c>
      <c><spanx style="verb">status</spanx></c>
      <c><spanx style="verb">final</spanx>, <spanx style="verb">provisional</spanx> or <spanx style="verb">undetermined</spanx> (<xref target="status"/>)</c>
      <c><spanx style="verb">bond_outcome</spanx></c>
      <c><xref target="bond-outcome"/></c>
      <c><spanx style="verb">deviation_codes</spanx></c>
      <c>the sorted set of deviation codes (<xref target="codes"/>). Empty when the verdict is not <spanx style="verb">deviation</spanx></c>
      <c><spanx style="verb">underspecified_reasons</spanx></c>
      <c>the sorted set of underspecified reasons (<xref target="codes"/>). Empty unless the verdict is <spanx style="verb">underspecified</spanx></c>
      <c><spanx style="verb">event_set</spanx></c>
      <c>the list of <xref target="event-order"/>, each item <spanx style="verb">{"kind", "sha256", "height"}</spanx> with sha256 the record_sha256. Empty when the walk did not run. A reason from <xref target="strict-terms"/>, rule 9 of <xref target="view-verify"/>, <xref target="triage"/>, <xref target="anchor-block"/> (below the lower bound) or <xref target="nenrin"/> (a cited walk) does not empty it</c>
      <c><spanx style="verb">dispositions</spanx></c>
      <c>an object mapping record_sha256 to one code, for every record the computation classified (<xref target="dispositions"/>)</c>
</texttable>

<t><spanx style="verb">deviation_codes</spanx> and <spanx style="verb">underspecified_reasons</spanx> are sets. The reference
may list a code more than once (one per action); that multiplicity is
informative.</t>

</section>
<section anchor="dispositions"><name>Dispositions</name>

<texttable title="Dispositions">
      <ttcol align='left'>code</ttcol>
      <ttcol align='left'>meaning</ttcol>
      <c><spanx style="verb">foreign</spanx>, <spanx style="verb">inconsistent</spanx>, <spanx style="verb">unbound</spanx></c>
      <c><xref target="binding"/></c>
      <c><spanx style="verb">record_not_strict</spanx>, <spanx style="verb">signature_not_strict</spanx>, <spanx style="verb">anchor_not_strict</spanx>, <spanx style="verb">revocation_not_provided_for</spanx></c>
      <c><xref target="triage"/>, set aside, or (for <spanx style="verb">record_not_strict</spanx> and <spanx style="verb">anchor_not_strict</spanx>) failing closed with the same reason added to the settlement</c>
      <c><spanx style="verb">duplicate_dropped</spanx></c>
      <c><xref target="collapse"/></c>
      <c><spanx style="verb">rejected</spanx></c>
      <c><xref target="triage"/> or <xref target="conformance"/></c>
      <c><spanx style="verb">ignored</spanx></c>
      <c><xref target="authentication"/>, a revocation whose <spanx style="verb">revoked_by</spanx> is not <spanx style="verb">"principal"</spanx></c>
      <c><spanx style="verb">anchor_incomplete</spanx>, <spanx style="verb">anchor_beyond_observed_chain</spanx>, <spanx style="verb">anchor_proof_invalid</spanx>, <spanx style="verb">anchor_batch_leg_refused</spanx></c>
      <c><xref target="anchor-block"/> to <xref target="batch-leg"/>, with the same reason added to the settlement</c>
      <c><spanx style="verb">anchor_below_lower_bound</spanx></c>
      <c><xref target="anchor-block"/>, with the same reason added</c>
      <c><spanx style="verb">orphaned</spanx></c>
      <c><xref target="anchor-block"/></c>
      <c><spanx style="verb">accepted</spanx></c>
      <c>in the event set</c>
</texttable>

<t>Every bound record is triaged (<xref target="triage"/>) whatever <xref target="contract"/> and
<xref target="view"/> find, so a record that <xref target="triage"/> rejects or sets aside always
has its disposition, and so does a copy that <xref target="collapse"/> drops
whenever <xref target="collapse"/> runs (<xref target="computation"/>). A bound record kept by
<xref target="triage"/> that the computation examined no further, because
<xref target="contract-door"/>, <xref target="settleable-terms"/>, rules 1 to 8 of
<xref target="view-verify"/> or <xref target="size-limit"/> failed, has no disposition
(revision 0.9). A
record that is not ours (<xref target="binding"/>, dropped silently) has no
disposition. A record set aside because it cannot be digested has no
record_sha256 and so no disposition; it is counted in the informative
<spanx style="verb">v1_16.refused</spanx>. Every record has at most one disposition.</t>

</section>
<section anchor="bond-outcome"><name>Bond Outcome</name>

<t>If <spanx style="verb">contract.bond</spanx> is absent or null: <spanx style="verb">n/a</spanx>. A bond of any other form
than an object with a positive integer <spanx style="verb">amount</spanx> makes the settlement
underspecified (<xref target="strict-terms"/>) and the outcome is then
<spanx style="verb">undetermined</spanx> when the value is one the lower layers read as a bond
(an object with a truthy <spanx style="verb">amount</spanx>), else <spanx style="verb">n/a</spanx>.</t>

<t>For a well formed bond: <spanx style="verb">undetermined</spanx> if the verdict is
<spanx style="verb">underspecified</spanx>; else <spanx style="verb">pending_finality</spanx> if the status is
<spanx style="verb">provisional</spanx>; else <spanx style="verb">held</spanx> for <spanx style="verb">within_grant</spanx> and
<spanx style="verb">nothing_to_settle</spanx>, and <spanx style="verb">forfeited</spanx> for <spanx style="verb">deviation</spanx>.</t>

</section>
<section anchor="informative-fields"><name>Informative Fields</name>

<t>Everything else the reference writes is informative: <spanx style="verb">schema</spanx>,
<spanx style="verb">settled_under</spanx>, <spanx style="verb">v1_16</spanx>, <spanx style="verb">signed_body_collapse</spanx>, <spanx style="verb">contract_id</spanx>,
<spanx style="verb">binding_mode</spanx>, <spanx style="verb">bound_by_label_only</spanx>, <spanx style="verb">ordering_rule</spanx>,
<spanx style="verb">approval_rule</spanx>, <spanx style="verb">anchor_rule</spanx>, <spanx style="verb">view_rules</spanx>, <spanx style="verb">chain_tip</spanx>,
<spanx style="verb">finality_horizon</spanx>, <spanx style="verb">pinned_blocks</spanx>, <spanx style="verb">revocations</spanx>, <spanx style="verb">approvals</spanx>,
<spanx style="verb">binding</spanx>, <spanx style="verb">orphaned</spanx>, <spanx style="verb">predates_contract</spanx>, <spanx style="verb">ignored</spanx>, <spanx style="verb">rejected</spanx>,
<spanx style="verb">duplicate_anchorings</spanx>, <spanx style="verb">establishes</spanx>, <spanx style="verb">does_not_establish</spanx>,
<spanx style="verb">spine_gate</spanx>, <spanx style="verb">convergence_gate</spanx>, <spanx style="verb">approval_gate</spanx>, <spanx style="verb">signatures</spanx>, the
full <spanx style="verb">underspecified</spanx> and <spanx style="verb">deviations</spanx> objects (heights, <spanx style="verb">why</spanx> texts,
details), and their order. The informative fields are useful when
comparing implementations. They are not part of conformance.</t>

</section>
</section>
<section anchor="grant"><name>Grant Evaluation</name>

<t>The walk runs over the event set (<xref target="event-order"/>) only. Let
<spanx style="verb">g = contract.grant</spanx>.</t>

<section anchor="revocation"><name>Revocation</name>

<t>For each accepted revocation at height h, its effective height E is:</t>

<t><list style="symbols">
  <t><spanx style="verb">effective_at</spanx> <spanx style="verb">"anchor"</spanx>: E = h.</t>
  <t><spanx style="verb">effective_at</spanx> <spanx style="verb">"delivery_ack"</spanx>: E = the smallest of h +
<spanx style="verb">ack_window</spanx> and the heights of every accepted acknowledgement whose
<spanx style="verb">revocation_sha256</spanx> names this revocation.</t>
</list></t>

<t>An acknowledgement names a revocation if its <spanx style="verb">revocation_sha256</spanx>
equals the lowercase hex body digest of the revocation, or the
record_sha256 of a copy of that revocation that is still in the input
after <xref target="collapse"/>. A copy <xref target="collapse"/> dropped no longer names the
revocation (WRITTEN BEHAVIOUR: settle v1.4 collects every anchored
copy, but since settle v1.15 the copies of one signed body are reduced
before v1.4 sees them). Vector <spanx style="verb">synth_ack_names_dropped_copy</spanx>.
An acknowledgement anchored
before its revocation makes E earlier than h (WRITTEN BEHAVIOUR, see
item N5 of <xref target="MUSUBI-NOTES"/>).</t>

<t>Sort the revocations by (E, record_sha256). An execution at height h
is revoked if some revocation has E &lt;= h. Authority ends at block E
inclusive: an act in the same block as an anchor-mode revocation is
revoked. For a revoked execution, every string in <spanx style="verb">performed_actions</spanx>
gives the deviation <spanx style="verb">revoked</spanx>, and nothing else in that record is
judged (no expiry, prohibited, conditional, unauthorized or
delegation check, and no approval nonce is consumed).</t>

<t>If the grant has no <spanx style="verb">revocation</spanx> object, a revocation never gets
here: it is set aside in <xref target="triage"/> (<spanx style="verb">revocation_not_provided_for</spanx>).</t>

</section>
<section anchor="expiry"><name>Expiry</name>

<t>If <spanx style="verb">g.expiry_height</spanx> is present and not null, every string action of
an execution with height h &gt; expiry_height gives <spanx style="verb">after_expiry</spanx>. This
does not stop the other checks for that action: an expired prohibited
action gives both <spanx style="verb">after_expiry</spanx> and <spanx style="verb">prohibited</spanx>. An execution at
exactly expiry_height is not expired.</t>

</section>
<section anchor="actions"><name>Actions and Delegation</name>

<t>For each accepted, unrevoked execution, for each string <spanx style="verb">a</spanx> in
<spanx style="verb">performed_actions</spanx>, in order, after <xref target="expiry"/>:</t>

<t><list style="numbers" type="1">
  <t>If <spanx style="verb">a</spanx> is in <spanx style="verb">g.prohibited_actions</spanx>: deviation <spanx style="verb">prohibited</spanx>. Next
action.</t>
  <t>Else if <spanx style="verb">a</spanx> is the <spanx style="verb">action</spanx> of some <spanx style="verb">g.conditional</spanx> entry: apply
<xref target="approvals"/>. Next action. (A conditional action does not need to
be in <spanx style="verb">authorized_actions</spanx>.)</t>
  <t>Else if <spanx style="verb">a</spanx> is not in <spanx style="verb">g.authorized_actions</spanx>: deviation
<spanx style="verb">unauthorized</spanx>.</t>
</list></t>

<t>Every entry of <spanx style="verb">performed_actions</spanx> and <spanx style="verb">delegated_to</spanx> is a string here
(<xref target="triage"/>). Comparison is exact string equality. There is no case
folding or Unicode normalization.</t>

<t>Then, for each entry <spanx style="verb">d</spanx> of <spanx style="verb">delegated_to</spanx> not in
<spanx style="verb">g.delegation.allowed</spanx> (empty when absent): deviation <spanx style="verb">delegation</spanx>. An
outcome (<spanx style="verb">outcomes</spanx>) never excuses an act. An attempt that errored is
judged like one that succeeded.</t>

</section>
<section anchor="approvals"><name>Conditional Actions and Approvals</name>

<t>Every approval carried by an accepted execution (revoked or not) is
classified. With <spanx style="verb">pk</spanx> the principal's key and <spanx style="verb">csha</spanx> this contract's
contract_sha256:</t>

<t><list style="numbers" type="1">
  <t>Not an object, or no principal key: <spanx style="verb">forged</spanx> (<xref target="forged"/>). (Once
<xref target="contract-door"/> accepts, the principal always has a usable key,
so in core v0 only "not an object" can apply.)</t>
  <t>If <spanx style="verb">valid_until_height</spanx> is an integer &gt;= 0, <spanx style="verb">nonce</spanx> is a string of
32 lowercase hex characters (a nonce with a line feed never gets
here, <xref target="triage"/>), <spanx style="verb">single_use</spanx> is a boolean, and <spanx style="verb">contract_sha256</spanx>
is 64 lowercase hex: compute the v2 approval bytes shown after this
list. If <spanx style="verb">sig_b64</spanx> verifies over them with <spanx style="verb">pk</spanx>: <spanx style="verb">counted</spanx> when its
contract_sha256 equals csha, else <spanx style="verb">other_terms</spanx>.</t>
  <t>Otherwise, if <spanx style="verb">sig_b64</spanx> verifies with <spanx style="verb">pk</spanx> over either older form,
it is <spanx style="verb">label_bound</spanx>:
  <list style="symbols">
      <t>the v1 approval bytes, with contract_id and payload_digest from
the contract, only when the field checks of step 2 other than
contract_sha256 hold, or</t>
      <t>the v0 approval bytes.</t>
    </list></t>
  <t>Otherwise <spanx style="verb">forged</spanx>.</t>
</list></t>

<t>The three forms of approval bytes are:</t>

<figure title="Approval signing bytes"><artwork><![CDATA[
v2: "a2a-approval-v2\n" + canonical({
      "contract_sha256": <its contract_sha256>,
      "action": <its action>,
      "valid_until_height": ..,
      "nonce": ..,
      "single_use": ..})

v1: "a2a-approval-v1\n" + canonical({
      "contract_id", "payload_digest", "action",
      "valid_until_height", "nonce", "single_use"})

v0: "a2a-approval-v0\n" + canonical({
      "contract_id", "payload_digest", "action"})
]]></artwork></figure>

<t>A missing field is canonicalized as null. The approval's <spanx style="verb">by</spanx> field
is not read. Only <spanx style="verb">sig_b64</spanx> is verified. An approval whose signature
is only under another key, such as <spanx style="verb">signature</spanx>, is <spanx style="verb">forged</spanx>
(<xref target="forged"/>), even when that value would verify. <spanx style="verb">label_bound</spanx> and <spanx style="verb">other_terms</spanx> approvals are listed
(informative) and never count. They are not deviations.</t>

<t>A conditional action <spanx style="verb">a</spanx> in an execution at height h with record sha
s is approved if some counted approval, taken in event set order
(records by <xref target="event-order"/>, approvals in list order within a record),
satisfies all of:</t>

<t><list style="symbols">
  <t>its <spanx style="verb">action</spanx> equals <spanx style="verb">a</spanx>,</t>
  <t>it is carried by the same record s, or by a record at a height
strictly below h (a record in the same block does not count),</t>
  <t>h &lt;= its <spanx style="verb">valid_until_height</spanx>,</t>
  <t>if its <spanx style="verb">single_use</spanx> is true, its <spanx style="verb">nonce</spanx> has not been consumed yet.</t>
</list></t>

<t>The first approval that satisfies all four is used. If it is single
use, its nonce is consumed. The set of consumed nonces is shared
across all actions and all executions (see item N5 of
<xref target="MUSUBI-NOTES"/>). If no approval satisfies all four: deviation
<spanx style="verb">conditional</spanx>. The value of <spanx style="verb">requires</spanx> in the conditional entry is
not read: every conditional action needs a counted approval.</t>

</section>
<section anchor="forged"><name>Forged Approvals</name>

<t>Every approval classified <spanx style="verb">forged</spanx> adds the deviation
<spanx style="verb">forged_approval</spanx>, whatever its action, whether or not the action was
needed.</t>

</section>
</section>
<section anchor="nenrin"><name>NENRIN Checks</name>

<section anchor="nenrin-cross"><name>Cross Check</name>

<t>Only when <spanx style="verb">nenrin_records</spanx> is a non-empty list. A walk that is not
strict (first paragraph of <xref target="triage"/>) is left out. Index every other
object in it by its record_sha256 (<xref target="digests"/>). For each execution
kept by <xref target="triage"/> and not rejected by <xref target="conformance"/>, left after
<xref target="collapse"/> (once per record_sha256), whatever <xref target="anchors"/> then
decides about its anchor (an orphaned execution is still read:
WRITTEN BEHAVIOUR, vector <spanx style="verb">synth_cited_walk_orphaned_execution</spanx>), look
up the walk whose
record_sha256 equals the execution's <spanx style="verb">nenrin_ref</spanx>. If found, and the
walk's <spanx style="verb">context</spanx> is an object whose <spanx style="verb">contract_sha256</spanx> is present, not
null, and not equal to csha: deviation
<spanx style="verb">evidence_names_other_contract</spanx>. If an execution's <spanx style="verb">nenrin_ref</spanx> names
a walk that was left out for not being strict: underspecified reason
<spanx style="verb">walk_not_strict</spanx>. The contractor chose what to cite, and leaving the
walk out silently could hide a deviation.</t>

<t>Records that were rejected or set aside are never checked here (item
N2 of <xref target="MUSUBI-NOTES"/>).</t>

</section>
<section anchor="nenrin-required"><name>Required Evidence</name>

<t>When <spanx style="verb">requirements.evidence</spanx> is <spanx style="verb">"nenrin_required"</spanx>, each accepted
execution whose <spanx style="verb">nenrin_ref</spanx> is not 64 lowercase hex gives the
deviation <spanx style="verb">evidence_missing</spanx>. This is applied after <xref target="verdict"/> found
<spanx style="verb">within_grant</spanx> or <spanx style="verb">deviation</spanx>.</t>

</section>
</section>
<section anchor="verdict"><name>Verdict</name>

<t><list style="symbols">
  <t><spanx style="verb">underspecified</spanx> if any underspecified reason was found
(<xref target="contract"/>, <xref target="view"/>, <xref target="triage"/>, <xref target="size-limit"/>, <xref target="anchors"/>,
<xref target="nenrin"/>). Then <spanx style="verb">deviations</spanx> is empty and status is
<spanx style="verb">undetermined</spanx>.</t>
  <t>Else <spanx style="verb">deviation</spanx> if any deviation was found (<xref target="conformance"/>,
<xref target="grant"/>, <xref target="nenrin"/>).</t>
  <t>Else <spanx style="verb">nothing_to_settle</spanx> if the event set is empty: no record was
read, so nothing was found within the grant or outside it. Status is
as <xref target="status"/> gives it (<spanx style="verb">final</spanx> with no orphaned record).</t>
  <t>Else <spanx style="verb">within_grant</spanx>.</t>
</list></t>

<t><spanx style="verb">within_grant</spanx> covers only the accepted records. One strictly signed
record whose anchor fails <xref target="anchor-block"/> (first, second or third
row), <xref target="anchor-proof"/> or <xref target="batch-leg"/> makes the whole settlement
<spanx style="verb">underspecified</spanx>. A record that is not strictly signed never does.</t>

</section>
<section anchor="written-behaviour"><name>Written Behaviour and Changes from Earlier Settle Versions</name>

<section anchor="wb-list"><name>Written Behaviour</name>

<t>These are what the reference still does. They are normative for core
v0 because the vectors pin them, or would.</t>

<texttable title="Written behaviour">
      <ttcol align='left'>item</ttcol>
      <ttcol align='left'>vector</ttcol>
      <c>An extra bad signature from a party refuses the contract even when that party also signed correctly</c>
      <c>none</c>
      <c>Revocations and acknowledgements cannot pass a typed batch leg (no kind)</c>
      <c>none</c>
      <c>An acknowledgement anchored before its revocation ends authority at the acknowledgement's height</c>
      <c>none</c>
      <c>An orphaned record is excluded without making the settlement underspecified, yet its claimed height still counts for status</c>
      <c><spanx style="verb">real_p31_anchor_names_other_height</spanx></c>
      <c>A header ending in a line feed passes rules 1 to 8 of <xref target="view-verify"/> and is refused only by rule 9 (<spanx style="verb">view_not_strict</spanx>); one in upper case is refused by both</c>
      <c><spanx style="verb">synth_view_header_trailing_newline</spanx></c>
      <c>A party-signed record that names this contract_sha256 with another contract_id or payload_digest is inconsistent (<xref target="binding"/>) and is never triaged, so a deviation it shows is not read</c>
      <c>none</c>
</texttable>

</section>
<section anchor="changes"><name>Changes in Settle v1.16</name>

<t>Each item below was written behaviour of settle v1.10 to v1.15;
<xref target="MUSUBI-NOTES"/> has the details under the item numbers given.</t>

<texttable title="Changes in settle v1.16">
      <ttcol align='left'>item</ttcol>
      <ttcol align='left'>before</ttcol>
      <ttcol align='left'>now</ttcol>
      <c>N2</c>
      <c>The NENRIN cross check read every bound execution, also rejected and unsigned ones: anyone could add a charge and forfeit a bond</c>
      <c>only records strictly signed and not rejected are read by any check</c>
      <c>N1</c>
      <c>A trailing line feed passed in block_hash, proof operands, headers, action names and approval nonces</c>
      <c>strict forms (<xref target="triage"/>, rule 9 of <xref target="view-verify"/>)</c>
      <c>N3, N6</c>
      <c><spanx style="verb">delegated_to</spanx> and <spanx style="verb">approvals</spanx> were not type checked; an object in <spanx style="verb">delegated_to</spanx> stopped the reference</c>
      <c><xref target="triage"/></c>
      <c>N4</c>
      <c>Floats and unsafe integers in records, views and walks were accepted</c>
      <c><xref target="triage"/>, rule 9 of <xref target="view-verify"/>, <xref target="nenrin-cross"/></c>
      <c>N6</c>
      <c>A revocation against a grant with no revocation block stopped the reference</c>
      <c>set aside</c>
      <c>N8</c>
      <c>The bond test used truthiness, so <spanx style="verb">[]</spanx> was no bond in Python and a bond in JavaScript</c>
      <c><spanx style="verb">bond_malformed</spanx></c>
      <c>N9</c>
      <c>Signatures were checked by OpenSSL alone</c>
      <c>strict signatures (<xref target="ed25519"/>)</c>
      <c>N10</c>
      <c>A copy with a junk signatures entry was a second body that blocked the verdict (fixed in v1.15)</c>
      <c>copies of one signed body collapse (<xref target="collapse"/>)</c>
      <c>N11</c>
      <c>An anchor below the checkpoint was dropped without a word; <spanx style="verb">parent_contract</spanx> and <spanx style="verb">requirements.evidence</spanx> were never read; an empty event set was <spanx style="verb">within_grant</spanx></c>
      <c><spanx style="verb">anchor_below_lower_bound</spanx>; <xref target="strict-terms"/>; <xref target="nenrin-required"/>; <spanx style="verb">nothing_to_settle</spanx></c>
</texttable>

<t>The vectors that pin each change are:</t>

<texttable title="Vectors for the changes in settle v1.16">
      <ttcol align='left'>item</ttcol>
      <ttcol align='left'>vectors</ttcol>
      <c>N2</c>
      <c><spanx style="verb">synth_nenrin_check_reads_rejected_record</spanx>, <spanx style="verb">synth_unsigned_copies_anchored_first</spanx></c>
      <c>N1</c>
      <c><spanx style="verb">real_p31_proof_operand_trailing_newline</spanx>, <spanx style="verb">real_p31_proof_operand_trailing_newline_beside_good</spanx>, <spanx style="verb">real_p31_block_hash_trailing_newline</spanx>, <spanx style="verb">synth_action_name_trailing_newline</spanx>, <spanx style="verb">synth_approval_nonce_trailing_newline</spanx>, <spanx style="verb">synth_view_header_trailing_newline</spanx></c>
      <c>N3, N6</c>
      <c><spanx style="verb">synth_delegated_to_object</spanx></c>
      <c>N4</c>
      <c><spanx style="verb">synth_float_in_signed_body</spanx>, <spanx style="verb">synth_unsafe_integer_in_signed_body</spanx>, <spanx style="verb">synth_unsigned_float_beside_good</spanx>, <spanx style="verb">synth_cited_walk_not_strict</spanx></c>
      <c>N6</c>
      <c><spanx style="verb">synth_revocation_irrevocable_grant</spanx></c>
      <c>N8</c>
      <c><spanx style="verb">synth_bond_empty_list</spanx>, <spanx style="verb">synth_bond_amount_string</spanx></c>
      <c>N9</c>
      <c><spanx style="verb">synth_signature_s_plus_l</spanx>, <spanx style="verb">synth_strict_signature_beside_malleated</spanx>, <spanx style="verb">real_p31_contract_bad_principal_signature</spanx></c>
      <c>N10</c>
      <c><spanx style="verb">real_p31_copy_with_junk_signature_entry</spanx></c>
      <c>N11</c>
      <c><spanx style="verb">real_p31_anchor_below_checkpoint</spanx>, <spanx style="verb">synth_parent_not_supplied</spanx>, <spanx style="verb">synth_parent_supplied_narrows</spanx>, <spanx style="verb">synth_parent_widened</spanx>, <spanx style="verb">synth_nenrin_required_without_ref</spanx>, <spanx style="verb">synth_nothing_to_settle</spanx></c>
</texttable>

</section>
<section anchor="clarifications"><name>Clarifications in Revisions 0.1 to 0.8 of the Source Text</name>

<t>None of the revisions below changed the reference. Each says what the
reference already did where the text had been silent, unclear or
wrong.</t>

<t>Revision 0.1. The first separately written implementation
(<xref target="impl-others"/>, <xref target="ISSUE-49"/>) asked which reading was meant
(item N13 of <xref target="MUSUBI-NOTES"/>). For each, what a reader could take
the text to mean, what the reference does, and the vector:</t>

<t><list style="symbols">
  <t>An approval signed only under <spanx style="verb">signature</spanx> counts. The reference
verifies only <spanx style="verb">sig_b64</spanx>, so such an approval is <spanx style="verb">forged</spanx>
(<xref target="approvals"/>, <xref target="forged"/>). Vector
<spanx style="verb">synth_approval_signature_key_only</spanx>.</t>
  <t>A strict <spanx style="verb">sig_b64</spanx> makes a malleated <spanx style="verb">signature</spanx> beside it harmless.
The reference checks each key for strictness on its own; the body is
not strict (<xref target="triage"/>). Vector
<spanx style="verb">synth_approval_lax_signature_beside_good_sig_b64</spanx>.</t>
  <t>(Control.) <spanx style="verb">sig_b64</spanx> and <spanx style="verb">signature</spanx> with the same strict value: the
approval is counted. Vector
<spanx style="verb">synth_approval_signature_key_same_value</spanx>.</t>
</list></t>

<t>Revision 0.2, from the 25 implementation notes of the same
implementation (item N14 of <xref target="MUSUBI-NOTES"/>):</t>

<texttable title="Clarified in revision 0.2">
      <ttcol align='left'>note</ttcol>
      <ttcol align='left'>what the text now says</ttcol>
      <ttcol align='left'>vector</ttcol>
      <c>A1</c>
      <c><spanx style="verb">contract_sha256</spanx> is null when the contract is not a JSON object (<xref target="core-fields"/>)</c>
      <c><spanx style="verb">synth_contract_not_object</spanx></c>
      <c>A3</c>
      <c>an integer token longer than 4300 digits stops the computation, no settlement (<xref target="reading-json"/>)</c>
      <c>none</c>
      <c>A6</c>
      <c>domain normalization strips all leading and trailing whitespace (rule 4 of <xref target="contract-door"/>)</c>
      <c><spanx style="verb">synth_domain_leading_tab</spanx></c>
      <c>A7</c>
      <c><spanx style="verb">delegation</spanx> must carry <spanx style="verb">allowed</spanx>; <spanx style="verb">ordering</spanx> may be <spanx style="verb">{}</spanx> (rule 10 of <xref target="contract-door"/>)</c>
      <c><spanx style="verb">synth_delegation_empty_object</spanx>, <spanx style="verb">synth_ordering_empty_object</spanx></c>
      <c>A10</c>
      <c>parent narrowing stated in full (<xref target="parent-narrowing"/>)</c>
      <c><spanx style="verb">synth_parent_full_*</spanx> (25)</c>
      <c>A12, A15</c>
      <c><spanx style="verb">NaN</spanx>, <spanx style="verb">Infinity</spanx>, <spanx style="verb">-Infinity</spanx> are read as numbers, in records and in batch bytes (<xref target="reading-json"/>, <xref target="batch-leg"/>)</c>
      <c><spanx style="verb">synth_nan_in_signed_body</spanx>, <spanx style="verb">synth_batch_bytes_with_nan</spanx></c>
      <c>A14</c>
      <c>the cross check reads an execution whatever its anchor result (<xref target="nenrin-cross"/>)</c>
      <c><spanx style="verb">synth_cited_walk_orphaned_execution</spanx></c>
      <c>A16</c>
      <c><spanx style="verb">view_not_strict</spanx> only for a view that is an object with a <spanx style="verb">headers</spanx> list (rule 9 of <xref target="view-verify"/>)</c>
      <c><spanx style="verb">synth_view_not_an_object</spanx>, <spanx style="verb">synth_view_headers_not_a_list</spanx></c>
      <c>A20</c>
      <c>what "carries" means for each out of scope key (<xref target="scope"/>)</c>
      <c><spanx style="verb">synth_limits_null</spanx>, <spanx style="verb">synth_admission_null</spanx></c>
      <c>A21</c>
      <c>a floor whose mantissa is zero decodes to no target: both the door and the view refuse</c>
      <c><spanx style="verb">synth_floor_zero_mantissa</spanx></c>
      <c>A22</c>
      <c>a copy dropped by <xref target="collapse"/> no longer names its revocation (<xref target="revocation"/>)</c>
      <c><spanx style="verb">synth_ack_names_dropped_copy</spanx></c>
      <c>A24</c>
      <c>step 1 "no principal key" of <xref target="approvals"/> cannot apply once <xref target="contract-door"/> accepts</c>
      <c>none</c>
</texttable>

<t>Revision 0.3, from 10 more notes of the same implementation (item N15
of <xref target="MUSUBI-NOTES"/>):</t>

<texttable title="Clarified in revision 0.3">
      <ttcol align='left'>note</ttcol>
      <ttcol align='left'>what the text now says</ttcol>
      <ttcol align='left'>vector</ttcol>
      <c>B2</c>
      <c>non-finite numbers are written <spanx style="verb">NaN</spanx>, <spanx style="verb">Infinity</spanx>, <spanx style="verb">-Infinity</spanx> in canonical form (item 6 of <xref target="canonical"/>)</c>
      <c><spanx style="verb">synth_infinity_in_signed_body</spanx>, <spanx style="verb">synth_negative_infinity_in_signed_body</spanx></c>
      <c>B7</c>
      <c>only a well formed approver list is carried; <spanx style="verb">{}</spanx>, <spanx style="verb">""</spanx>, <spanx style="verb">0</spanx>, null and <spanx style="verb">[]</spanx> refuse in <xref target="contract-door"/> (<xref target="scope"/>, rule 10 of <xref target="contract-door"/>)</c>
      <c><spanx style="verb">synth_approvers_object</spanx>, <spanx style="verb">synth_approvers_string</spanx>, <spanx style="verb">synth_approvers_zero</spanx></c>
      <c>B5</c>
      <c>rule 14 of <xref target="parent-narrowing"/> applies when the parent pins approvers and the child has none</c>
      <c><spanx style="verb">synth_parent_approvers_child_drops</spanx></c>
      <c>C3</c>
      <c>rule 14 of <xref target="parent-narrowing"/> compares each approver entry as canonical bytes, so the order of <spanx style="verb">actions</spanx> inside an entry counts; core v0 cannot reach it</c>
      <c>none (out of scope)</c>
      <c>B8</c>
      <c>exactly three spellings; <spanx style="verb">nan</spanx> and <spanx style="verb">+Infinity</spanx> stop the reader; <spanx style="verb">1e400</spanx> is an infinity (<xref target="reading-json"/>)</c>
      <c>none (no settlement to compare)</c>
</texttable>

<t>Revision 0.4 (item N17 of <xref target="MUSUBI-NOTES"/>) checked which
underspecified reasons the vectors actually produce. <xref target="exercise"/> has
the result.</t>

<t>Revisions 0.5 to 0.8:</t>

<t><list style="symbols">
  <t>0.5 (item N18): two statements about the reference were wrong. The
file list named the copy of agreement_verify.py vendored into
another package instead of the copy the reference imports (this
document already gave the right one), and the source text said
settle_v1_15.py gives the same core fields outside a listed set of
vectors, which it does not (<xref target="not-established"/>).</t>
  <t>0.6 (item N20): a strictly signed record whose body is not strict is
set aside as well as adding its reason (<xref target="triage"/>); failing
<xref target="contract-door"/> with the parent supplied adds only
<spanx style="verb">parent_narrowing_failed</spanx> (<xref target="strict-terms"/>). Both say what the
vectors already expected.</t>
  <t>0.7 (item N21): the same-block order is not a signed fact
(<xref target="same-block"/>, <xref target="sec-unsigned"/>).</t>
  <t>0.8 (item N22): two files the reference imports changed after
revision 0.4, in layers outside settle core; on the 128 vectors the
complete output of the reference is the same before and after. The
file list now names its commit (<xref target="impl-ref"/>).</t>
</list></t>

</section>
<section anchor="revisions-09-11"><name>Revisions 0.9 and 0.11 of the Source Text</name>

<t>Revision 0.9 (item N23 of <xref target="MUSUBI-NOTES"/>). A differential run of
5,128 inputs (5,000 generated and the 128 vectors) through the
reference and the two implementations of <xref target="impl-others"/> found that
the reference added <spanx style="verb">unscoped_approvals_refused</spanx> when
<xref target="contract-door"/> refused the contract, while <xref target="settleable-terms"/>
applies only when <xref target="contract-door"/> accepts. Both implementations
followed the text. The reference was fixed (settle_v1_16.py, whose
sha256 prefix changed from <spanx style="verb">ec332ecb387e9b56</spanx> to <spanx style="verb">a7a1621224a63ef2</spanx>).
On the 128 vectors its complete output, not only the core fields, is
the same before and after. The same run found that <xref target="computation"/>
and <xref target="dispositions"/> could be read as giving no bound record a
disposition when <xref target="contract"/> or <xref target="view"/> fails; both now state the
reading of the reference and the vectors: triage runs on every bound
record, and its dispositions are reported.</t>

<t>Revision 0.11 (items N25 and N26). A blind run in which programs were
written from the text and the vector inputs alone <xref target="ERRATA-K001"/>
found two places the text left open. The 8 hex characters of
<spanx style="verb">max_target_bits</spanx> are the 32 bit value written big-endian
(<xref target="contract-door"/>, <xref target="view-verify"/>, <xref target="compact-target"/>). A reason
added by <xref target="settleable-terms"/> stops the examination of records, as
<xref target="computation"/> already said; <xref target="settleable-terms"/> now says so
itself. No rule of the reference, reason or vector changed.</t>

</section>
</section>
<section anchor="vectors"><name>Test Vectors</name>

<section anchor="vectors-contents"><name>Contents</name>

<t>The vectors are published beside the source text of this document,
unchanged since commit dd64d0e2 (revision 0.4) and cited here at
commit e41715d7 <xref target="MUSUBI-VECTORS"/>, in the folder
settle-core-v0/vectors/ under
<eref target="https://github.com/ogasurfproject-jpg/horizon-shield/tree/e41715d75704e92455400707df895efd82bf4ecf/workers/hs-ledger/nenrin/musubi-v0/settle-core-v0">https://github.com/ogasurfproject-jpg/horizon-shield/tree/e41715d75704e92455400707df895efd82bf4ecf/workers/hs-ledger/nenrin/musubi-v0/settle-core-v0</eref>.
They are not reproduced here.</t>

<t>There are 128 vectors: 22 real and 106 synthetic. Their expected
verdicts are <spanx style="verb">within_grant</spanx> for 38, <spanx style="verb">deviation</spanx> for 21,
<spanx style="verb">nothing_to_settle</spanx> for 8 and <spanx style="verb">underspecified</spanx> for 61. Their expected
statuses are <spanx style="verb">final</spanx> for 65, <spanx style="verb">provisional</spanx> for 2 and <spanx style="verb">undetermined</spanx>
for 61.</t>

<t>The 76 vectors of the first release are unchanged, byte for byte.
Revision 0.1 of the source text added 3, revision 0.2 added 38,
revision 0.3 added 6 and revision 0.4 added 5 (<xref target="clarifications"/>,
<xref target="exercise"/>). All added vectors are synthetic.</t>

<t>The folder holds one subfolder per vector:</t>

<t><list style="symbols">
  <t><spanx style="verb">contract.json</spanx>, <spanx style="verb">view.json</spanx>, <spanx style="verb">events/eNN.json</spanx> (the events, in any
order), <spanx style="verb">nenrin/nNN.json</spanx> when the vector supplies walks,
<spanx style="verb">parent.json</spanx> when it supplies a parent contract,</t>
  <t><spanx style="verb">expected.json</spanx>: the core fields (<xref target="core-fields"/>).</t>
</list></t>

<t>The index vectors/vectors.json lists, for each vector, its id, its
family (real or synthetic), the section of the source text it tests,
a description, the expected verdict, status and code sets, and the
sha256 of each of its files. It also lists the four synthetic public
keys.</t>

<t>Real vectors use the real records, contracts and Bitcoin headers of
two settled engagements, with one change each (the record bodies
cannot change without breaking the real Bitcoin anchors).
<spanx style="verb">real_p31_baseline</spanx> is the first MUSUBI contract settled with no
party from the project that publishes the reference
(agent-passport.org and kuangmi-bit.github.io, Bitcoin block 970480),
with its 359 header view.</t>

<t>Synthetic vectors use keys named synthetic-principal,
synthetic-contractor, synthetic-witness and synthetic-stranger,
derived from fixed seeds, and a synthetic header chain at regtest
difficulty (<spanx style="verb">max_target_bits</spanx> <spanx style="verb">207fffff</spanx>, checkpoint height 800001).
The public keys are:</t>

<texttable title="Synthetic keys">
      <ttcol align='left'>name</ttcol>
      <ttcol align='left'>public_key_ed25519_b64</ttcol>
      <c>synthetic-principal</c>
      <c><spanx style="verb">rrXBDf5mF249t6wdmMvKtkcBXZmVVfcfYFBBjIpSttE=</spanx></c>
      <c>synthetic-contractor</c>
      <c><spanx style="verb">hl3xmxz86HFYXTczn/5HZw4Jkb3i0nL7iig/2GXZrcs=</spanx></c>
      <c>synthetic-witness</c>
      <c><spanx style="verb">gMKIYPPIjG6r9mt8ybr3dIGGm6LYxkngSym9ltvHNtY=</spanx></c>
      <c>synthetic-stranger</c>
      <c><spanx style="verb">YmzNu/v6K9wUOtcmnssHABVLBDuLBqMcZnM0NL+32rU=</spanx></c>
</texttable>

<t>gen_vectors.py regenerates every vector from the two real inputs and
the reference; it is deterministic. run_reference.py runs the
reference over every vector and checks every file hash and every core
field. exercise_inputs.py <xref target="MUSUBI-EXERCISE"/> builds in memory the
inputs the vector layout cannot carry (events that are not a list,
10001 records) and the inputs of the four rows of
<xref target="settleable-terms"/> that cannot fire, runs the reference on them,
and compares its lines with EXERCISE.txt (<spanx style="verb">--check</spanx>).</t>

</section>
<section anchor="projection"><name>The Core-Field Projection</name>

<t>expected.json holds exactly the eight core fields of <xref target="core-fields"/>
and nothing else. The reference's full settlement object is projected
onto them as follows:</t>

<t><list style="symbols">
  <t><spanx style="verb">contract_sha256</spanx>, <spanx style="verb">verdict</spanx>, <spanx style="verb">status</spanx> and <spanx style="verb">bond_outcome</spanx> are copied.</t>
  <t><spanx style="verb">deviation_codes</spanx> and <spanx style="verb">underspecified_reasons</spanx> are the sorted sets
of the codes in the reference's <spanx style="verb">deviations</spanx> and <spanx style="verb">underspecified</spanx>
lists. Repetitions are removed.</t>
  <t><spanx style="verb">event_set</spanx> keeps, for each accepted record in the order of
<xref target="event-order"/>, only <spanx style="verb">kind</spanx>, <spanx style="verb">sha256</spanx> and <spanx style="verb">height</spanx>.</t>
  <t><spanx style="verb">dispositions</spanx> is one object from record_sha256 to one disposition
code (<xref target="dispositions"/>), with keys sorted. A record that would
receive two different codes is an error of the projection.</t>
</list></t>

<t>An implementation conforms on a vector when its eight core fields are
equal to expected.json as JSON values. The informative fields
(<xref target="informative-fields"/>) are not compared.</t>

</section>
<section anchor="exercise"><name>Exercised and Unexercised Rejection Paths</name>

<t>Revision 0.4 of the source text checked every underspecified reason
against the expected results of the vectors (item N17 of
<xref target="MUSUBI-NOTES"/>). It was prompted by
<xref target="I-D.he-agentproto-verifier-evidence"/>, under which a rejecting path
counts as exercised only when an identified input drove it.</t>

<t>All ten deviation codes (<xref target="codes"/>) are produced by at least one
vector. Of the 28 underspecified reasons, the 123 vectors of revision
0.3 produced 18 and did not produce 10. Each of the ten was then
driven through the reference. For each, what was missing, and the
vector or exercise that now drives it:</t>

<t><list style="symbols">
  <t><spanx style="verb">grant_action_name_invalid</spanx>: no vector; <xref target="contract-door"/> accepts any
non-empty string, and <xref target="settleable-terms"/> applies the name pattern.
Vector <spanx style="verb">synth_grant_action_name_invalid</spanx>.</t>
  <t><spanx style="verb">too_many_witnesses</spanx>: no vector; 33 witnesses refuse, 32 are
accepted. Vectors <spanx style="verb">synth_too_many_witnesses</spanx> and
<spanx style="verb">synth_witnesses_32</spanx>.</t>
  <t><spanx style="verb">ack_window_missing</spanx>: no vector; <xref target="contract-door"/> allows
<spanx style="verb">delivery_ack</spanx> without <spanx style="verb">ack_window</spanx>. Vector
<spanx style="verb">synth_delivery_ack_without_window</spanx>.</t>
  <t><spanx style="verb">finality_policy_missing</spanx>: no vector; <xref target="contract-door"/> allows
<spanx style="verb">finality</spanx> without <spanx style="verb">depth</spanx>. Vector <spanx style="verb">synth_finality_without_depth</spanx>.</t>
  <t><spanx style="verb">input_not_settleable</spanx>: events that are not a list cannot be written
as event files. <spanx style="verb">exercise_inputs.py</spanx> cases E1 to E3.</t>
  <t><spanx style="verb">input_too_large</spanx> (records): 10001 records would be 10001 files; the
count is after <xref target="identical-bytes"/>. <spanx style="verb">exercise_inputs.py</spanx> cases E4 to
E6.</t>
  <t><spanx style="verb">input_too_large</spanx> (headers): more than 200000 verified headers. None:
not exercised.</t>
  <t><spanx style="verb">authorized_actions_missing</spanx>, <spanx style="verb">witness_key_invalid</spanx>,
<spanx style="verb">revocation_policy_not_recomputable</spanx>, <spanx style="verb">bad_expiry_height</spanx>: cannot
fire, because <xref target="contract-door"/> refuses first
(<xref target="settleable-terms"/>). <spanx style="verb">exercise_inputs.py</spanx> cases D1 to D6 show
<spanx style="verb">contract_not_verified</spanx>.</t>
</list></t>

<t>So the 128 vectors produce 22 of the 28 underspecified reasons. Four
reasons are reachable and now have vectors; <spanx style="verb">synth_witnesses_32</spanx>
(<spanx style="verb">within_grant</spanx>) pins the boundary beside <spanx style="verb">synth_too_many_witnesses</spanx>.</t>

<t>Two reasons are reachable only through inputs the vector layout cannot
carry, and exercise_inputs.py <xref target="MUSUBI-EXERCISE"/> drives them in
memory. Cases E1 to E3 pass events that are an object, null and the
string <spanx style="verb">"[]"</spanx>, and each gives <spanx style="verb">input_not_settleable</spanx>. Case E4 passes
10001 distinct records and gives <spanx style="verb">input_too_large</spanx>; case E5 passes
10000 distinct records and does not give it (the limit is more than
10000); case E6 passes 10001 copies of one record and does not give
it either, because identical bytes are one record
(<xref target="identical-bytes"/>) and are counted once (<xref target="size-limit"/>). Driving
these inputs is how the last point was found; <xref target="size-limit"/> did not
say it before revision 0.4. The headers branch of <spanx style="verb">input_too_large</spanx>
(<xref target="view-verify"/>, more than 200000 verified headers) is still not
exercised by any vector or exercise.</t>

<t>Four rows of <xref target="settleable-terms"/> can never produce their reason in
core v0: <spanx style="verb">authorized_actions_missing</spanx>, <spanx style="verb">witness_key_invalid</spanx>,
<spanx style="verb">revocation_policy_not_recomputable</spanx> and <spanx style="verb">bad_expiry_height</spanx>. Each
row describes an input that rules 9 or 10 of <xref target="contract-door"/>
already refuse, and <xref target="settleable-terms"/> applies only after
<xref target="contract-door"/> accepts. Cases D1 to D6 run those inputs (an
<spanx style="verb">authorized_actions</spanx> that is the string <spanx style="verb">"read"</spanx>, an absent
<spanx style="verb">authorized_actions</spanx>, a witness key that decodes to 31 bytes, an
<spanx style="verb">effective_at</spanx> of <spanx style="verb">"immediately"</spanx>, an <spanx style="verb">expiry_height</spanx> of -1, an
<spanx style="verb">expiry_height</spanx> that is the string <spanx style="verb">"800010"</spanx>) and each gives
<spanx style="verb">contract_not_verified</spanx>. The rows are kept so that no code changes
meaning.</t>

<t>Each line of EXERCISE.txt names the case, the input, what the text
says, and the verdict, status and codes the reference returns.
<spanx style="verb">python3 exercise_inputs.py --check</spanx> recomputes the lines and compares
them with EXERCISE.txt byte for byte. These lines are a record of
which input drove which outcome. They are not vectors and have no
expected.json (<xref target="projection"/>).</t>

<t>Every unexercised path was on the refusal side. A verifier that never
returns <spanx style="verb">underspecified</spanx> for a malformed policy goes on to compute a
verdict it has no basis for, so these are the paths worth pinning.</t>

</section>
<section anchor="disagreement"><name>Reporting a Disagreement</name>

<t>A disagreement is valuable. It means either the text, the reference
or the implementation is wrong, and only a comparison finds out
which.</t>

<t><list style="numbers" type="1">
  <t>Run the implementation over the vectors. For each vector, compare
its core fields (<xref target="core-fields"/>) with <spanx style="verb">expected.json</spanx>.</t>
  <t>Run <spanx style="verb">python3 run_reference.py</spanx> from the folder holding the source
text, to confirm the reference still gives the expected values on
the same machine.</t>
  <t>For each vector with a disagreement, write down: the vector id;
the core field; the expected value and the implementation's; the
section of this document followed; and, if possible, the step
where the computation first diverged (for example "conformance: the
name was treated as nonconforming").</t>
  <t>Say whether the vector is listed in <xref target="written-behaviour"/>. If it
is, say whether the behaviour should change.</t>
  <t>For a disagreement on an input that is not a vector, attach the
input files (contract, events, view, walks) so the reference can be
run on them.</t>
  <t>Open an issue on the repository that publishes the source text (at
the time of writing github.com/ogasurfproject-jpg/horizon-shield)
titled <spanx style="verb">settle core v0 disagreement: &lt;vector id&gt;</spanx>, with the above.</t>
</list></t>

<t>An implementer should not change an implementation to match a vector
believed to be wrong without saying so. A silent match hides a
defect. A reported disagreement is resolved by changing this text,
the reference, or the vector, and the change is recorded with a new
version of this document.</t>

</section>
</section>
<section anchor="admission"><name>Admission Records and Settle v1.19 (Outside Core v0)</name>

<section anchor="adm-why"><name>Why This Is Outside Core v0</name>

<t>Core v0 settles no contract that carries <spanx style="verb">requirements.admission</spanx>
(<xref target="scope"/>). Under such a contract, a relying party (a shop, an API,
the principal itself) decides at its own door, before an action,
whether an agent acting under the contract may go ahead, and signs
what it decided: an admission record <xref target="MUSUBI-ADMISSION"/>. The
execution record then names the admission it acted on. This section
describes the admission record of version 0.4, which carries the
relying party's account of a check it made of the content that came
with the request, and the rules settle v1.17, v1.18 and v1.19 apply to
admission records <xref target="MUSUBI-REF-V119"/>. It follows sections 17 to 19 of
revision 0.13 of <xref target="MUSUBI-SETTLE-CORE"/>.</t>

<t>Nothing here changes core v0. The 128 core vectors are unchanged, file
for file, and an implementation of core v0 that does not read
admissions is unaffected. When no accepted execution names an
admission these rules read, each version gives the output of the
version below it byte for byte, schema line included. This section
does not define the content check itself: the rule set a relying
party runs is named only by a digest, and the record carries no
content.</t>

</section>
<section anchor="adm-record"><name>The Admission Record</name>

<t>An admission record has the schema <spanx style="verb">a2a-admission-v0</spanx>. Its signed
bytes are <spanx style="verb">"a2a-admission-v0\n"</spanx> followed by the canonical bytes
(<xref target="canonical"/>) of the record without <spanx style="verb">signatures</spanx>, and
<spanx style="verb">admission_sha256</spanx> is the sha256 of those bytes. An execution record
under a contract that requires admission carries
<spanx style="verb">admission_ref: [{action, admission_sha256, executed}]</spanx>. The record
is signed by the relying party with a key served at
<spanx style="verb">relying_party.key_url</spanx>, an https URL on <spanx style="verb">relying_party.domain</spanx>. Its
<spanx style="verb">decision</spanx> is <spanx style="verb">admit</spanx>, <spanx style="verb">escalate</spanx> or <spanx style="verb">refuse</spanx>, with <spanx style="verb">reasons</spanx> from a
closed list, and <spanx style="verb">rules.version</spanx> names the version of the record.
Settle v1.19 reads versions 0.1 to 0.4:</t>

<t><list style="symbols">
  <t>0.1 carries no account of a content check.</t>
  <t>0.2 adds <spanx style="verb">immune</spanx>, the verdict of the immune rules over the request,
as digests and codes, and the reasons <spanx style="verb">immune_hold</spanx> and
<spanx style="verb">immune_reject</spanx>.</t>
  <t>0.3 adds <spanx style="verb">content</spanx> in the shape <spanx style="verb">a2a-content-check-v0</spanx> (digests and
three flags).</t>
  <t>0.4 keeps <spanx style="verb">immune</spanx> and carries <spanx style="verb">content</spanx> in the shape
<spanx style="verb">a2a-content-check-v1</spanx>, below, with the reason <spanx style="verb">content_off_topic</spanx>.</t>
</list></t>

<t>Each version has its own <spanx style="verb">does_not_establish</spanx> lines, and a verifier
refuses a record that carries another version's lines.</t>

</section>
<section anchor="adm-content"><name>The Content Block of Version 0.4</name>

<t>The block is a JSON object with exactly these members:</t>

<texttable title="The content block of version 0.4">
      <ttcol align='left'>member</ttcol>
      <ttcol align='left'>value</ttcol>
      <c><spanx style="verb">rules</spanx></c>
      <c><spanx style="verb">"a2a-content-check-v1"</spanx></c>
      <c><spanx style="verb">input_commit</spanx></c>
      <c>64 lowercase hex or null: <spanx style="verb">sha256(r + c)</spanx>, where c is the canonical bytes of the content the door received and r is a 32 byte salt the door returns to the requester alone</c>
      <c><spanx style="verb">check_sha256</spanx></c>
      <c>64 lowercase hex or null: the digest of the check's own record</c>
      <c><spanx style="verb">rules_sha256</spanx></c>
      <c>64 lowercase hex or null: the identifier of the rule set the door's policy names</c>
      <c><spanx style="verb">scope_sha256</spanx></c>
      <c>64 lowercase hex or null: the digest of the task scope the check was given</c>
      <c><spanx style="verb">decision</spanx></c>
      <c><spanx style="verb">"stop"</spanx>, <spanx style="verb">"pass"</spanx> or null</c>
      <c><spanx style="verb">marks</spanx></c>
      <c>a list, in this order, of those of <spanx style="verb">"off_topic"</spanx> and <spanx style="verb">"rule_conflict"</spanx> that apply</c>
      <c><spanx style="verb">unread</spanx></c>
      <c>boolean: part of the content was not read</c>
      <c><spanx style="verb">answered</spanx></c>
      <c>boolean: the check returned an answer</c>
      <c><spanx style="verb">checked</spanx></c>
      <c>boolean: a check covered these bytes</c>
</texttable>

<t><spanx style="verb">input_commit</spanx> is never the content's own digest, so the content
cannot be guessed from the public record and tested against it; only
a holder of the content and the salt can recompute it.</t>

<t>The content reasons of version 0.4 are <spanx style="verb">content_flagged</spanx>,
<spanx style="verb">content_scope_conflict</spanx>, <spanx style="verb">content_off_topic</spanx> and
<spanx style="verb">content_unchecked</spanx>. A record whose block breaks any of these rules is
malformed, and the verifier refuses it:</t>

<t><list style="numbers" type="1">
  <t>When <spanx style="verb">checked</spanx> is false, <spanx style="verb">decision</spanx> is null, <spanx style="verb">marks</spanx> is empty,
<spanx style="verb">unread</spanx> and <spanx style="verb">answered</spanx> are false, and the reasons include
<spanx style="verb">content_unchecked</spanx>.</t>
  <t>When <spanx style="verb">checked</spanx> is true, <spanx style="verb">input_commit</spanx>, <spanx style="verb">check_sha256</spanx> and
<spanx style="verb">rules_sha256</spanx> are not null.</t>
  <t>When <spanx style="verb">answered</spanx> is false, <spanx style="verb">decision</spanx> is null, <spanx style="verb">marks</spanx> is empty,
<spanx style="verb">unread</spanx> is false and the reasons include <spanx style="verb">content_unchecked</spanx>; an
answer has a decision.</t>
  <t><spanx style="verb">decision</spanx> is <spanx style="verb">stop</spanx> if and only if the reasons include
<spanx style="verb">content_flagged</spanx>.</t>
  <t><spanx style="verb">rule_conflict</spanx> is in <spanx style="verb">marks</spanx> if and only if the reasons include
<spanx style="verb">content_scope_conflict</spanx>.</t>
  <t><spanx style="verb">off_topic</spanx> is in <spanx style="verb">marks</spanx> if and only if the reasons include
<spanx style="verb">content_off_topic</spanx>.</t>
  <t>The reasons include <spanx style="verb">content_unchecked</spanx> if and only if <spanx style="verb">unread</spanx> is
true, <spanx style="verb">answered</spanx> is false or <spanx style="verb">checked</spanx> is false.</t>
  <t>A <spanx style="verb">stop</spanx> has no marks and <spanx style="verb">unread</spanx> false; <spanx style="verb">unread</spanx> true goes only
with <spanx style="verb">pass</spanx>. A mark needs a <spanx style="verb">scope_sha256</spanx>.</t>
  <t>An <spanx style="verb">admit</spanx> carries none of <spanx style="verb">content_flagged</spanx>,
<spanx style="verb">content_scope_conflict</spanx> and <spanx style="verb">content_unchecked</spanx>; it may carry
<spanx style="verb">content_off_topic</spanx>.</t>
</list></t>

<t>Whether a stop, a mark or an unread part refuses or escalates is the
relying party's policy, and the record says which it did. Under the
policy <spanx style="verb">on_off_topic: mark_only</spanx>, an off_topic mark does not stop the
request. A door whose policy requires no content check writes a
version 0.1 record with no content block.</t>

</section>
<section anchor="adm-reads"><name>What Settle v1.19 Reads</name>

<t>A contract may name the content check both parties expect in
<spanx style="verb">requirements.furui</spanx>: an object whose keys are only <spanx style="verb">rules_sha256</spanx>
(64 lowercase hex, or null) and <spanx style="verb">scope_text</spanx> (a non-empty string, or
null). Any other shape refuses the contract (<spanx style="verb">bad_requirements_furui</spanx>,
with <spanx style="verb">contract_not_verified</spanx>). Core v0 does not check this shape
(<xref target="contract-door"/>).</t>

<t>The rules apply only to a contract that requires admission. They read
the admissions that settle v1.11 reads: an admission named by an
accepted execution's <spanx style="verb">admission_ref</spanx> for the action performed, that
verifies under the relying party's key in hand and was not used by an
earlier action. Executions are taken in event set order
(<xref target="event-order"/>), actions in <spanx style="verb">performed_actions</spanx> order, and each
admission is read once.</t>

<t>For each such action and admission, nothing is added where settle
v1.11 already raised <spanx style="verb">executed_after_refusal</spanx> for the same execution,
action and admission. Otherwise:</t>

<t><list style="numbers" type="1">
  <t>The content gate, only when the contract names
<spanx style="verb">requirements.furui.rules_sha256</spanx>. <spanx style="verb">content_rules_mismatch</spanx> when the
admission is version 0.3 or 0.4 and its <spanx style="verb">content.rules_sha256</spanx> is
not the contract's. Otherwise <spanx style="verb">content_not_checked</spanx> when the
decision is <spanx style="verb">admit</spanx> and the admission carries no content check
(version 0.1 or 0.2), or its block says <spanx style="verb">checked</spanx> false (0.3 and
0.4) or <spanx style="verb">answered</spanx> false (0.4). A <spanx style="verb">content_not_checked</spanx> that settle
v1.17 raised for a version 0.4 admission, a version it did not
know, is taken back, and this gate stands in its place.</t>
  <t>The escalate rule, over versions 0.3 and 0.4 in one pass.
<spanx style="verb">executed_after_content_escalate</spanx> when the decision is <spanx style="verb">escalate</spanx>,
the reasons name a content reason (in version 0.3
<spanx style="verb">content_flagged</spanx>, <spanx style="verb">content_scope_conflict</spanx> or
<spanx style="verb">content_unchecked</spanx>; in version 0.4 also <spanx style="verb">content_off_topic</spanx>), and
the execution carries no approval that counts.
<spanx style="verb">executed_after_immune_escalate</spanx> when the decision is <spanx style="verb">escalate</spanx>,
the reasons name <spanx style="verb">immune_hold</spanx> and no content reason, and the
execution carries no approval that counts.</t>
  <t><spanx style="verb">content_marks</spanx>: for the version 0.4 admissions read, the number
with each mark and the number with <spanx style="verb">unread</spanx>. A mark is counted, not
a deviation.</t>
</list></t>

<t>An approval counts for an action when it is the principal's
a2a-approval-v2 over this contract's sha256, or a pinned approver's
a2a-approval-v2 or a2a-approval-v3 for an action the approver lists,
and its <spanx style="verb">valid_until_height</spanx> is at or above the execution's height. A
<spanx style="verb">single_use</spanx> approval counts once in the pass: an approval already
counted for an earlier escalate does not count again
(<spanx style="verb">approval_reused</spanx>; an expired one is <spanx style="verb">approval_expired</spanx>). This spend
is the rule's own and is not shared with the approvals of
<xref target="approvals"/>.</t>

<t>These rules run only when the settlement below them has the verdict
<spanx style="verb">within_grant</spanx>, <spanx style="verb">deviation</spanx> or <spanx style="verb">nothing_to_settle</spanx>. Any of their
deviations makes the verdict <spanx style="verb">deviation</spanx>, and the bond outcome follows
<xref target="bond-outcome"/>; when every deviation is taken back, the verdict and
the bond outcome are those of core v0. When the rules read an
admission of version 0.4 or an escalate on the immune rules alone, the
schema line is <spanx style="verb">a2a-settlement-v1.19</spanx>, <spanx style="verb">settled_under</spanx> names the
schema below it, and the output carries <spanx style="verb">escalate_gate</spanx> (the rows of
both escalate clauses), <spanx style="verb">content_marks</spanx> and, when the contract names a
check, <spanx style="verb">content_gate_v04</spanx>.</t>

</section>
<section anchor="adm-limits"><name>What These Rules Do Not Establish</name>

<t><list style="symbols">
  <t>That any content check ran, or that the same check would answer the
same way again. The content block is the relying party's own signed
statement; the rules compare digests, flags and marks, and do not
read the content, run a check, or recompute <spanx style="verb">input_commit</spanx> or
<spanx style="verb">rules_sha256</spanx>.</t>
  <t>That the content was off topic or in conflict with the scope. A mark
says what the relying party's own check marked.</t>
  <t>That an approval answers a particular escalate. An approval signs no
admission's digest, so an approval signed before the door escalated
and still valid counts the same as one signed after.</t>
  <t>That rates measured for a named rule set hold on a contract's scope
(<xref target="impl-use"/>).</t>
</list></t>

</section>
<section anchor="adm-vectors"><name>Vectors and Reference</name>

<t>The rules are kept apart from the core vectors, in
settle-core-v0/vectors_v1_17/ (13 vectors), vectors_v1_18/ (14) and
vectors_v1_19/ (19) <xref target="MUSUBI-REF-V119"/>. Each folder holds the core
files of <xref target="vectors-contents"/>, the admissions with their anchors and
the relying party's keys; expected.json adds the counts of each gate.
The reference is settle_v1_19.py, called as
<spanx style="verb">settle_v1_19.settle_v1_19(contract, events, view, admissions=...,
relying_keys=...)</spanx>; run_reference.py checks these vectors after the
core vectors. Later settle versions in the same repository read later
versions of the admission record and are not described here.</t>

</section>
</section>
<section anchor="impl-status"><name>Implementation Status</name>

<t>Note to the RFC Editor: please remove this section before
publication.</t>

<t>This section records the status of known implementations of the
protocol defined by this specification at the time of posting of this
Internet-Draft, and is based on a proposal described in <xref target="RFC7942"/>.
The description of implementations in this section is intended to
assist the IETF in its decision processes in progressing drafts to
RFCs. Please note that the listing of any individual implementation
here does not imply endorsement by the IETF. Furthermore, no effort
has been spent to verify the information presented here that was
supplied by IETF contributors. This is not intended as, and must not
be construed to be, a catalog of available implementations or their
features. Readers are advised to note that other implementations may
exist.</t>

<t>According to <xref target="RFC7942"/>, "this will allow reviewers and working
groups to assign due consideration to documents that have the benefit
of running code, which may serve as evidence of valuable
experimentation and feedback that have made the implemented protocols
more mature. It is up to the individual working groups to use this
information as they see fit".</t>

<t>The sections below separate implementations written by others
(<xref target="impl-others"/>), runs of the reference by others
(<xref target="impl-reruns"/>), the reference itself (<xref target="impl-ref"/>), and the
operator's own use of the records of <xref target="admission"/> (<xref target="impl-use"/>). A
public table of the rows of the first two, with a link for each, is
kept at <xref target="INTEROP-TABLE"/>.</t>

<section anchor="impl-others"><name>Implementations by Others</name>

<t>Two implementations of settle core v0 by authors other than the
publisher of the reference agree with it on all 128 vectors, on the
eight core fields (<xref target="core-fields"/>). Neither was written from the
source text alone: one also read NOTES.md, and the other had the
expected outputs and let them decide where the text could be read two
ways. The entries "Relation to the reference" below say which. No
implementation has yet been written from the text and the vector
inputs with the expected outputs withheld and its code fixed by hash
before comparison; that test is open. The 128 core vectors have not
changed since revision 0.4 (<xref target="changes-since-02"/>,
<xref target="changes-since-01"/>). Both implementations were run in the
differential run behind revision 0.9 (<xref target="revisions-09-11"/>). Neither
reads admission records; the rules of <xref target="admission"/> have no
implementation by others yet.</t>

<section anchor="impl-aeoess"><name>aeoess/musubi-settle-core-check</name>

<dl>
  <dt>Organization:</dt>
  <dd>
    <t>aeoess (Agent Passport System).</t>
  </dd>
  <dt>Implementation:</dt>
  <dd>
    <t>musubi-settle-core-check <xref target="AEOESS-CHECK"/>, written from
SETTLE_CORE_v0.md and NOTES.md alone, with an AI coding agent under
its author's direction, from an input copy without the expected
outputs.</t>
  </dd>
  <dt>Maturity and coverage:</dt>
  <dd>
    <t>The eight core fields on every vector.</t>
  </dd>
  <dt>Implementation experience:</dt>
  <dd>
    <t>Updated to revision 0.4 by its author (commit 53cfafb), it agrees
with all 128 vectors with no field left unassessed (item N19 of
<xref target="MUSUBI-NOTES"/>). Its earlier runs (76, 117 and 123 vectors) and
the notes that led to revisions 0.1 to 0.3 are recorded in items
N13 to N16.</t>
  </dd>
  <dt>Relation to the reference:</dt>
  <dd>
    <t>Written without the expected outputs, but from NOTES.md as well as
the source text. Its author declared that, and the public tally
<xref target="ERRATA-K001"/> does not count it under that tally's rule for that
reason.
Run by its author, who maintains a party to the contract behind
<spanx style="verb">real_p31_baseline</spanx> (<xref target="vectors-contents"/>).</t>
  </dd>
  <dt>Licensing:</dt>
  <dd>
    <t>Apache-2.0.</t>
  </dd>
  <dt>Contact:</dt>
  <dd>
    <t>aeoess, through <xref target="ISSUE-49"/>.</t>
  </dd>
  <dt>Last updated:</dt>
  <dd>
    <t>2026-10-22</t>
  </dd>
</dl>

</section>
<section anchor="impl-kuangmi"><name>kuangmi-bit/nenrin-independent-verifiers</name>

<dl>
  <dt>Organization:</dt>
  <dd>
    <t>kuangmi-bit.</t>
  </dd>
  <dt>Implementation:</dt>
  <dd>
    <t>The settle-core-v0 directory of nenrin-independent-verifiers
<xref target="KUANGMI-CHECK"/>, at commit bbe455c, in Python. Written from
revision 0.5 of the source text and the published vectors only,
without reading the reference or musubi-settle-core-check. Standard
library at run time; Ed25519 through the Python package cryptography
when present and through its own RFC 8032 verifier otherwise.</t>
  </dd>
  <dt>Maturity and coverage:</dt>
  <dd>
    <t>The eight core fields on every vector.</t>
  </dd>
  <dt>Implementation experience:</dt>
  <dd>
    <t>Agrees with all 128 vectors, including <spanx style="verb">real_p31_baseline</spanx>,
<spanx style="verb">real_cross_record_names_other_contract</spanx> and the 25 parent vectors.
Its author also ran the reference on the same machine and got 128 of
128. The publisher of the reference reran it at bbe455c and got the
same. Its README lists the readings where the text could be read two
ways and the vectors decided; revision 0.6 states two of those
places in the text, and neither changed a vector (item N20 of
<xref target="MUSUBI-NOTES"/>).</t>
  </dd>
  <dt>Relation to the reference:</dt>
  <dd>
    <t>Written without reading the reference or musubi-settle-core-check,
as its author records (item N20 of <xref target="MUSUBI-NOTES"/>). The expected
outputs were available to it and settled the readings listed in its
README, so it is not a test of the text without the vectors. It is
the one implementation that <xref target="ERRATA-K001"/> counts under that tally's
rule, which allows the vectors.</t>
  </dd>
  <dt>Licensing:</dt>
  <dd>
    <t>MIT.</t>
  </dd>
  <dt>Contact:</dt>
  <dd>
    <t>kuangmi-bit, through <xref target="ISSUE-49"/>.</t>
  </dd>
  <dt>Last updated:</dt>
  <dd>
    <t>2026-10-22</t>
  </dd>
</dl>

</section>
</section>
<section anchor="impl-reruns"><name>Runs by Others</name>

<t>These are not further implementations of this document. They are
outside runs of the reference, or outside checks of records that a
settlement reads, listed because each was done on the reader's own
machine from published bytes.</t>

<dl>
  <dt>Recomputation of a live settlement:</dt>
  <dd>
    <t>Poke-nushi recomputed the settlement of the 22 October run of the
cA2A companion pilot (horizon-shield Issue #51, set 2) from the
published bundle at commit b85029de, with the companion's settler
(pilot_v0.py, which reads records of the kinds this document
defines but is not settle core v0 itself) and ca2a-runtime 0.4.0,
and reproduced the published output sha256. The
same reader checked the signatures and the OpenTimestamps proof of
the log checkpoint, and rebuilt the header view from Blockstream and
mempool.space headers up to block 972564: the verdicts did not
change and only <spanx style="verb">anchor_depth</spanx> moved, from 9 to 14 <xref target="RUN-POKE"/>.
This shows which part of a settlement rests on the reader's header
range, a point that applies to settle core as well (<xref target="horizon"/>).</t>
  </dd>
  <dt>Byte identity of the audits a settlement names:</dt>
  <dd>
    <t>In the same run, aeoess, which ran the same calls through its own
row of the testbed (Agent Passport System), reported that the A and B audit files of its
run are byte identical to those of ours (sha256 <spanx style="verb">c2920477...</spanx>,
1,532 bytes, and <spanx style="verb">40bc2a46...</spanx>, 1,101 bytes) <xref target="RUN-AEOESS"/>.</t>
  </dd>
  <dt>Acceptance of the executions:</dt>
  <dd>
    <t>babyblueviper1 reported issuing four acceptance records for the
same run, A and B of each of the two rows, and that A passes all four
criteria and B fails three, matching the expected A accept and B
reject <xref target="RUN-BBV"/>. The four records were published on 24 October
2026 <xref target="BBV-SET"/>; the publisher of this document fetched the two
records of the MUSUBI row and their sha256 match those stated with
them (<spanx style="verb">204915d0...</spanx> for A, verdict pass; <spanx style="verb">7e05c6c1...</spanx> for B,
verdict fail). In a later run of a newer version of the verifier
that author used, the records as published gave a problem code about
how the producer's key is listed, which goes away when that key is
added; the verdicts did not change <xref target="RUN-BBV-RERUN"/>.</t>
  </dd>
  <dt>A neighbouring record format:</dt>
  <dd>
    <t>Pico, the AI agent that maintains Agent Errata, reproduced its part
of a decision record exchange (E6 of the agent governance vocabulary
boundary map) on fresh clones: 12 of 12 cases agree with the record
published by this document's author <xref target="RUN-PICO"/>. This does not
exercise settle core; it is listed because it uses the same
publish-and-recompute procedure as <xref target="disagreement"/>.</t>
  </dd>
</dl>

</section>
<section anchor="impl-ref"><name>Reference Implementation</name>

<dl>
  <dt>Organization:</dt>
  <dd>
    <t>The HORIZONs Co., Ltd.</t>
  </dd>
  <dt>Implementation:</dt>
  <dd>
    <t>settle_v1_16.py (output schema a2a-settlement-v1.16) and the modules
it imports, at commit e41715d7 <xref target="MUSUBI-REF"/>; settle_v1_19.py for
the rules of <xref target="admission"/> <xref target="MUSUBI-REF-V119"/>. It is called as a
library, <spanx style="verb">settle_v1_16.settle_v1_16(contract, events, view,
parent=..., nenrin_records=...)</spanx>. Its command line (<spanx style="verb">--settle</spanx>,
<spanx style="verb">--event</spanx>, <spanx style="verb">--view</spanx>, <spanx style="verb">--nenrin</spanx>, <spanx style="verb">--parent</spanx>) first checks each
view's strict form and runs fork choice, and for a view that does
not pass prints why and writes no settlement. The vectors use the
library.</t>
  </dd>
  <dt>Description:</dt>
  <dd>
    <t>The implementation from which this document was written. It covers
all of settle core v0 and the gates outside its scope (<xref target="scope"/>).</t>
  </dd>
  <dt>Maturity:</dt>
  <dd>
    <t>Reference. Its behaviour defines the expected values of the vectors.</t>
  </dd>
  <dt>Coverage:</dt>
  <dd>
    <t>All of settle core v0. The canonical form (<xref target="canonical"/>) is also
implemented in JavaScript (canonical_v0.mjs), which covers only that
section.</t>
  </dd>
  <dt>Licensing:</dt>
  <dd>
    <t>MIT.</t>
  </dd>
  <dt>Implementation experience:</dt>
  <dd>
    <t>At commit e41715d7, run with Python 3.14.7 and the Python package
cryptography 50.0.1 with outgoing network connections blocked,
run_reference.py reports "128/128 vectors agree with the reference",
then 13/13, 14/14 and 19/19 for the vectors of <xref target="adm-vectors"/>, and
<spanx style="verb">exercise_inputs.py --check</spanx> reports "EXERCISE.txt agrees"
(<xref target="exercise"/>). <spanx style="verb">settle_v1_19.py --selftest</spanx> passes its 5 checks and
<spanx style="verb">admission_verify_v0.py --selftest</spanx> its 10.</t>
  </dd>
  <dt>Contact:</dt>
  <dd>
    <t>Toshikatsu Oga, thehorizon.nnovation@icloud.com</t>
  </dd>
  <dt>Last updated:</dt>
  <dd>
    <t>2026-11-09</t>
  </dd>
</dl>

<t>Since draft -02, settle_v1_16.py changed once, for revision 0.9
(<xref target="revisions-09-11"/>), and two imported files changed in layers
outside settle core: admission_verify_v0.py, which reads the admission
versions of <xref target="admission"/>, and agreement_verify.py. On the 128 vectors
the complete output of settle_v1_16.py is the same before and after
the fix of revision 0.9 (item N23 of <xref target="MUSUBI-NOTES"/>). The reference, as run at commit
e41715d7, consists of these files (first 16 hex digits of each file's
sha256):</t>

<texttable title="Reference files">
      <ttcol align='left'>file</ttcol>
      <ttcol align='left'>sha256 prefix</ttcol>
      <c>settle_v1.py</c>
      <c><spanx style="verb">0e51350331c8931c</spanx></c>
      <c>settle_v1_1.py</c>
      <c><spanx style="verb">a2391a5db199bcda</spanx></c>
      <c>settle_v1_2.py</c>
      <c><spanx style="verb">6c674fd54a1a5c89</spanx></c>
      <c>settle_v1_3.py</c>
      <c><spanx style="verb">17119bbc7802647f</spanx></c>
      <c>settle_v1_4.py</c>
      <c><spanx style="verb">ef3779d05fc98e40</spanx></c>
      <c>settle_v1_5.py</c>
      <c><spanx style="verb">f9b313566713cdc4</spanx></c>
      <c>settle_v1_6.py</c>
      <c><spanx style="verb">6752a748ed7b6e3b</spanx></c>
      <c>settle_v1_7.py</c>
      <c><spanx style="verb">878a3a3697775d3a</spanx></c>
      <c>settle_v1_8.py</c>
      <c><spanx style="verb">fe61f67c1593c776</spanx></c>
      <c>settle_v1_9.py</c>
      <c><spanx style="verb">7a84302be1ec38a1</spanx></c>
      <c>settle_v1_10.py</c>
      <c><spanx style="verb">34beeebaa1397059</spanx></c>
      <c>settle_v1_11.py</c>
      <c><spanx style="verb">fb03687cbacd2713</spanx></c>
      <c>settle_v1_12.py</c>
      <c><spanx style="verb">40a8fbb82a8e7680</spanx></c>
      <c>settle_v1_13.py</c>
      <c><spanx style="verb">eb02b3a6544b1bf1</spanx></c>
      <c>settle_v1_14.py</c>
      <c><spanx style="verb">7d0704fa060adafd</spanx></c>
      <c>settle_v1_15.py</c>
      <c><spanx style="verb">a2be5de8ba60656e</spanx></c>
      <c>settle_v1_16.py</c>
      <c><spanx style="verb">a7a1621224a63ef2</spanx></c>
      <c>settle_v1_17.py</c>
      <c><spanx style="verb">03a0915e2fcc8f64</spanx></c>
      <c>settle_v1_18.py</c>
      <c><spanx style="verb">c3c356011aa5cadf</spanx></c>
      <c>settle_v1_19.py</c>
      <c><spanx style="verb">eeab089991e37887</spanx></c>
      <c>admission_verify_v0.py</c>
      <c><spanx style="verb">28751e3405f86033</spanx></c>
      <c>agreement_verify.py</c>
      <c><spanx style="verb">7c54b644a0ca8f32</spanx></c>
      <c>anchor_compose.py</c>
      <c><spanx style="verb">a661f673c05dbe6e</spanx></c>
      <c>approval_v3.py</c>
      <c><spanx style="verb">e8dde985dbe6fbf4</spanx></c>
      <c>clause_eval_v0.py</c>
      <c><spanx style="verb">fee48c3a66c28fe1</spanx></c>
      <c>contract_v0.py</c>
      <c><spanx style="verb">211a568783705e11</spanx></c>
      <c>convergence_v0.py</c>
      <c><spanx style="verb">65a267380fa623f5</spanx></c>
      <c>corroboration_v0.py</c>
      <c><spanx style="verb">62ae1d6c27a92228</spanx></c>
      <c>independence_v0.py</c>
      <c><spanx style="verb">50964883c27896de</spanx></c>
      <c>recovery_verify.py</c>
      <c><spanx style="verb">fd4c93f015f6b495</spanx></c>
      <c>spine_verify.py</c>
      <c><spanx style="verb">ad2076d49be6763e</spanx></c>
      <c>terms_v0.py</c>
      <c><spanx style="verb">6532d5c6160a3974</spanx></c>
      <c>requirements_furui_v0.py</c>
      <c><spanx style="verb">96b169e14e188110</spanx></c>
</texttable>

</section>
<section anchor="impl-use"><name>Records of Version 0.4 in Use</name>

<t>These are uses by the publisher of this document, not by others, and
are listed with their limits.</t>

<dl>
  <dt>The measured check:</dt>
  <dd>
    <t>Version 0.4 of the admission record is the shape in which a content
check measured in a blind test, FURUI blind-06, reports its answer.
The plan was anchored before the draw (ledger entry 231, Bitcoin
block 974072), the result is ledger entry 232 (block 974128), and
the draw manifest, salt and counts are in commit 200d5d96
<xref target="BLIND-06"/>. The result states that all six lines set in the plan
held. Limits: the rates were measured on the scopes of public
datasets, not on a scope rendered from a contract; the test was run
by the publisher; and it measures the check, not settlement.</t>
  </dd>
  <dt>One call through every layer:</dt>
  <dd>
    <t>The publisher's daily self-check of 2026-11-09 <xref target="TOOSHI-1109"/>
found, for the one call on the paid path in its window (2026-11-07
03:47Z to 2026-11-08 03:47Z), the content check envelope (FURUI), the
door's admission record (SEKI), the MUSUBI settlement and the NENRIN
ledger entry, one of each (its item 17). The admission record of
that call is version 0.4, decision admit, ledger entry 268. Its item
25 found no part of the records of that call missing, and two of its
checks for that call not yet possible (not_yet). Limits: the call
was made to the publisher's own demonstration store
(hs-seki-demo), paid in pilot tickets that carry no monetary value;
the relying party is the publisher; and one call is a demonstration
of the path, not a measure of it.</t>
  </dd>
</dl>

</section>
</section>
<section anchor="related"><name>Relationship to Other Work</name>

<section anchor="related-audit"><name>The Agent Auditing Architecture</name>

<t><xref target="I-D.kuehlewind-audit-architecture"/> describes roles and four record
types for auditing agents across domains. Its Auditor "determines
whether recorded behaviour matched the User's intent and the
authorization in force at the time", without trusting the Agent. The
work items it proposes specify record formats, a profile of SCITT and
a query interface; none of them specifies the procedure by which an
Auditor reaches that determination, so that two Auditors holding the
same records reach the same one. This document specifies such a
procedure for one record set, with test vectors, and is offered as an
input to that question, not as a replacement for the architecture's
records.</t>

<texttable title="Roles">
      <ttcol align='left'>architecture role</ttcol>
      <ttcol align='left'>in this document</ttcol>
      <c>User</c>
      <c>the principal, who signs the contract and its grant</c>
      <c>Agent</c>
      <c>the contractor, which signs execution records; a witness named in the grant</c>
      <c>Sub-Agent</c>
      <c>a party to whom work may be delegated under <spanx style="verb">grant.delegation</spanx></c>
      <c>Audit Recorder, Auditing Service</c>
      <c>no separate role: each record is signed by the party that acted (<xref target="sec-unsigned"/>)</c>
      <c>Audit Store</c>
      <c>any source of records; a settler trusts none of them</c>
      <c>Transparency Log</c>
      <c>not required by core v0; see <xref target="related-scitt"/></c>
      <c>Auditor</c>
      <c>the settler</c>
</texttable>

<texttable title="Record types">
      <ttcol align='left'>architecture record type</ttcol>
      <ttcol align='left'>in this document</ttcol>
      <c>Delegation Record</c>
      <c>the contract's grant: authorized, prohibited and conditional actions, delegation, expiry and revocation policy, signed by both parties (<xref target="contract"/>)</c>
      <c>Action Record</c>
      <c>an execution record, <spanx style="verb">a2a-execution-v0</spanx></c>
      <c>Authorization Transition Record</c>
      <c>a revocation, <spanx style="verb">a2a-revocation-v0</spanx>, and its acknowledgement, <spanx style="verb">a2a-revocation-ack-v0</spanx>; a principal's approval carried by an execution, and its consumption when single use (<xref target="approvals"/>); expiry by height (<xref target="expiry"/>)</c>
      <c>Interaction Record</c>
      <c>none; settlement reads no conversation</c>
</texttable>

<t>The two also differ in ways a profile would have to settle. The
records here are JSON with this document's canonical form and
detached Ed25519 signatures, where the architecture asks for CBOR/COSE
or JSON/JWS. Parties are identified by raw Ed25519 keys, not by
workload identity credentials. Records are linked by the contract and
by digests, not by a shared audit context identifier. And a settler
reads every record in full: selective disclosure, which the
architecture treats as a core design consideration, is not addressed
here.</t>

</section>
<section anchor="related-scitt"><name>Transparency: SCITT</name>

<t>SCITT <xref target="RFC9943"/> registers signed statements with a Transparency
Service, whose receipts show that a statement existed and in what
order, and prevent an inconsistent history from being shown to
different readers. This document does not use a Transparency Service.
It orders records by Bitcoin anchors <xref target="OPENTIMESTAMPS"/>, which bound
when a record existed, but it does not prevent a holder from
withholding records (<xref target="sec-unsigned"/>).</t>

<t>The two are meant to be combined, not substituted. Records can be
registered as signed statements, so that a receipt gives existence and
order within a Transparency Service; a Bitcoin anchor of the record,
or of the service's checkpoint, then adds evidence of time that does
not rest on any one Transparency Service. A profile that carries
SCITT receipts as anchors is not defined in core v0.</t>

</section>
<section anchor="related-authz"><name>Authorization: GNAP and OAuth</name>

<t>GNAP <xref target="RFC9635"/>, and OAuth with token exchange <xref target="RFC8693"/> and its
<spanx style="verb">act</spanx> claim for delegation, define how a party obtains authorization
and how a delegation is carried in a token. They do not define how a
party outside the grant decides, after the fact, whether a recorded action stayed
within that authorization. Settlement does the second, over a grant
both parties signed. A profile could take the grant from a GNAP access
right, or from an OAuth access token's scope, <spanx style="verb">authorization_details</spanx>
<xref target="RFC9396"/> and <spanx style="verb">act</spanx> chain, instead of from <spanx style="verb">a2a-contract-v0</spanx>; such a
profile is not defined in core v0.</t>

</section>
<section anchor="related-drafts"><name>Other Internet-Drafts</name>

<t><list style="symbols">
  <t><xref target="I-D.gilda-audit-agent-audit-record"/> defines a record format for
an agent's authorization decisions. Such records are inputs of the
kind settlement evaluates.</t>
  <t><xref target="I-D.he-agentproto-verifier-evidence"/> defines evidence of what a
verifier exercised and how completely. <xref target="exercise"/> and the vectors
of this document are among the material it discusses.</t>
  <t><xref target="I-D.kroehl-agentic-trust-aae"/> defines a machine-evaluable
authorization envelope, a candidate form for the grant a settler
reads.</t>
  <t><xref target="I-D.gravit-gevp"/> defines verification records of its own. It and
this document are separate rows of the same testbed (<xref target="ISSUE-55"/>),
answering the same questions on the same traces.</t>
  <t><xref target="I-D.birkholz-verifiable-agent-conversations"/> defines records of
agent conversations, the architecture's Interaction Records, which
settlement does not read.</t>
</list></t>

<t>None of these defines a procedure for reaching an Auditor's
determination from a set of records, which is the part this document
specifies.</t>

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

<section anchor="sec-unsigned"><name>Unsigned Parts of a Record</name>

<t>A record has three parts (<xref target="triage"/>): the body, which its signer
signed, and the anchor and the signatures list, which nobody signed.
Any party that holds copies of genuine records, the contract and
the walks can therefore change anchors and signature lists at will,
but cannot change a body without the signer's key. The rules are
arranged around this:</t>

<t><list style="symbols">
  <t>A copy whose signatures list holds no strict, verifying entry from
an allowed role is never examined: it is set aside or rejected in
<xref target="triage"/>, and no later step (collapse, size limit, conformance,
anchors, NENRIN checks) sees it. An unsigned record cannot add a
charge or forfeit a bond.</t>
  <t>A copy that keeps the genuine signature but adds junk entries to
<spanx style="verb">signatures</spanx> has the same record signing bytes as the genuine
record. Junk entries are ignored, and copies of one signed body are
collapsed (<xref target="collapse"/>), keeping the copy that anchors validly at
the smallest (height, record_sha256).</t>
  <t>A copy with an anchor that is not strict is set aside when a
validly anchored copy of the same body is present. A copy whose
anchor is strict but wrong is dropped by the collapse when a
validly anchored copy is present.</t>
  <t>Binding is by contract_sha256 alone (<xref target="binding"/>): a record made
under earlier terms with the same contract_id is foreign and is
never triaged.</t>
</list></t>

<t>What remains is a liveness property, not a safety one. When the
genuine copy is not in the input, a copy with an incomplete or wrong
anchor from a strictly signed record makes the whole settlement
<spanx style="verb">underspecified</spanx> (<xref target="verdict"/>); a party can delay a verdict in the
same way by handing the settler such a record. A copy whose
<spanx style="verb">block_hash</spanx> was edited to disagree with the view at a held height is
<spanx style="verb">orphaned</spanx> and excluded without an underspecified reason
(<xref target="wb-list"/>); a settler that has only such a copy does not see the
act. Settlers that collect records from more than one source reduce
both risks. <spanx style="verb">underspecified</spanx> must not be read as a statement that
the terms are bad.</t>

<t>The record_sha256 covers the unsigned parts, so it identifies a copy,
not an act. The record signing bytes identify the act. One
consequence reaches the listing: the same-block order of
<xref target="same-block"/> sorts by record_sha256, so the holder of the copies
chooses which of two same-block executions is blamed for a single use
approval. A later version that wants this order to be a signed fact
would sort same-height records by the body digest instead; that
changes the expected results of the vectors with two records in one
block, so it is not done in core v0.</t>

</section>
<section anchor="sec-ed25519"><name>Signature Malleability and Strict Ed25519</name>

<t>Ed25519 implementations differ on adversarial signatures: a
non-canonical S (for example S + L), a small order or mixed order R or
public key, and cofactored versus cofactorless verification equations
<xref target="EDDSA-TAMING"/>. Honest signatures verify everywhere. Settle core v0
counts a signature only when it is strict (<xref target="ed25519"/>): S, read as a
little-endian integer, is below L, and R and the public key are
canonical encodings of points of prime order that are not the
identity. With all points in the prime order subgroup and S below L,
the libraries that disagree on lax signatures agree.</t>

<t>A lax signature that a library accepts is not ignored silently: the
record is set aside as <spanx style="verb">signature_not_strict</spanx>, and a contract
signature that is not strict makes the settlement underspecified
(<spanx style="verb">contract_signature_not_strict</spanx>). Anyone can turn a genuine
signature into a lax one (S + L) without a key; that changes the
record_sha256 but not the body, and the genuine copy, if present,
still settles (vectors <spanx style="verb">synth_signature_s_plus_l</spanx> and
<spanx style="verb">synth_strict_signature_beside_malleated</spanx>).</t>

<t>Item N9 of <xref target="MUSUBI-NOTES"/> records that the behaviour of the
reference's OpenSSL on published Ed25519 edge cases was not tested
beyond the strictness rule. The strictness rule is what conformance
depends on.</t>

</section>
<section anchor="sec-numbers"><name>Number and Canonicalization Divergence</name>

<t>Settlement depends on byte-exact digests computed by different
runtimes. JSON readers differ: some accept the tokens <spanx style="verb">NaN</spanx> and
<spanx style="verb">Infinity</spanx>, some read <spanx style="verb">970480.0</spanx> as a float and others as the integer
970480, integers above 2^53 are exact in some and rounded in others,
duplicate keys are resolved differently, and key order by code point
differs from order by UTF-16 code unit above U+FFFF. A pattern ending
in <spanx style="verb">$</spanx> in some regular expression engines also accepts a trailing line
feed, and some hex decoders skip whitespace.</t>

<t>The rules of <xref target="reading-json"/>, <xref target="canonical"/> and <xref target="triage"/> remove
these differences for the values that matter: only integers within
plus or minus (2^53 - 1) are strict, keys in the contract and in
strict values are ASCII, fixed forms such as hex digests are matched
whole, and a value that is not strict is set aside or makes the
settlement underspecified instead of being read two ways. An
implementer has to parse numbers to an exact integer representation,
read the tokens <spanx style="verb">NaN</spanx>, <spanx style="verb">Infinity</spanx> and <spanx style="verb">-Infinity</spanx> as non-integer
numbers rather than as errors (<xref target="reading-json"/>), write them in the
spellings of item 6 of <xref target="canonical"/> when a signature over a value
that is not strict is checked, match fixed forms against the whole
string, and refuse whitespace in hex.</t>

<t>Some divergence remains possible outside strict values. The
record_sha256 of a record that is set aside, foreign or unbound is still
reported in <spanx style="verb">dispositions</spanx>, and its canonical form can depend on the
runtime for values that are not strict. A record whose record_sha256
"cannot be computed at all" is set aside without a disposition; in
the reference this happens, for example, for an integer too long for
the Python runtime to write out, and another runtime may reach that
limit at a different length. Item N4 of <xref target="MUSUBI-NOTES"/> also records
that JSON batches reached through a proof (<xref target="batch-leg"/>) are parsed
as the lower layers of the reference parse them.</t>

</section>
<section anchor="sec-reorg"><name>Reorganizations Deeper Than the Finality Depth</name>

<t>Settle core v0 takes one header view as given. It checks linkage,
proof of work against each header's own nBits, the grant's difficulty
floor and the signed checkpoint, and nothing else of Bitcoin
consensus. It does not establish that the view is the heaviest chain.
A status of <spanx style="verb">final</spanx> means that every accepted and orphaned record is
at least <spanx style="verb">grant.finality.depth</spanx> blocks below the tip of the view; a
reorganization deeper than that depth can still change a final
settlement. Choosing between several views (fork choice) is out of
scope (<xref target="scope"/>). The difficulty floor and the checkpoint limit what
a party can forge cheaply, but a view supplied by an adversary with
enough work is accepted.</t>

</section>
<section anchor="not-established"><name>What the Verdict Does Not Establish</name>

<t><list style="symbols">
  <t>That any record describes what happened. A signature proves which key signed, not
that the act happened as described.</t>
  <t>That no key was stolen. A stolen key signs validly. Key rotation
and compromise are outside settle.</t>
  <t>That acts nobody recorded did not happen. <spanx style="verb">within_grant</spanx> covers the
accepted records only.</t>
  <t>That a foreign, unbound, inconsistent, rejected, orphaned or
predated record is false. Only that it was not settled here.</t>
  <t>Full Bitcoin consensus validity. Only linkage, proof of work against
each header's own nBits, the grant's difficulty floor and the signed
checkpoint are checked.</t>
  <t>That the view is the heaviest chain. Core v0 takes one view as
given. A reorganization deeper than the finality depth can change a
final settlement.</t>
  <t>That the key at a party's <spanx style="verb">key_url</spanx> is the key in the contract.
Settle never fetches.</t>
  <t>That the walk cited by <spanx style="verb">nenrin_ref</spanx> was checked. <xref target="nenrin"/> only
compares the contract it names.</t>
  <t>That anyone judged liability or fault, or that this is a legal
contract. The verdict is a function anyone can recompute from the
same bytes.</t>
  <t>That the parts out of scope (<xref target="scope"/>) hold. A contract that relies
on the spine gate, convergence, pinned approvers or limits must be
settled by a settler that reads them.</t>
  <t>That settle layers before v1.16 agree. They do not: over the 128
vectors, settle_v1_15.py gives different core fields (dispositions
aside) on 51 and stops with an exception on 3. This document
specifies v1.16 only.</t>
  <t>That a record set aside or rejected is false. Only that it was not
strictly signed, or not readable strictly, so it was not settled
here.</t>
  <t>That a content check ran, or what a request contained. The rules of
<xref target="admission"/> read digests, flags and marks that the relying party
signed (<xref target="adm-limits"/>).</t>
</list></t>

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

<t>A contract names both parties by domain and pins their public keys,
and records carry action names, delegation targets, payload digests
and references to walks. Anyone who is given these bytes can settle,
so anyone who is given them learns who contracted with whom and what
the agent did. Anchoring commits only a digest to the Bitcoin block
chain, but the commitment is public and permanent, and publishing a
record later links it to a block and so to a time. Parties should
treat the inputs of a settlement as disclosed to every settler they
hand them to. Settlement itself fetches nothing (<xref target="scope"/>), so
computing it reveals nothing to a network observer.</t>

</section>
<section anchor="iana"><name>IANA Considerations</name>

<t>This document has no IANA actions.</t>

</section>


  </middle>

  <back>


<references title='References' anchor="sec-combined-references">

    <references title='Normative References' anchor="sec-normative-references">




<reference anchor="RFC2119">
  <front>
    <title>Key words for use in RFCs to Indicate Requirement Levels</title>
    <author fullname="S. Bradner" initials="S." surname="Bradner"/>
    <date month="March" year="1997"/>
  </front>
  <seriesInfo name="BCP" value="14"/>
  <seriesInfo name="RFC" value="2119"/>
  <seriesInfo name="DOI" value="10.17487/RFC2119"/>
</reference>

<reference anchor="RFC8174">
  <front>
    <title>Ambiguity of Uppercase vs Lowercase in RFC 2119 Key Words</title>
    <author fullname="B. Leiba" initials="B." surname="Leiba"/>
    <date month="May" year="2017"/>
  </front>
  <seriesInfo name="BCP" value="14"/>
  <seriesInfo name="RFC" value="8174"/>
  <seriesInfo name="DOI" value="10.17487/RFC8174"/>
</reference>

<reference anchor="RFC8259">
  <front>
    <title>The JavaScript Object Notation (JSON) Data Interchange Format</title>
    <author fullname="T. Bray" initials="T." role="editor" surname="Bray"/>
    <date month="December" year="2017"/>
  </front>
  <seriesInfo name="STD" value="90"/>
  <seriesInfo name="RFC" value="8259"/>
  <seriesInfo name="DOI" value="10.17487/RFC8259"/>
</reference>

<reference anchor="RFC3629">
  <front>
    <title>UTF-8, a transformation format of ISO 10646</title>
    <author fullname="F. Yergeau" initials="F." surname="Yergeau"/>
    <date month="November" year="2003"/>
  </front>
  <seriesInfo name="STD" value="63"/>
  <seriesInfo name="RFC" value="3629"/>
  <seriesInfo name="DOI" value="10.17487/RFC3629"/>
</reference>

<reference anchor="RFC4648">
  <front>
    <title>The Base16, Base32, and Base64 Data Encodings</title>
    <author fullname="S. Josefsson" initials="S." surname="Josefsson"/>
    <date month="October" year="2006"/>
  </front>
  <seriesInfo name="RFC" value="4648"/>
  <seriesInfo name="DOI" value="10.17487/RFC4648"/>
</reference>

<reference anchor="RFC8032">
  <front>
    <title>Edwards-Curve Digital Signature Algorithm (EdDSA)</title>
    <author fullname="S. Josefsson" initials="S." surname="Josefsson"/>
    <author fullname="I. Liusvaara" initials="I." surname="Liusvaara"/>
    <date month="January" year="2017"/>
  </front>
  <seriesInfo name="RFC" value="8032"/>
  <seriesInfo name="DOI" value="10.17487/RFC8032"/>
</reference>

<reference anchor="FIPS180-4" >
  <front>
    <title>Secure Hash Standard (SHS)</title>
    <author >
      <organization>National Institute of Standards and Technology</organization>
    </author>
    <date year="2015" month="August"/>
  </front>
  <seriesInfo name="FIPS" value="PUB 180-4"/>
  <seriesInfo name="DOI" value="10.6028/NIST.FIPS.180-4"/>
</reference>
<reference anchor="MUSUBI-REF" target="https://github.com/ogasurfproject-jpg/horizon-shield/tree/e41715d75704e92455400707df895efd82bf4ecf/workers/hs-ledger/nenrin/musubi-v0/settle_v1_16.py">
  <front>
    <title>MUSUBI reference settler: settle_v1_16.py and the modules it imports</title>
    <author fullname="Toshikatsu Oga">
      <organization>The HORIZONs Co., Ltd.</organization>
    </author>
    <date year="2026" month="October" day="23"/>
  </front>
</reference>


    </references>

    <references title='Informative References' anchor="sec-informative-references">




<reference anchor="RFC7493">
  <front>
    <title>The I-JSON Message Format</title>
    <author fullname="T. Bray" initials="T." role="editor" surname="Bray"/>
    <date month="March" year="2015"/>
  </front>
  <seriesInfo name="RFC" value="7493"/>
  <seriesInfo name="DOI" value="10.17487/RFC7493"/>
</reference>

<reference anchor="RFC7942">
  <front>
    <title>Improving Awareness of Running Code: The Implementation Status Section</title>
    <author fullname="Y. Sheffer" initials="Y." surname="Sheffer"/>
    <author fullname="A. Farrel" initials="A." surname="Farrel"/>
    <date month="July" year="2016"/>
  </front>
  <seriesInfo name="BCP" value="205"/>
  <seriesInfo name="RFC" value="7942"/>
  <seriesInfo name="DOI" value="10.17487/RFC7942"/>
</reference>

<reference anchor="RFC8785">
  <front>
    <title>JSON Canonicalization Scheme (JCS)</title>
    <author fullname="A. Rundgren" initials="A." surname="Rundgren"/>
    <author fullname="B. Jordan" initials="B." surname="Jordan"/>
    <author fullname="S. Erdtman" initials="S." surname="Erdtman"/>
    <date month="June" year="2020"/>
  </front>
  <seriesInfo name="RFC" value="8785"/>
  <seriesInfo name="DOI" value="10.17487/RFC8785"/>
</reference>

<reference anchor="MUSUBI-SETTLE-CORE" target="https://github.com/ogasurfproject-jpg/horizon-shield/tree/e41715d75704e92455400707df895efd82bf4ecf/workers/hs-ledger/nenrin/musubi-v0/settle-core-v0">
  <front>
    <title>MUSUBI settle core v0 (SETTLE_CORE_v0.md), revision 0.13</title>
    <author fullname="Toshikatsu Oga">
      <organization>The HORIZONs Co., Ltd.</organization>
    </author>
    <date year="2026" month="November" day="02"/>
  </front>
</reference>
<reference anchor="MUSUBI-NOTES" target="https://github.com/ogasurfproject-jpg/horizon-shield/tree/e41715d75704e92455400707df895efd82bf4ecf/workers/hs-ledger/nenrin/musubi-v0/settle-core-v0">
  <front>
    <title>Notes on settle core v0 (NOTES.md)</title>
    <author fullname="Toshikatsu Oga">
      <organization>The HORIZONs Co., Ltd.</organization>
    </author>
    <date year="2026" month="November" day="02"/>
  </front>
</reference>
<reference anchor="MUSUBI-VECTORS" target="https://github.com/ogasurfproject-jpg/horizon-shield/tree/e41715d75704e92455400707df895efd82bf4ecf/workers/hs-ledger/nenrin/musubi-v0/settle-core-v0">
  <front>
    <title>MUSUBI settle core v0 test vectors</title>
    <author fullname="Toshikatsu Oga">
      <organization>The HORIZONs Co., Ltd.</organization>
    </author>
    <date year="2026" month="October" day="18"/>
  </front>
</reference>
<reference anchor="MUSUBI-EXERCISE" target="https://github.com/ogasurfproject-jpg/horizon-shield/tree/e41715d75704e92455400707df895efd82bf4ecf/workers/hs-ledger/nenrin/musubi-v0/settle-core-v0">
  <front>
    <title>MUSUBI settle core v0 exercised inputs (exercise_inputs.py and EXERCISE.txt)</title>
    <author fullname="Toshikatsu Oga">
      <organization>The HORIZONs Co., Ltd.</organization>
    </author>
    <date year="2026" month="October" day="18"/>
  </front>
</reference>



<reference anchor="I-D.he-agentproto-verifier-evidence">
  <front>
    <title>Exercise and Completeness Evidence for Agent Protocol Verifier Results</title>
    <author fullname="James Siyuan He">
      <organization/>
    </author>
    <date day="17" month="October" year="2026"/>
  </front>
  <seriesInfo name="Internet-Draft" value="draft-he-agentproto-verifier-evidence-00"/>
</reference>

<reference anchor="AEOESS-CHECK" target="https://github.com/aeoess/musubi-settle-core-check">
  <front>
    <title>musubi-settle-core-check: a second implementation of MUSUBI settle core v0</title>
    <author >
      <organization>aeoess (Agent Passport System)</organization>
    </author>
    <date year="2026" month="October"/>
  </front>
</reference>
<reference anchor="ISSUE-49" target="https://github.com/ogasurfproject-jpg/horizon-shield/issues/49">
  <front>
    <title>horizon-shield issue 49: call for a second implementation of settle core v0</title>
    <author >
      <organization>The HORIZONs Co., Ltd.</organization>
    </author>
    <date year="2026" month="October"/>
  </front>
</reference>
<reference anchor="MUSUBI-REF-V119" target="https://github.com/ogasurfproject-jpg/horizon-shield/tree/e41715d75704e92455400707df895efd82bf4ecf/workers/hs-ledger/nenrin/musubi-v0">
  <front>
    <title>MUSUBI settle v1.17, v1.18 and v1.19 (settle_v1_17.py, settle_v1_18.py, settle_v1_19.py) and their vectors</title>
    <author fullname="Toshikatsu Oga">
      <organization>The HORIZONs Co., Ltd.</organization>
    </author>
    <date year="2026" month="November" day="02"/>
  </front>
</reference>
<reference anchor="MUSUBI-ADMISSION" target="https://github.com/ogasurfproject-jpg/horizon-shield/tree/e41715d75704e92455400707df895efd82bf4ecf/workers/hs-ledger/nenrin/musubi-v0">
  <front>
    <title>Admission records (a2a-admission-v0): ADMISSION.md and admission_verify_v0.py</title>
    <author fullname="Toshikatsu Oga">
      <organization>The HORIZONs Co., Ltd.</organization>
    </author>
    <date year="2026" month="November" day="08"/>
  </front>
</reference>
<reference anchor="BLIND-06" target="https://github.com/ogasurfproject-jpg/horizon-shield/tree/200d5d96a30ba8b2605c2486686f7059bbfb403e/data/furui-blind-06">
  <front>
    <title>FURUI blind-06: plan (ledger entry 231) and result (ledger entry 232), with the draw manifest, salt and counts in commit 200d5d96</title>
    <author >
      <organization>The HORIZONs Co., Ltd.</organization>
    </author>
    <date year="2026" month="November" day="01"/>
  </front>
</reference>
<reference anchor="TOOSHI-1109" target="https://github.com/ogasurfproject-jpg/horizon-shield/blob/e41715d75704e92455400707df895efd82bf4ecf/tooshi/days/2026-11-09.json">
  <front>
    <title>TOOSHI daily self-check, day file of 2026-11-09 (tooshi/days/2026-11-09.json)</title>
    <author >
      <organization>The HORIZONs Co., Ltd.</organization>
    </author>
    <date year="2026" month="November" day="09"/>
  </front>
</reference>
<reference anchor="BBV-SET" target="https://github.com/babyblueviper1/preaction-governance-conformance/tree/235aaab/examples/hs51-acceptance-set">
  <front>
    <title>preaction-governance-conformance: hs51-acceptance-set (four acceptance records of the 22 October run)</title>
    <author >
      <organization>babyblueviper1</organization>
    </author>
    <date year="2026" month="October" day="24"/>
  </front>
</reference>
<reference anchor="RUN-BBV-RERUN" target="https://github.com/ogasurfproject-jpg/horizon-shield/issues/51#issuecomment-6618068660">
  <front>
    <title>horizon-shield issue 51: the four acceptance records rerun under a later verifier</title>
    <author >
      <organization>babyblueviper1</organization>
    </author>
    <date year="2026" month="November" day="03"/>
  </front>
</reference>



<reference anchor="RFC9943">
  <front>
    <title>An Architecture for Trustworthy and Transparent Digital Supply Chains</title>
    <author fullname="H. Birkholz" initials="H." surname="Birkholz"/>
    <author fullname="A. Delignat-Lavaud" initials="A." surname="Delignat-Lavaud"/>
    <author fullname="C. Fournet" initials="C." surname="Fournet"/>
    <author fullname="Y. Deshpande" initials="Y." surname="Deshpande"/>
    <author fullname="S. Lasker" initials="S." surname="Lasker"/>
    <date month="June" year="2026"/>
  </front>
  <seriesInfo name="RFC" value="9943"/>
  <seriesInfo name="DOI" value="10.17487/RFC9943"/>
</reference>

<reference anchor="RFC9635">
  <front>
    <title>Grant Negotiation and Authorization Protocol (GNAP)</title>
    <author fullname="J. Richer" initials="J." role="editor" surname="Richer"/>
    <author fullname="F. Imbault" initials="F." surname="Imbault"/>
    <date month="October" year="2024"/>
  </front>
  <seriesInfo name="RFC" value="9635"/>
  <seriesInfo name="DOI" value="10.17487/RFC9635"/>
</reference>

<reference anchor="RFC8693">
  <front>
    <title>OAuth 2.0 Token Exchange</title>
    <author fullname="M. Jones" initials="M." surname="Jones"/>
    <author fullname="A. Nadalin" initials="A." surname="Nadalin"/>
    <author fullname="B. Campbell" initials="B." role="editor" surname="Campbell"/>
    <author fullname="J. Bradley" initials="J." surname="Bradley"/>
    <author fullname="C. Mortimore" initials="C." surname="Mortimore"/>
    <date month="January" year="2020"/>
  </front>
  <seriesInfo name="RFC" value="8693"/>
  <seriesInfo name="DOI" value="10.17487/RFC8693"/>
</reference>

<reference anchor="RFC9396">
  <front>
    <title>OAuth 2.0 Rich Authorization Requests</title>
    <author fullname="T. Lodderstedt" initials="T." surname="Lodderstedt"/>
    <author fullname="J. Richer" initials="J." surname="Richer"/>
    <author fullname="B. Campbell" initials="B." surname="Campbell"/>
    <date month="May" year="2023"/>
  </front>
  <seriesInfo name="RFC" value="9396"/>
  <seriesInfo name="DOI" value="10.17487/RFC9396"/>
</reference>

<reference anchor="I-D.kuehlewind-audit-architecture">
   <front>
      <title>An Architecture for Auditing Agent Delegation and Interactions</title>
      <author fullname="Mirja Kühlewind" initials="M." surname="Kühlewind">
         <organization>Ericsson</organization>
      </author>
      <author fullname="Henk Birkholz" initials="H." surname="Birkholz">
         <organization>Fraunhofer SIT</organization>
      </author>
      <date day="7" month="September" year="2026"/>
   </front>
   <seriesInfo name="Internet-Draft" value="draft-kuehlewind-audit-architecture-01"/>
   
</reference>

<reference anchor="I-D.gilda-audit-agent-audit-record">
   <front>
      <title>An Audit Record Format for AI Agent Authorization Decisions</title>
      <author fullname="Sankalp Gilda" initials="S." surname="Gilda">
         </author>
      <date day="16" month="October" year="2026"/>
   </front>
   <seriesInfo name="Internet-Draft" value="draft-gilda-audit-agent-audit-record-00"/>
   
</reference>

<reference anchor="I-D.kroehl-agentic-trust-aae">
   <front>
      <title>Agent Authorization Envelope (AAE): A Machine-Evaluable Authorization Structure for Autonomous AI Agents</title>
      <author fullname="Lars Kersten Kroehl" initials="L. K." surname="Kroehl">
         <organization>CryptoKRI GmbH</organization>
      </author>
      <date day="6" month="September" year="2026"/>
   </front>
   <seriesInfo name="Internet-Draft" value="draft-kroehl-agentic-trust-aae-02"/>
   
</reference>

<reference anchor="I-D.gravit-gevp">
   <front>
      <title>Gravit Epistemic Verification Protocol (GEVP)</title>
      <author fullname="Alex Konviser" initials="A." surname="Konviser">
         <organization>Gravit Open Network</organization>
      </author>
      <date day="22" month="October" year="2026"/>
   </front>
   <seriesInfo name="Internet-Draft" value="draft-gravit-gevp-10"/>
   
</reference>

<reference anchor="I-D.birkholz-verifiable-agent-conversations">
   <front>
      <title>Verifiable Agent Conversation Records</title>
      <author fullname="Henk Birkholz" initials="H." surname="Birkholz">
         </author>
      <author fullname="Tobias Heldt" initials="T." surname="Heldt">
         </author>
      <author fullname="Orie Steele" initials="O." surname="Steele">
         </author>
      <date day="31" month="August" year="2026"/>
   </front>
   <seriesInfo name="Internet-Draft" value="draft-birkholz-verifiable-agent-conversations-01"/>
   
</reference>

<reference anchor="ISSUE-55" target="https://github.com/ogasurfproject-jpg/horizon-shield/issues/55">
  <front>
    <title>horizon-shield issue 55: testbed for decision records and settlement records on the same traces</title>
    <author fullname="Toshikatsu Oga">
      <organization>The HORIZONs Co., Ltd.</organization>
    </author>
    <date year="2026" month="October"/>
  </front>
</reference>
<reference anchor="KUANGMI-CHECK" target="https://github.com/kuangmi-bit/nenrin-independent-verifiers/tree/bbe455c/settle-core-v0">
  <front>
    <title>nenrin-independent-verifiers: settle-core-v0, a third implementation of MUSUBI settle core v0</title>
    <author >
      <organization>kuangmi-bit</organization>
    </author>
    <date year="2026" month="October"/>
  </front>
</reference>
<reference anchor="ERRATA-K001" target="https://github.com/piiiico/agent-errata/blob/a28e50d76a8a59fcec844a10869d7b16d04de919/claims/K001.md">
  <front>
    <title>Agent Errata claim K001: three programs compute the same MUSUBI settlement on all 128 test vectors</title>
    <author >
      <organization>Agent Errata</organization>
    </author>
    <date year="2026" month="October" day="22"/>
  </front>
</reference>
<reference anchor="INTEROP-TABLE" target="https://github.com/ogasurfproject-jpg/horizon-shield/blob/main/docs/INTEROP_TABLE.md">
  <front>
    <title>Interop table: implementations and runs checked against the same bytes</title>
    <author fullname="Toshikatsu Oga">
      <organization>The HORIZONs Co., Ltd.</organization>
    </author>
    <date year="2026" month="October" day="22"/>
  </front>
</reference>
<reference anchor="RUN-POKE" target="https://github.com/ogasurfproject-jpg/horizon-shield/issues/51#issuecomment-6332669484">
  <front>
    <title>horizon-shield issue 51: recomputation of the set 2 settlement</title>
    <author >
      <organization>Poke-nushi</organization>
    </author>
    <date year="2026" month="October" day="22"/>
  </front>
</reference>
<reference anchor="RUN-AEOESS" target="https://github.com/ogasurfproject-jpg/horizon-shield/issues/51#issuecomment-6330218168">
  <front>
    <title>horizon-shield issue 51: byte identity of the set 2 audits</title>
    <author >
      <organization>aeoess</organization>
    </author>
    <date year="2026" month="October" day="22"/>
  </front>
</reference>
<reference anchor="RUN-BBV" target="https://github.com/ogasurfproject-jpg/horizon-shield/issues/51#issuecomment-6331414511">
  <front>
    <title>horizon-shield issue 51: acceptance records for set 2</title>
    <author >
      <organization>babyblueviper1</organization>
    </author>
    <date year="2026" month="October" day="22"/>
  </front>
</reference>
<reference anchor="RUN-PICO" target="https://github.com/aeoess/agent-governance-vocabulary/pull/179#issuecomment-6332546428">
  <front>
    <title>agent-governance-vocabulary pull request 179: Agent Errata confirmation of E6</title>
    <author >
      <organization>Agent Errata</organization>
    </author>
    <date year="2026" month="October" day="22"/>
  </front>
</reference>
<reference anchor="NENRIN-VERIFIER" >
  <front>
    <title>NENRIN provenance verifier: normative procedure (VERIFIER.md), interoperability vectors and reference verifiers</title>
    <author fullname="Toshikatsu Oga">
      <organization>The HORIZONs Co., Ltd.</organization>
    </author>
    <date year="2026" month="October" day="04"/>
  </front>
  <seriesInfo name="DOI" value="10.5281/zenodo.23136978"/>
</reference>
<reference anchor="BITCOIN" target="https://bitcoin.org/bitcoin.pdf">
  <front>
    <title>Bitcoin: A Peer-to-Peer Electronic Cash System</title>
    <author fullname="Satoshi Nakamoto">
      <organization></organization>
    </author>
    <date year="2008"/>
  </front>
</reference>
<reference anchor="BITCOIN-HEADERS" target="https://developer.bitcoin.org/reference/block_chain.html">
  <front>
    <title>Bitcoin Developer Reference: Block Headers</title>
    <author >
      <organization>Bitcoin Project</organization>
    </author>
    <date />
  </front>
</reference>
<reference anchor="OPENTIMESTAMPS" target="https://opentimestamps.org">
  <front>
    <title>OpenTimestamps</title>
    <author >
      <organization>OpenTimestamps</organization>
    </author>
    <date />
  </front>
</reference>
<reference anchor="A2A" target="https://a2a-protocol.org">
  <front>
    <title>Agent2Agent (A2A) Protocol Specification</title>
    <author >
      <organization>A2A Project</organization>
    </author>
    <date />
  </front>
</reference>
<reference anchor="EDDSA-TAMING" >
  <front>
    <title>Taming the many EdDSAs</title>
    <author initials="K." surname="Chalkias">
      <organization></organization>
    </author>
    <author initials="F." surname="Garillot">
      <organization></organization>
    </author>
    <author initials="V." surname="Nikolaenko">
      <organization></organization>
    </author>
    <date year="2020"/>
  </front>
  <seriesInfo name="DOI" value="10.1007/978-3-030-64357-7_4"/>
</reference>


    </references>

</references>


<?line 2668?>

<section anchor="codes"><name>Codes</name>

<t>Underspecified reasons in core v0:
<spanx style="verb">contract_not_verified</spanx>, <spanx style="verb">input_not_settleable</spanx>,
<spanx style="verb">contract_signature_not_strict</spanx>, <spanx style="verb">bond_malformed</spanx>,
<spanx style="verb">parent_not_supplied</spanx>, <spanx style="verb">parent_narrowing_failed</spanx>, <spanx style="verb">view_not_strict</spanx>,
<spanx style="verb">record_not_strict</spanx>, <spanx style="verb">anchor_not_strict</spanx>, <spanx style="verb">anchor_below_lower_bound</spanx>,
<spanx style="verb">walk_not_strict</spanx>, <spanx style="verb">authorized_actions_missing</spanx>,
<spanx style="verb">grant_action_name_invalid</spanx>, <spanx style="verb">too_many_witnesses</spanx>,
<spanx style="verb">witness_key_invalid</spanx>, <spanx style="verb">key_conflict</spanx>, <spanx style="verb">tie_policy_refused</spanx>,
<spanx style="verb">revocation_policy_not_recomputable</spanx>, <spanx style="verb">ack_window_missing</spanx>,
<spanx style="verb">finality_policy_missing</spanx>, <spanx style="verb">bad_expiry_height</spanx>,
<spanx style="verb">unscoped_approvals_refused</spanx>, <spanx style="verb">chain_view_rejected</spanx>,
<spanx style="verb">input_too_large</spanx>, <spanx style="verb">anchor_incomplete</spanx>,
<spanx style="verb">anchor_beyond_observed_chain</spanx>, <spanx style="verb">anchor_proof_invalid</spanx>,
<spanx style="verb">anchor_batch_leg_refused</spanx>.</t>

<t>Of these, <spanx style="verb">authorized_actions_missing</spanx>, <spanx style="verb">witness_key_invalid</spanx>,
<spanx style="verb">revocation_policy_not_recomputable</spanx> and <spanx style="verb">bad_expiry_height</spanx> cannot be
reached in core v0 (<xref target="settleable-terms"/>, <xref target="exercise"/>).</t>

<t>Deviation codes in core v0:
<spanx style="verb">revoked</spanx>, <spanx style="verb">after_expiry</spanx>, <spanx style="verb">prohibited</spanx>, <spanx style="verb">conditional</spanx>,
<spanx style="verb">unauthorized</spanx>, <spanx style="verb">delegation</spanx>, <spanx style="verb">forged_approval</spanx>,
<spanx style="verb">nonconforming_record</spanx>, <spanx style="verb">evidence_names_other_contract</spanx>,
<spanx style="verb">evidence_missing</spanx>.</t>

<t>Verdicts: <spanx style="verb">within_grant</spanx>, <spanx style="verb">deviation</spanx>, <spanx style="verb">nothing_to_settle</spanx>,
<spanx style="verb">underspecified</spanx>.</t>

<t>Dispositions: <xref target="dispositions"/>.</t>

<t>Outside core v0 (<xref target="admission"/>): the deviation codes
<spanx style="verb">content_rules_mismatch</spanx>, <spanx style="verb">content_not_checked</spanx>,
<spanx style="verb">executed_after_content_escalate</spanx> and <spanx style="verb">executed_after_immune_escalate</spanx>,
and the contract refusal <spanx style="verb">bad_requirements_furui</spanx>.</t>

</section>
<section anchor="changes-since-02"><name>Changes since -02</name>

<t>Note to the RFC Editor: please remove this section before
publication.</t>

<t>This version follows revision 0.13 of <xref target="MUSUBI-SETTLE-CORE"/> at
commit e41715d7; -02 followed revision 0.8 at commit d7f5a91e. The
128 core vectors did not change. The changes in meaning are:</t>

<t><list style="symbols">
  <t><xref target="computation"/> and <xref target="dispositions"/>: triage runs on every bound
record whatever <xref target="contract"/> and <xref target="view"/> find, and its dispositions
are reported (revision 0.9). The reference no longer adds
<spanx style="verb">unscoped_approvals_refused</spanx> when <xref target="contract-door"/> refuses the
contract.</t>
  <t><xref target="contract-door"/>, <xref target="view-verify"/> and <xref target="compact-target"/>:
<spanx style="verb">max_target_bits</spanx> is the 32 bit value written big-endian (revision
0.11).</t>
  <t><xref target="settleable-terms"/>: any of its reasons stops the examination of
records (revision 0.11).</t>
  <t><xref target="contract-door"/>: <spanx style="verb">requirements.furui</spanx> is not read in core v0.</t>
</list></t>

<t>New section <xref target="admission"/>, outside core v0: the admission record of
version 0.4 and its content block, and the rules of settle v1.17 to
v1.19 (revisions 0.10, 0.12 and 0.13).</t>

<t>Other changes: the date; the pinned commit of every reference to the
source repository; the title of <xref target="MUSUBI-SETTLE-CORE"/>; <xref target="why"/> and
<xref target="scope"/> point to the new sections; new section <xref target="revisions-09-11"/>;
in <xref target="impl-status"/>, the publication of the acceptance records in
<xref target="impl-reruns"/>, the reference's run results and file list at the new
commit, and new section <xref target="impl-use"/>; a line in <xref target="not-established"/>
and in <xref target="codes"/>; new references <xref target="MUSUBI-REF-V119"/>,
<xref target="MUSUBI-ADMISSION"/>, <xref target="BLIND-06"/>, <xref target="TOOSHI-1109"/>, <xref target="BBV-SET"/> and
<xref target="RUN-BBV-RERUN"/>; and the acknowledgements.</t>

</section>
<section anchor="changes-since-01"><name>Changes since -01</name>

<t>Note to the RFC Editor: please remove this section before
publication.</t>

<t>This version follows revision 0.8 of <xref target="MUSUBI-SETTLE-CORE"/> at commit
d7f5a91e; -01 followed revision 0.4 at commit dd64d0e2. No rule of the
reference and no vector changed between the two commits. The changes
in meaning are:</t>

<t><list style="symbols">
  <t><xref target="triage"/>: the row "strictly signed, body not strict" sets the
record aside as well as adding <spanx style="verb">record_not_strict</spanx>.</t>
  <t><xref target="strict-terms"/>: the parent row adds only
<spanx style="verb">parent_narrowing_failed</spanx>; <spanx style="verb">contract_not_verified</spanx> is decided on the
contract alone.</t>
  <t><xref target="same-block"/> and <xref target="sec-unsigned"/>: the same-block order can be
chosen by whoever holds the copies; the verdict does not move.</t>
  <t><xref target="not-established"/>: the earlier settle layers disagree with v1.16
on 51 of the 128 vectors and stop on 3.</t>
</list></t>

<t>New informative section <xref target="related"/>: settlement as an Auditor's
evaluation in the agent auditing architecture, with the roles and
record types mapped; SCITT combined with, not replaced by, the Bitcoin
anchors; GNAP and OAuth as the source of a grant; and related
Internet-Drafts. <xref target="what"/> carries one sentence pointing to it.</t>

<t>Other changes: the date; the pinned commit of every repository
reference; the title of <xref target="MUSUBI-SETTLE-CORE"/>; <xref target="why"/> and
<xref target="clarifications"/> carry revisions 0.5 to 0.8; <xref target="impl-status"/> is
rewritten, with a second implementation by another author
(<xref target="impl-kuangmi"/>) and what each implementation read before it was
written, runs by others (<xref target="impl-reruns"/>), the run of the
first implementation against all 128 vectors, the reference's run
environment and its file list at the new commit; new references
<xref target="KUANGMI-CHECK"/>, <xref target="ERRATA-K001"/>, <xref target="INTEROP-TABLE"/>, <xref target="RUN-POKE"/>,
<xref target="RUN-AEOESS"/>, <xref target="RUN-BBV"/>, <xref target="RUN-PICO"/>, <xref target="RFC9943"/>, <xref target="RFC9635"/>, <xref target="RFC8693"/>,
<xref target="RFC9396"/>, <xref target="ISSUE-55"/> and five Internet-Drafts of related work;
and the acknowledgements.</t>

</section>
<section anchor="changes-since-00"><name>Changes since -00</name>

<t>Note to the RFC Editor: please remove this section before
publication.</t>

<t>This version follows revision 0.4 of <xref target="MUSUBI-SETTLE-CORE"/> at commit
dd64d0e2; -00 followed the first release at commit 4100d52a. The
reference did not change between the two commits. The changes in
meaning are:</t>

<t><list style="symbols">
  <t><xref target="scope"/>: what it means for a contract to carry an out of scope
key; a malformed <spanx style="verb">approvers</spanx> value refuses the contract instead of
making the input out of scope.</t>
  <t><xref target="reading-json"/>: <spanx style="verb">NaN</spanx>, <spanx style="verb">Infinity</spanx> and <spanx style="verb">-Infinity</spanx> are read as
non-integer numbers; other spellings are parse errors; <spanx style="verb">1e400</spanx> is an
infinity; an integer token over 4300 digits produces no settlement.</t>
  <t><xref target="canonical"/>: item 6, how a non-integer number is written.</t>
  <t><xref target="contract-door"/>: rule 4 (domain whitespace); rule 10
(<spanx style="verb">delegation</spanx> needs <spanx style="verb">allowed</spanx>, <spanx style="verb">ordering</spanx> may be <spanx style="verb">{}</spanx>, well formed
<spanx style="verb">approvers</spanx>).</t>
  <t><xref target="settleable-terms"/>: four rows that cannot fire are named.</t>
  <t><xref target="parent-narrowing"/>: new; parent narrowing stated in full, replacing
the pointer to the reference's <spanx style="verb">grant_subset</spanx>.</t>
  <t><xref target="view-verify"/>: rule 9 applies only to an object view with a
<spanx style="verb">headers</spanx> list.</t>
  <t><xref target="triage"/>: an approval's <spanx style="verb">sig_b64</spanx> and <spanx style="verb">signature</spanx> are each checked
for strictness on their own.</t>
  <t><xref target="size-limit"/>: records are counted after identical bytes are
merged.</t>
  <t><xref target="batch-leg"/>: rule 4 reads <spanx style="verb">NaN</spanx> as a number.</t>
  <t><xref target="core-fields"/>: <spanx style="verb">contract_sha256</spanx> is null for a contract that is
not an object.</t>
  <t><xref target="revocation"/>: a copy dropped by <xref target="collapse"/> no longer names its
revocation.</t>
  <t><xref target="approvals"/>: step 1 can apply only to an approval that is not an
object; only <spanx style="verb">sig_b64</spanx> is verified.</t>
  <t><xref target="nenrin-cross"/>: an orphaned execution is still read.</t>
  <t><xref target="codes"/>: the four unreachable reasons are named.</t>
</list></t>

<t>Other changes: the date, the pinned commit of every repository
reference, the vector counts and verdict breakdown (<xref target="vectors-contents"/>),
new sections <xref target="clarifications"/> and <xref target="exercise"/>, the first
implementation by another author in <xref target="impl-others"/>, the reference's run results in
<xref target="impl-ref"/>, references <xref target="MUSUBI-EXERCISE"/>,
<xref target="I-D.he-agentproto-verifier-evidence"/> and <xref target="AEOESS-CHECK"/>, the
comparison with RFC 8785 in <xref target="canonical"/>, the advice to implementers
in <xref target="sec-numbers"/>, and the acknowledgements.</t>

</section>
<section numbered="false" anchor="acks"><name>Acknowledgements</name>

<t>The parties behind agent-passport.org and kuangmi-bit.github.io
published the contract, the anchored record and the header view of
the first MUSUBI contract settled with no party from the project that
publishes the reference; the real vectors are built from them. Two
adversarial reviews of the reference found defects that this
document now describes as closed. Tima (aeoess) wrote the first
separately written implementation, and its questions and notes led to
revisions 0.1 to 0.3 of the source text. Revision 0.4 was prompted by
<xref target="I-D.he-agentproto-verifier-evidence"/>. kuangmi-bit wrote a third
implementation from the text and the vectors alone, and its two notes
led to revision 0.6. Poke-nushi recomputed a live settlement from the
published bytes with headers fetched from two block explorers. Pico,
the agent that maintains Agent Errata, ran the blind run that led to
revision 0.11. babyblueviper1 published the acceptance records of the
22 October run.</t>

</section>


  </back>

<!-- ##markdown-source:
H4sIAEOo8WoAA+y9e3fbVpIv+j8+BUY564acgBRJvalOz5FtuaOJI/tactI9
3X1IkAQlRCTAAUDJiuz57Kee+wGCsnt6Tt85a93MmrZIAvtZu3Y9flXV6XSC
Kq0WyTDc+enD1YcXF+FVUsHnZZJVw/B9Ms2Xq3WVZjfhL7dJdZsUYZyFZzfw
a3g2rZJZ+Eta3aZZGIdX6U0Gn/9QxFm1E8STSZHcN7W6E8zyaRYvoctZEc+r
Tn4Td5brcj1JO6V5qtPbC6ZxldzkxeMwTLN5HsATy7Qs0zyrHlfw9sX59esg
XRXDsCrWZTXo9U56g2AGLw3DQW9w2On3O/3DAAaxF9wljw95MRsGYdgJbS/0
McbZlPTn+WxwcNA/ob9fpNU0x5ll09u8gBWgb6dxlmfpNF6E/3r19jKI1xX8
SM3C/4fhfL1Y8Nyu8/I2vYurch2+vYnpx7y4ibP0t7iCKcDKXN8m4Q9v31/8
29vLMnyZd6PwTTXr7tCjyTJOFzCx2wT7/i3PulmW39Ob/zOdLvL1rAs7Q49O
83VW4SL9a7yKsyDI8mIJD94nOKr3r18O+v0T+fO4f7Svfw4O9Nu9w4H+uX+4
f6wP9PYG+Ofri3dX/eNeh14MQyWWq2S6LmD8cXkbXlVxNouLWdi6+uGqzROw
CyMzH4aXNHxYuYushGbWVRLmc/NyCQs9C6+T6W2WL/KbR3pTN7N/0Okd0zdl
UqRJifSgbeMAh+G7Dy9CGqZ8++rtxTDs97qHvcHx7uXF1XUXn+vqI0yVnffn
r/1pCbUWyTwpkmyaCK0AjfEfo/v+qH/YXT3SaGF7wmU+Wy+SMkyrMF2u8qIq
mxbgWbr4myjDpe9eZ7DHw4+LmwTO621Vrcrh7u4NHMr1BElkF05XuS7mqyL/
NZlWnV9XN7tCUh0YSLKY7VZFkuwm+/2j/sHs6OCot5+cDPYPDvZ7vaPe0Wx+
fHKQzGfHg8l8P5nOd+EY3SVFuXtbdhbJ7CYpdrMkg+OxK0f4vrdbW6ogwO3y
ifJo/2RP/zzZHyjNHR0fDO3uXJ1fX78577x8+/68cZe4HzgAQIj3PaA+en6E
z4/ue93lrB3BTt6nyDLCXre/94/bmX4HWNF/153p4IrBR7vSl2+vz6/8Nb7M
K6BqWLj6KtOzuLj//2puWc2fz19ev31/9TU0C2tchfcw4Lz4h/KN/vH/Ret5
/sfz9y8vrr6KCSQfk2KaliCIpBkILmXY0m9G/IXybm20W32s/oGk/H/L0l90
XnVvkw7JRjCQKu/cw807T5OiAxx1hncjLtTZ+dvzq6vOyx/OX/7ob48n0XHD
09tkejcEUbEEsRK2AO5LFsNoEVEYaNzTrQJFnORJCTvM4ui7uCzx/g2vHssq
WbY3l/5L687t7W4bOa7K1dWH887+iT9Vf3NCkFHXSQgPgbC4WIRw9T035a+c
69cT2N9PXTSBcnf/xBOUOj+LILntCN73u/2jiP45piOGf52ELUccOILTF7mi
1HH9ixP4oq3CVVr843njf8+7xu7E2aufgAov3l76W3E2E90IJB6gJRCnW/Eg
7sT6NbTRHobmZbi/aZHN7yM63o8oN60e/6Gr/d+QHcKIXry5uHzV6R36q/z6
w/sPF+FkkWYz/C1cLUAbbnErYYJqWDjY6zP9Fkm5XlQbvw5AKH2AyZHyAArw
Q7iEhZuDHADnIIYX8F3S6UCnyOCv5RJUC1BuZwezk8O/nz/Aivf/i1ZcBxXv
9Sbx8WRw2DuYDvaPDw+PD+dHvYOTyWQ+2e/tJbvQf7w7XxfrtKNrB0O4fvv2
6ocLGFGvxlb4Bxh1ungE3rCYM/+N4JvHcJ4uSG80swEOU+VIktDNY7lrv+/+
WubZdnX0b1mxk79/xSaLfPL1NPrMjJA4X/yMypG/aqsiiad4/Do3ORzmLIYr
Gm4vVr3wug5vy4N+J55Ok1VFPwLbDVvzfA33k/nSsA9YYqTQwSB8Cxx4AhRc
rJ9ZzUk8eZws1iAbrJKi36Sp7n9pDf0mdr80H6HBvYM4jmFlP8Z4teKp3pgk
KpcfLju4au/P4a+vuL0P+mR8CbetTpHAaoTrbIYGsXABM8WrigWk/+wa9dHm
9V91eR/0v6G/kH+gOe3wsH/cg4N52GNV++RkXxXwk8O9A1XAD41afrJ3QswP
BcG7dXK7SB7w5MbrWVp14mJ6m1YwgnWR6EM36WIW6+8okcnfvGSmqSKHtviB
dNohy10njm0rRXwPL90k9yv9apIWd7f54jeRQOPJQuRSpAb4rqRLpxwa+ezg
4Gt2+GBI2tcEVAUU0WbJNPXuT+TD1lJoz0VGhFHCVRhWRTxN/pF6238deRxA
Uz9+OLv8w08XTdI734gd2PFklcD/wGIreZdqC1NtIYITUN2mxd8v0N+t4+xm
CZdEWv3tc3de3n1u+Mw3JpMEuO90U/E5f//+7Pqs82Ov169JWKRlnBcFXGbh
dBGny5AegrlDgyGsOdDusgzZYp5YKvHWgGgJ1gY1g/7g+IsGAFoXt+sm1vpF
gXWVwn/TfJePTUIN8Y0UD46Tg97s6DA+jg9O5tNkery/H/d7wAlmR5P+4ay3
P0tO+ie7NONyF6cMwiOetcvr8/dv33Wuz168qanmFxnww3wFI5rgZ58s+GQB
+4Slwmsdzl98E6cZrINZsslj9Q89V1+xhF95wy9hJruzfFruyvqMaH14yfAW
evf2x/OvvIAK8b6Y00TrA5f2wCGmrUTzLr9LOtkaWv0/NN9tN83e3uDw8GT/
eF9mzDaCr5wz7nyI9gV48tGfM10n2w8Ja+//8Ln2Bv3j/uGxlTG+cqINQgVe
QzTVv0PK+j861/5+f/+g31dKvnj51p8ssxdHXLvPp/FkvYhB81nByYWZ/vsa
GV7/6MTnaiEKdin5CJjUz7erOn83NxQ7zzOj3cXR7sIoN0n7YP9wf4DbfXl+
+f7isvPz+fuL1xfn72vWc/oRL4X7hBo30uEwNA46/HmazNCP1tJm2GuRMgNN
iniSLvAgyA0hiqU6p8yd9o9klT0W5DcdcepyOxgc93d/S7J8lndBG947PDnC
9Xpxcf3y7UVN+BYnK2xq+C5Jik6Vd/Df8HwBEy7Q0xq+JBcj2fSen+ZVXOFE
w8v4Ll7mVe6NfYuZYcL9d2E1zN+r2dwOt/PD+dmr87oxX33Dr5L7ZIH7FL7X
TRmGLxb59C78IYlnW3aGqFibeMcH0RnsPF6USeNoZ9pd1x23oQe8gaZ3o+kt
3EHd22q5gEbevju/vL746fzq+uynd7VZvAXZ6DpdwoEE3Wn7SP3HvmqgMEZg
4PoKjhIeOxucNUhUAz7OLfi1jYtR5dN8EV6tQByfp1Miz+2cYHD2N60fWsRW
0oUM6vzVq6szkGF+urj8Q80KES8R9kA+3jh7DM9n8GTjKnXk3xDOLUjHP3bD
l7fx4i6Ny+YHXnfDP8RFuljkVfMDP3fDy/QuX8RJdueT8aD3hcPX7/WOduHA
dfZAm+x1Dvf3Do46R6P9IOh0OmE8KVFnqYJgUyhNgbuAFgSMB+adlqCdwdnK
SAOHRYhBOovvEnwGWDU3UjLiY/IYVg95uIqLCgYVkbG7ItM2/+5YFB6wnThj
yEU4S2cBq9DUvrYbsWUyvE+TB3xJTwpRd3jL54oeCkTWLqFhBqfgbnF/KFZO
qzIEEnyEvx8YpoK/3yBCBf8KtMfwBrhx2Q3Pssc8S0JQNme69VYcReiHEclI
vA/ExIdt5+uKYSj6Ii7HI08MSG62nqJLquqG17ew0CAgrgl/UjKdk6c1CR8S
uCNnyTzFZcMGWACCJnQnopoqFQWrAhXXZPEYAsdd39zy89scDjiHCXRUpFWV
ZOG8yJcBjrZKPlYG0rBaTxZpeQtjcLUTUFlgjN3w7boqQUBzBxKgEphCA4sy
h/GX0yKdwJSwrbhml8YZoc7OLvl9vOpg89LpbRjDci4ecfl46WDjKmmE9ATs
YAl7jy3EKkl8WxLZIDVNHoNZegPf8S25jIs7phLYzwfXR3ECL8ezknbjyqP/
DA4kKCbIuWA1QRyb8k/YCP8MPf+6nt2IGjcPF6nc0F0+YMt0NlskQfBNCDoQ
bTst+9M3KX78DD98E/6CG+R0fFHC73gy4OczIJY0m6areCGnQGkUZUM4T843
SEuJ+QRbWyBTgHUkAmf7FVvQZBVtS3jWiYWlv+Eu53S2I/eNIKUVAdK9TeGy
QSMJ0Ao+BzsU1dqG+WQJHjhYZtg82LcgXqHsA7Pg15CqdGKROR84XOj84RYe
Qas8bNojkGcwSxbJDbQyw/0Lk4+rFATI2yS9ua1kVYoEhTVa21W+SKePXVhW
tAy7xxom2iLnGxzQDD2FKCegY9iygTb1WxL14vKWyq0C5VYp8SkmpCQGOhU6
TkuxzVeygtCkx6kCuofDpyeRJD5/RksBHdFYUWXd8A2aDrF14jywEsR9Sm8e
kcPWyhpz5F8WyT3yNI9D4lkPmOyH4QPcR14rsgjIKJlr8rs8ZmKH+O0yIhYb
A6+huSOdGecFLSAvjHA4HFdSzFLi4XSA1zTeIA4nyIyASUJzCZyVC+4eLxtj
auZLgbQ85AKujTF4evqiHfLz56h+nWXhGT6WF9+WAazQYm3UC70v3LuCiHnz
utCDQu8G8CXxhTBmcwXyilPYZNgDJFgeF2z1Ml7xqYOrEeZ7bZY+LIGbLWOi
xZIO7aMcD+D7H0GHGsQ7cAam8bpMxASN6gLsYAnnQq5U8ikTbJFPH+kLuDmo
zNBbm7KVCj6wlvD582eP+TFHzHISqWBgJbROOzsDVYkYH1vS0HQFj3SFkT0i
7pMvmgv/orkEfoDCwjVeLMjeHoG7fQDWukBbob0BmbZ9M6ERO2DO5jrMM7ji
gMe/e4TtQL8Y3ATWtYyoDjiENeRZxFS7gC0tBJ0n13MAp62EZV7A1ZACF3h6
so5uXJqzmuxDveMJFTMfHS4VBUhyqvIcOBtfEUbG9G97vdL0qg/qV31abdzv
eFBXCzyUizi7WcO24OBQistukiIoQfBYzPBKR1MbMVO42pOt979wYxiU3vh0
/XtXfGQEGqQKlWeMliHUTO/jBrmyhP8kbRKvCf2Y0kXdDX4p+ISjxCCWP2fj
0AGNfhdkKwu0r+NhLRK5wmZwtZDlRBE+eCECfwI1cMuUH2iFcCogH8xpaNXi
kfkW7Sh2In3EekizfJLPHlWCnVITwBthzWW9iMuFLVhrYGCXfVz5y34fN9GQ
EgHlPn9udzewo3iWddngxeki57O+jEQGc4nGylMgsMDhlRc7k+Q2vk/zdQHc
BsS1quT7ikc5w/tQV81Bs8L5A7LDM10CI0YedxOW6wLu5lLudY9e40d80MpT
MITgl/cX19fnl+GL8x/Ofr54++E9dCVzMUPCI0F9OcYO5FlC1cNgG33SYeNt
IldLivcNCsUpcCQ0s+vtEwjJEjW6Y2a+H8/ilbWbmi1xEKWwakCPdG8HZK/O
kqrzCgHpwnHdFUT9hkeF53GZxJmsViwdBklR5AXfqs5ohAGowDvPQed7KH1Q
qsr4AZ6obvhefivxR6SOXncPOigfYCyzkIReY0SvLV6GkM3AnMN5WtDJYokM
+JeSnP9a0Hp6wm86Od6IJRDsqTvAfWOlZzGdFDZVW2Ygs8elCpiGi8B9HC+g
QxULoAdF4XnN4xwPYI5Br3uMtAGHD881sckl3W7xhBQrl4gj/n1Gm1ImyFOm
JPY2+qDi8o7vS/MaCdAlqg94IkT2WGd81AOWy/gcsrjIfIdQGEbekqGA9kFw
BDx+oLLnYbEmcELgnzpqK5fl0QMaUmiBmriXXSDS6SIujNEDNiLgY013GHEp
Sx2wcCfQ98dkk4f5faOOTQyYSC9UXS8SS6Klw767rML4S/N8cBvPwkUChwNN
O6cN02mhDCTb2umddPp9Yn2GnAO0TUTY04D6JtKPZyh7FQSgz0WvNNBYPFw4
EKtDBiK9osBlvoSD7Kmcy+42LesyR1EEVgNEkQ3FL2xQ/LrhBWo2KFOVVb4q
remCFGeUGbObU2RSLD8ZEX1eiRg2R6GMPMnISpusFAFZKfB0b9gouk3D3KaA
gigPva0XsrWotojoFixBufM0QIQsgIK1qKmSequRPGPkH1ZNAt/6YZiMtY8g
BUNvHTT6if3Av5mYa3MXRqw0T/OuXU2BvnBeL4UKnr4p8SuzY4Y8YJDJRxg6
8jVp3ZM8rGEo8209qidneHDkRPPpgM0cghofjtG4kBbMgrrlCmS0cdii2eLf
IaqmkWNROG5HG28xNOEGBwPvIs6L7g24a8l4ZV8+4Zd5w1VrHolSy5/hDWgD
ukZZxHzlNtLvua0sUmBW+ApzbCLnsjY+c3zGDYO3P6qITZ3QdYR/7PO5TKtT
WtrQntBShDI0HcRlSeyTdV4gJ+YUbVT/KtGfcZOIFU5AqUNhZIly/Q7It0Al
O6zO6DaTaY50W9J+oZm3a7YzIoUIyyZNhaVq5Ia0n8DRcxRvQD6pHpIEQxqg
XSB/p7USkVAFKsB5OkWKZoME3D/3OEFjhVqBfl/RkvkjxvM0fdwJW8oDJiTB
QjtKeKOUlRh6eZ5UUxK+4HSRFKbamKFvZjtVvoZumQ+DgIJj6obn8AvGR90l
j3jOQVEscTlBkEb+6R70LlqU6LFS6H7GJ0KtOwleoCxUge4OGjIcCWFRcN6z
9WIRBSxywa2QfETBG3e523xGUJ145higGcEdRaLKqoxQhhMF2C9MZ7pYA4Mc
ChL4Rs2BmYhD4Zi6HTkdjuClHAWAKhkHSA9jZwDbHkS7L4z7iyfQWUNSB52F
jFmfQ8HROaLE/IAiUCjo91gQ1Z3pzPK8IGGIhABac1r/QIVKXQ3aBSLu8Z//
OjbLgqFNc9p6j7OnWbDRSdjakNnhZv5ZpLVx+ZhVtyNmGiPsbBwF8qWFB/P3
+rBZlVE+Qe+L84b5Bc9wdtP0zm9JkY+RC8wda6a1XxqhCOmi5CNY10zM3caH
Xbg/WuNxN+tCIEj/1yGoYwLbkSMWOGBC9r2vp2SbIxbW3SZjRJ6kQqZv6p2E
GBcKr9zyBMUcCQMwsxXphgiyDExHeAvC3Qc0m1kx/5pcMhQVCffh1P7K9uT3
zoEL34iVAB6cTFf9/c+sreMJeyC+hKD9652I/8VFwb/fn/+/Hy7en7/Cv69+
OHvzxvzBTwTw4e2HN/I7/mXffPn2p5/OL1/xy7jIta9+OvvTDntrdt6+u754
e3n2ZmdDUSLeAOs1SdiWBcRPsnpp1h1vieDFy3ch3D5PTxLXCvtCf2NgK/yN
J1IVezmgtDmPeCSTmPQzUExAllmlVbxA9wBs2W0OdzNeFrD4P+VlJVvZOESR
kUVKwxtgytptayctcZ3IdIrThj0vd9okWD3Sq1YTRuyXyk5DHp/63dSAXaD6
lqAhuIXHuUg6/BE4xvMOHtxVEJNV85K7k4Urn47I9grkoSQ5DNALj1HNIR9q
vhbEXDlG36nhK/c9ONfW/Rdb4z4KAzW/hejJer5d8dbxCdjncSDXonQ7XgLb
ZCQAIVRAyftI98O4yBd8B6kDT+5tCeoWSuntDYBSiLOIRCWb6dyZ4i/gFcGj
YxwHckvIAwncCjQ39FuQdwbU3Om6cmCk0JoCYIfImewDLVpS8xnXtB15rg0Y
Pz9kv6KnQmIljgsknt5l+QMFGkhU+8Z78Ai9C+NxpB6eIV1TwHIxGOG4xxCs
RqerdWYoLAIPHVAJrzRsAjtpcNboceDmBQtDLggxrz091dAz0M6UvExITc4q
UcMsQI4Z1Am3+BwvD0/ex462BWijhu6GZ0dquherTEnzZ/4ShmILYd8FsRfP
MBxsmr94ktb0VTcSOBTChjY2M+KCRkBcK9wu196vVhSgTjIIw2aTCbOLcdRV
cpMUzkllkU0NO3HJ1xc9BWfmLskQbgTaWize4wyluBVI0nj8wnBcFWs5M2OC
TbCQxpKZtFOKwdfoSHjvMt0Bd1yX4o4gquE7vxQTNdl4M2Cg68kiCf59jeap
SG5ZkDt2aizlL9kOcBU42bn6fc+uXl5ciPddZHfad22KDR7jv4CeUmIaAYbQ
oZ6wQFY6R59kq/exd9YmY14Gw8Kpvghwku6AQUg5C78LX5CQx5QMA8P8ExnN
FOTd8jYeHBy2PrbpGbgbO/ARZC2THwGIRbaeB/yxG77ALshSX5pp7/Cv+8jS
9g53cL0DFIg6kxiXktyMH0nEY8PUCtFFpExEaIql2yIj8Dn8ujeQzuIqyOdz
OALUspyYvYNuuPMmB1F5Cm3D2fy4Q7ZLdarHuPB4rMe9MQ5nfCJCc8wf57Ad
xIQT3t4u+7UxijZCGADF1+oxu7qNV0affungVkGDx59YVuG3yR+Of8CX/S4J
6UQE4yFtnnsHtbbdP8DJBvBqgrJQOcaoUmVj9mAAIZ6j8Qw2DOddiYShVv4h
Gm7D8G/iyc0cOSBE0rNM+RmWjLgTUgJpMHAKSZu7AYEB7fhvyUVrrRZqfu4G
ezB/5OKby4YNjZ92hG3vDMM/4wdkzfD3DyCdIC0Mw53f9Q974cKlkN/vfI7C
brf7188g00MrX7wTYICkvLCsDudKBx7sdx2eTZchDnQlSUAaditsOXdEyRcC
jgHuBLMlJS7XpFRURrJcVY+0ItYaJfr0mTBHsq0jskrtZ9QznnNgwiJkxtRi
N4ADM0Zvd1Y5QzV4BezUyiCIeHBVQVpzI0Tw4yND2a66DSIdGzg6LIN9brMy
NPaepsFmsqE8Bt+97Vvi+R6mp2D6C7yA0DJvGx1BgyNBjM5A6cUO5fT4axMJ
9qOh/XBM55aa4sGg23GMtm/2JQn1Bq5pJ0mJQJDCYJ4/Kp1wh+qeUKXr2kWK
fFtKNKSKht+WgUpzBFKaWrIbi6j45/SvXf5tBL+NEn58NDncHzsWEgdbgT0o
QMS2VpPznmm1G4zxq3Wx4HWku5PsO3h6X6LRiH2oInagPFkmxT1SXhWadwPQ
4uAGZlpjwVTMIvkm4Korih+7Z+kIPX0jRvsOBgl+RogF8+eCaLdkQudHnyQB
kEhvxD2Sj7AepPoMLdkFYhmarVcL1HSMqQZlNPJ4iW/Dp0wFvHXJ/R/AEiIp
zJJkpZalg/4gXCCaldAf4z+PkX2On8Zhi0Jfk5mdNd/Q7QB7XZAzkruGNbhc
Lyd4f5EqK6ojSixqSV8t1iU2DFoP/NEa/K+DvbAT9ts+cgpbVh8Xyg+k9RFy
p2ZHiVz0zNMTDAwUbfyaLNsIyKHv8S8ON3qEBSbyArYmLzFDZIiBK8HROgdZ
jkFLLL5lNDk9msYkKtoKMJIyYZOSjkMd8yj2wZ1+GV+i8eUim6dZWj2KlNdx
PheJ+FdK6YzRJIGeTrOcm/ajJleYyLSC0IAvkBAlP1Lb8mNFVyLRsbQFjeTS
Jc8S/S6B4UlEvbLndatVFmcjuF9YFR2hz96aolKZ6dYHMgSZgUK+9UkQH8T2
DKMk8bxIEsRsLhbkHdMVjIi50Z6k2CBefHGGq/+dXW4kcO1oTNTsziz60kk6
U4Kg/UWkWAlERyuKq5jn4QLRzmJnYpmbbNqhhMaG436y30P1MS3NvseIaOIx
kXiZVtQUjW6TGNGcXNMvQHrIbvRU7+/1euEsvUmr0mUSOregeW5DoSTSSos4
RZtmo8XyGkmF6UK0pMjVnIqED4u7hIiuYAAbmcKdE8fj25HJNFik6GJP52Lh
FRUr+NtUrC0KVtCkYJVqQyQzOUNCcpjguqJzYMSbGITvi45l5JhKiw4/Wnhq
sjdM6i0cs3dwxcKFtrol86H5GS6Jl+LPM0ICqzkoogiYFi1OOHx0dxj+N0HH
AwIHxAlCrt5r4kMhTu1R3B+i7rcsk8LsB8Bj8CHHUC18knknfACKnA1Nb5Yx
LB59qxMChQM+hb8ydkBCqGK6O27Vqyo63SwtVzkeUfRH0xTtZxqanEGO54ZL
awWNqVeZr3LPVATiCCMYW+5UeL34AqDHHYMLPslzbKsfB44BHNWlQJ7gDM7j
gr1jiwXN6dR1VS7FE03gTST6gK4ua65c6V6ThMcURY67Y74t/NuJnXSI8lws
4lWJU1sRGn1O/FkXG7FQNEd+SraRndmw4miI9zaT50JAOJwbDcV8ItQYP9Hk
rYhgkFa2VBEZu6lNJqhNho6eReci2yM/4Hxd4ECZzbhbTjjXLABSsvQpdGhn
ypYFlr/HqL3d4e0A/Jyw8TQKRIhSODufKpqB41swjTK2BCczyxVywMIDyeFE
uDhw1Gk0CMI7R2gtCjeHS36QIl+t+Fx4o7+NiSU600YXpQWRtEV/ewC2O2Ra
Cl2/SKs2Fdh4fkgmgw+YeRHeg88sBz4AdRG/w8Uzc4wYdazGPnxXwd3k3VOc
LHEiPKb0JiZyoB5+USizXTQ3fpijO+At+oNP9YxRjmKIlAPTssKYnHxMATBT
I5wA/IDG+E98KhBIMylwbCAU5uigm/FX/NyRz8wb1X9cMsn4+l/EwKGThhMa
uZvdMvhEktPmcQryM10VGDSoG9Eh/RwfZ3wrzod0fCamtlCjTj6AVtn2imJq
m68f5T1G3MSb5yF+FCig4EmcRKq+sngqdioDnVEUVT2a2ewh319sVIJbKqc/
OugvJHbJoi1/SzqRb+9AZx3/aCAn3KH1o8TBdodJuOkwYebsKLUVxfuoTZdQ
tuJ1N7kt2XlnEsS+hu+j8JUTk4IpcmPEqqMFbFo8rioJCXnpZZUNW5KJyCSb
xexN+Ip+hrfG5kPrvq12y/DD9evOMcsFmNYVVgOGm7O8PQ8N8HdIdre3tHbl
kNXeMi/EAP8Bms1njLYNCfBANnO+2ElrQ9uxQX2On3budob3Ubfb/Tzm6zbD
oA5coRWi1tBQd1YU8SN0Nf5zHE3w0b82P7rXDa9Y4xsas7bCNsY7KJGXUzI0
onQGX5Al+S9oOB7/RT78hS1XuOAfvuvBf3hh4F/91zBnXAU4MBH/dowLN/7L
ZCyfX/LnOTdBX53xV5k+8oo/F/r5hD9XYwXvkiea3IVkEvrLutf7+FFm69na
hMjE+2/ssY64GXIIIcjTyWIeie8fNxN7PnqNIxj0Bsfy7wnfGYS6zTpkPEe9
ATHd2Eq5BmEcURS8sWsQ+sPWh+9eHesKvXr9+jXdS4pHJ0hIivj0tEC1AVvR
kcG8Zh+3zAtNfxci3A6JrS7RjDbujHnfUCtR5auka3shpgy0hNM32NUWt8WB
itWRytTwB+ETcP8XKYZCLECoPnSVJgFViFFN5XcEtTG0fzpds3vYsRrKjK34
ydI4W/LlOWyBkc4TIy454i5ZCXUP6XfCIaHNvk3oQm+vwybVWXgNtzHc1Osj
V6lXsYE+Au9b5Ogrtg0TRSK3G4etcb97gG/3k+/6h/RHl6zRCJMU/TWjP3Fp
yaWj7IvM3feU+ohEQ1elFwtjKesqEB/Ujo3a3BIXFw1mux4fflGPd97/kirf
NkE4Bgt/m65KwdBKrlBj/cHrOGjZ9JiG144Udk65xWSxL/E4VQ9APS3nwV53
+WvZDjYciIQIxztD6RE5BrF9cwtIsFF4hb6PROx1R8cHZDZagsIdMMvQeIYp
niJ4JsSHiJGLKROZOdwI/UN75Mltlk5vAw7yik0jlQSD4SLQy3SwHS8RAXGA
TQCXeM2ADgxDxtEymrJmLCjFNa/w3nl4/vKns6tpka4qGgNF68ERo7WQC7XJ
6iCaONtua+eTZF89nCqrGEKjkBpQOA4jG8sECha1R740PByBIWum1o3jPUnU
PMhCygtgc4f71CIZs5++mdA3HVw1uJnldwqBkMTo/AAyhquEUbX7ODjJuI5S
JC32CiRVCizAxVWjpiEp4qS5BuSgnNwxl7uJ2BMXCIUSnAXWSI7Dsd5Cwidr
axSrZ9YAhilw3jMFpgWLJLupbilKkfbHJsFnLDnZgfkndfuW4c46W5eowu2Q
6cUfClly9QFrZdHjYDtwxZc4IGEE40KS2QPmjCcrfGTfJ1COJo8RyQXDMljY
5LeB4ti4yd8ViPlmyofTflPk61XYKpd4i/K3ZD8moD19pEbKtrW3qID39I1a
TILgE7toPzEWldWuT8GnTqdD/w+/GyCYdyfAG01u8fA7uyIt86ZGaeHNsmM2
sNzBNQaJ0O1FzTqf/ItadCwVxreNCRpitePLzThSKb/SHrPq/AB3hAmBtHLM
mFUWsZDZSfDw6XLhVYU+n+nFXwtuU9aBb1kGTZZeF0jhpHyC7J/P6bQWlUG2
l4533a7AxnZR9PnHCnZo+6C8zRF3E/8gI8VRvpSGUkVGuF5ohkbw957rGX8g
y3L9J3Yu08/KhmlQIFuIR/1T8DTkZBff7wgN7zhwMN3r+qkSWiHNldbWQPkZ
382+UJej4HHHOA8UIax5hg0bHq5LYzw3bYt02NTz93Nio2Tg1IkvDlUiVppb
wGrQM0cINf5DIlrhVBQJjCljBTKfU+kQ0AVIjIwCZUg0L33ScCkNJSAeJuH4
PENSxOGmd++Bz45hRbiTTuDenUDLgaW1NXZImHBYu5rwJw3OFSbtyzSKB9Yo
2TErmWQPfBzD/TK9w3dbmNLl6upNG1HnCd+RlydsmfCDF5HXLtJJEQt0z0Tn
sZ9hRukXUQiwQy2J3W/OAK8EdWIJdBqk9RiOtp6zyLolAhDjMS9fAiTgyOtX
EZknrF2D+TVx5uBN+H04+F+DgwEcxsHR0f5gr38E/+0dDfYO8P+OYe33jo5O
esfHe3Dr7p/saahMKbFye4OAz/V7Nl3RBWHvLbH4hbWNRrOMOfuBd2HJlQN/
OvdMZK45PVSwDztXYuAOVIB2sTreKvIiukZmJlUMtgBq7gZv6hsmsZO4l/FH
Z6dIVzS/SbsYGQ1bSKopA6xouaNwpz6yxt7D13iWSIIxJ4f2DESJcMcwC353
J5KggjoIM0RbcNDaxMuL9Crh82QQd6J6yQLDThBph/DSbJZ2PCTy46scU0SY
jy5jaTtvct+sP1igfRmaQGR1FbGFhO2jTKNkXxdZCrkEKk+OrMZIjRrQ4nng
Bi5Kg3k8ldCI8i5FO3C7zhlA5yO1gCEemN2CwgcoBmyWcOT3DUrYp2wf0a+R
5BwLV/gWphpsGRp7LcjVhqguvmfG9kZzhJqdMcO3nOAYkQHgHGmsCAU+AnPw
hQ2rkTACSnAfY3ZaKqxIg4YQv8Y2wpJ9pooTFsiRWGncGKPxjgGg7IxNmPh4
x8JQcPT7gi0jHxTcquNZjrkfx4zzZqWIbrcJTgV6rjhLmVo68IBgtqkCqAI/
WyMY7MwyR3hIS+lJDUPGPDCGs9rF87pqtccE06JXSmYtRI5oIRr02hGThVnC
mXxB9x9ItjnP3wxjllelDkBpiJIciSNhvIuCyP/E/xmS5EFDRvsgNgOTQzP8
A7rQ4ACUJgEDfAiBhKaKchv/rz/Hnd96nZO/tuSPzl//Wb9q/8v/GHcNBNys
JypwuMRkN+BDT8YgfxtcLA7cHGPN/TUOP7x/0w1bWsYMuT7uihh5NA9VIp1I
ECSLIJrXQY5Mt422Jb/bbZAjUrKQzkpHqeHrg9Y9N7cfjO3aKCnMQxruGhzW
arVAPkFeSjKC0HXqsvXC2GFmMNgjNJjF5Z2uiQvvZ4WbYXSiu8FcihWclDFZ
eenN7ip+XOTxbMTC+DjiLhXQppF3FC+GncDtbygOGyGj4HFXYFX+OFDYcpBX
SLxD0AtMeqKRpGYRe4/NSGR/ID4ySxmuhx8lgRBGMOKnuIpHyKxLbWQZfxzd
5it6FQ77fTwle5OVn8m2mEJrYmGzgH9ugHMSjdhthQ/QzS7xTrUYMvxKQjJd
pTrkAL6mKK5ucNI1IaGbC8GMkn9uWA5z3dKJJGcIEFJ9m0vj7eb32JdB73WD
fg/O3yP6E9mdhrsjZ1mRUjYErokKOC1dx1/7oeHOG4Pp6vPuzviLPDRS4O+/
D3vmBWerhx5luywdxo/7ggB2uKcie09I9rzN8YTjp8/j5hA7MV/Kq2KAtMMY
UUMaGNc2A3VJ1K6E3E2yuFuOpHbFOzU+NdmK1ln672u4sjXosqydTE7DVBBa
WIZhqNpfLTp/bJxHBRiWdVXdMrHrgeFsiqMJE7I+UevP2aN+d9vb9ZeO/Ss+
Mtj2VBMk+nEOk/RGNYPWeNA7muN/LGj0PprP7chgf/HCEEZMMuVyhWIID4lC
qnKS1Rm9JWkjzYrZs/+f3Da8+MdiK956UbCoLoN0rwUdhvKp4UY/5hHDhPy9
hTNOKmtlAnvHaCW03IsNwpKAsrQY7RYdAkfSjcIm2tdua5R/utmNR5nSiNw6
6BICxmHmUmeiz5MrHe3ROqNQ8NlYmm5NcpD1YjHUj20sb9f9UB/Ycg17jGEC
0oruOQFP+odNe18fDu24OTzbtpwHJVx7+CUCikJHzVTy2Np4JBqSiJl82SjD
gP0sctFvCfHtD8Ubic8oZ5gYNJtWTdcJLQfBZIRPkQqcSeUb04THCG3sA3vZ
heF2NSrNwEjI2S4taBR2zKALkwKAcBmULeKzZbvOnT50hR/WBRKQIqfkOIor
1VTYJofCrY54B7h7irI43GR3aIlr3HSvrcjy7LsRZqzLH4Bvu5+e55z9PiWU
EQ9InIkqXleTW47Up5HhBPBk4nWcKVF4GV9G7PxUwCfZ7U3Hirj2UNPUjIOc
DvoD47bmlUTWsnV8BPOEDsgMQB5pbpC1E3XMH51jw6jM2eQhqtD5lEiLYkPL
KCUoIu1LXm+8QWDCBojLeb6M0XIVV5joyUT8aZ6WlL+BMWDMCtooSLE1g9ky
lDeOUkXuXT4JnEU4v2E9A5cixgzXguEbS54ZNttKshkgDf2ZsrsI+YxNdhd+
mvK7IBnNkvs0xr/o3FIul7FoJCajyZrJlyxkpWFqoGUS7KoSXY1QfnKL9NGz
7trgjUbdRTALH3DZdVr8IU1bNV+hOKP9cuKDfb1dRVPaUOeIDF2iCeQaYl4w
Y/3AjouGpdoNpZYSA7B60cylr13S5af202Ynh5qtYChqtaImmgTAiFBL6lQi
fxtnrCWnFurJcXiT58x3reWuEdIMF9eNCb9B6yfsgqdbcnvebpkWiTmqUU78
24bEcwZ/hv29iFNEUT4KzO4XU1qU//iP/wj+MoFvOWnFXybwSdhcK5l9emzT
N0t49TFskTngE+I2kVfxT3NQmct/Mb/h6ccf+S/MT1iCgl4AO6KnV3EKfbT+
5Z/+Un4HPcL/tvDcfYIlbMMHcRbRg48cLge9fcLJT5IELvAWpv79pDk7Zp/Y
BDb7BC8mMGbpRbeIPoj/gP6+WceoyCQJ/I0JBfIHfh5ORT57pL/LfIFgNPp7
ET/M1wueZpwW/GxScMg5fMLFs96Td7rmQrEU9GQZGXpVLmV71RmDFkWbsYbN
3g4P1Tx5qgONw1aMEdULRPtN79qozKDZdkZXTS0ZC64lMmgKdEgW7HAe2VIE
rI8KYxmRHyYuGG9h8HwY1eQ1StXbPEp1xkkgUJsrw2TICDB+WZMtI7WTmM9d
mrZUocS5n1oIuMKaq/pItPzo2MQxeIE1MA1EX0oyms0wPHzWR19K5iljU6U8
CejP3TCzBsFbo/VuZllhIVkFIebDaAmfSTrfZjMvpZ8YBuysnYmrWH/0fcbP
WAT84MZPTfaTEe0PKAfkzsRMZqz3k6caUxs0mVwa7S10FzlS3J//2lWJz9j3
6VoWM99fxbA36g47f33qRYd7n/8HZZZCaEtTOkv2yqBTqjmjZqhrIUMa4RxG
aUZoNZ4fGjZZ1wFJ2Whw+GKV5yNM6D9ykjjQirhRgG7eW4LOOBoZNiKPktjt
9YsWS9tdeRtTVlEcn4FJkEkaZmUzfZN5wslaTZ9Vu9AhfXaGYi4ed0w4FsRq
gzpQ8WCEYFQ/67rqmBuQqvnNJVKTEg5l5FWmVInFPRn/QSTG5UsTUclGwu68
vhxx27c2kqSyIW4TwMwVuTeEbezTNqpdo3hmyiNhCCqPoUGcr7Umio6RwhtA
gSiA0xEyT/lHR2aqhpQuG0K2NdSyYIc4lHYcKwxNmVsgstYfdKJNPft2MT9X
lLHC6t3ij6eHfUzi2ajWBrVeU6KdfCd1ZZzDp4FfPGJ7+sZIHytd0rD3pMNi
iZ/ixXhmpHOECJrQFgcqPnw+rCMQg47GNBFaABOtkKDexNSYR7ILM5O45dRN
iC/AzECWT6NNX2N2jwLztxooNc0FlHJ6cqY1i4V5O/d48Bw/jprZSRR8Fd0T
QW/uaDd8T7C7E8rosCX7mAgslp0oDhK94TYXccDzZL9D6Vh+n7kDOVUFNyhx
f+QE2RqOTq4mXFzkmBToogi9LOe8z4wvKTX/wSmRdkP6YI46SNDgLnTFVTxR
YCoDk77e+kbdgJh9sjvXK8aPSbMsw1dkA3p1GAVeRluRINh/bqQHV8gIgl+8
cKn/L8SH2FXu4M52NCrH7qj4P3MpGPRzTdf65OylaYozA9DjwlRYGKtdNj6b
qic7UONMvMRYcL1HlIuRsYX2Qa0SYUvCLDdYMI0SRwA3/oIVHRnWc/kZNkeY
5Zr/gXSgtXjgbDs0bfn6P9VDZVNMbOZ+wPKkKo8QMGWTenLG0wAxbOjTX3zX
HAkZgJlfqx7ijS+w/5h0XZHR6dTV59v1XPtucgN67tuywflvciRMb9MFQtXY
+0RmiHI9kbJCbiNDp5X7XpeFQn50bDI+k8zPb3SyGLj5A8PK3P3Tr0eI0ti4
uJj+zaXF0boye4nlM1FOGEV1Wg9QQ/SxG3mIsAMOGzCZMjbvsVCqZBBIqiAI
jreVDZket23lDi4imZZQBQ2aloMdFLOZMPitK3OqkZ+BHwJG4tZ2YAgH10nv
DdTrUaylZhssIM4GGRaWfRv985hZ7zcY4YxDuNTBhi2XFBDMszFhQeLrrUYX
2lcfWrqvbE5hsVD/DUcrnZs8hQHDl7e24R7txgHGWeAzzr/x6LXiwAim3g+E
a2xbBkXHUg4lRUrZjNWBTJB/fNf1fn/Hq9uYcZnr3SCtB1xeB8QW46A1CfY5
44/4iELKEULihMTK7txqLgOJ5rfQOT+3rLjqnAh+NzAoEH8Rz3jH3bekQuwu
tfaFtoPGNLrd8AUC4Gh5yhBUXUwDidaLR02UbNPENsVEA7dlNJAk0YhLIgo4
VYsYQ2mrx5Xgrs6sxyUcv3zOWpByzqJ3Tc9w0J7b1LtGrIHf1MumZwimVRuV
CwuwbWAn7i/oBiKaI5CLgKbHG+/SkVZc5QNwmYLgWe/Ydmr32oILwhVv8UuF
NTU8wu4fJFkOfFmhqGFGFErJupWmtLRnyzYh8Sky+NQws7CBlzEWdBovxnLg
tLg8P6SNuM33SDSm0ZizRoio+mIbbEJXIRAbJNDwDEXQGbu/Swg+2oaHhUA/
tNOrJYjqC+BXEnIm+AQEyxpmY7/F4ZAslJt8UWhAbA8J0BgyDoCYiPzaStvc
J6FWmgmPJKFWap7UYD5+YRuwxglD8R2Zj7wGRBW0ru4ysKzEmyEKIQgG2C/G
GC0WbR8OLgnJsI5XvpBcHXMKc3T2zWsfCRW7EKsdtBxJ1jn8Ns/EHcCMAVZ5
agQNcfpGLocJw4Zc1OK7cLbFYWIFkVXs1ER6SbnZkKuZpdTSe1go0hm9J3zY
CDvg8XcJh+oKOqGLWLXr7S7kjVXh7RfD1XZuRtYsu3plAx0TPm2TZdTsJEkz
32jAhIWSiZKZR0KQrnd2ha2VjK/u03ABcuKSsljhMfIsWbQmGS84OfL4/mhq
x1xw9RYu5uFSzoA1uEXhy2+tb9B+bx5v8H2/5ChSLtXnGu9asvAVHpQZRtrB
/It2k1HvVBgPS5AUlo8KtrjEN3fBNdiBGPzSrgKTT/OTSDuwpgjyfCCQPp1O
vUck/xQ5ys0pFN7rDCOjili4TE1mQ0kuNCO4gQ9l8I83Y+oMATj2PwFB8g47
2wv9nvrrwY5RnamYHWfecnhrIU/wMrADcXZKjXCzelm50KxncVKWXKiVSeLJ
xtY8HQuEinPpFDiYFotwkaaLITCJCUD016rtC0ub8lHYxL5IRjI917Lo05+C
72JwhW6FWtgFw2UxsM3J/pyT6K6666D1WpRoUrLoU8JKEuLoW/iOmoH9ecdt
kahOScl1IJENOJorRAD3wLgYGOfyi/E4cPYOJkpktvAqK6MCByoSgWZ7MMYG
eURNkvANJr8eC/TD4gDmFpLGRkiDPnNqTg1t1JSVY0iaptzGQrmU+IaSdSGv
x0VwMnXBvUMwjS/tv9jLBNePWoUE+io+ScMAFBLFAgcK1tgjsq1UasoMQwW+
UYkaGpDUEx8/7ThelJ0htfKZMmBKgGhqaxZyPDlljnvIsSuGhmhG+NsExkit
093qLCAubnNHDHMx+g+u3DNCpfA6g9AbK87m5Wahi7rxv3SSl/HU333NS7hi
ctidceqRxxbheLcurO84pi83vQ+qHlG4DA/gGUdDc9aobpsgPdc2p6aB/UkE
wdAT3rmbZ2qAeJrwtwb0Ft9x/obqlmEjDnSO6F+ZnzmZvHUkh7ixzhVGbbFt
gTgDByBLqhz+1XOxwjRPXfE8Ld0u+L6ovMgDMu0ibzepZJCti7Xe689FM6WZ
K9yX7OclRjUXp2aM+V3q2hHnLKYW7Lqn5VbQHqcgtKRROQYae285ajlsauk0
7dsp9PYeuncLw88aT4wtUELvjzClVskJMWzVKfJnMNQMI/AV8mhWbmy1XsHj
0ECJzZHN9Iv2W5PjstniRcRm1EJeaCIlvlnij6kkzeVqlSSC4neI/+FpmyC5
Hzgv3c+Yl+7pG8mz9M03HAf3qIU/nZ81KxVPjTOe2typEkLzkDtVTvAYauK1
+TydrhdwLuYLDO6ru1E3JBG7m1p3yQeQBK2xjb+Azb6VVNoaaCcheDbcrtGD
Eoyp5vEIZzPSNILij6rPkHBfNc/X58+Bcio2kEr025dmB81+Afa+DfMeOdLW
M3h2kq/pG26YVp2j7uownHpMkCNDqjbD6HUOWvLj8XjpbTbyriQGbwZNmrg5
SYr4db2T51oA9NDnWHTwjezlTnBgZGDslHCIbRooEWl5AMowRBUCtma35axp
VD64PNUcbSb4OAtqAEFbdICPgpqwmdDaZJr5gQtkiJyWlRTjf08AHEZaIyMZ
Mk+X0PsVuifzdUnAXLFJsW3GFOFjFUqWFBMmeUUGdDq3UnSiQFZBaR+RMZdM
eyLluP3x89+iuRLr84lqbr7FqWUvcJIt7u5ogP0dHUahH72txTUxvhrVBiy4
vKksnIr7h+V3NcmS9mzIV+vN6VRxYDZo3I0LwdQF0FEti7I2xwBRjfM4dk47
V1+PtYPY5zldJUg2tFP3bF8nvdp7kk9FYJeMH09LSWTRkn+Oe5Rkn3NCtXnr
OmaL8Br3SZwMTesSdwtdSkCdTDQS8rtMijvMtJHnFRfnmDwGG7kaO5TyAjkF
ZmKgzQMyQe/vscnsDGNSgkHpTUckgZlolCuFqih1ox4Z1IL4PWdVl/FjwOk1
bOi+Zev8IIfgG+pSKuzQqlESP8VkaDYeDORGqPe6lBxDk+Qx50ujJK8+bzCr
UekSIZnLVRkFBE6Ve7ptctHbHD9PQ84F8v3Oyc5ntNtcGX1GLBRcdqHGvNj3
4nM/zqePEiHdzn8T03P4nSg9GtL2PN9z1EzRIRmiT1K8ngcWo/SCUwc/74FJ
IevnF0eH5FxsqFLzTCLV6e50Pf9uyg+dcya1TDQmtDKJv/mGJkti42XM6mW9
k1M1zxPbIIsT5W+zXJjbZW9m/W3KRw7f8UjiLV73hj116iRY1aR52NSz+i9V
2DQDiTNbPM+q4PSz9MePjbAvFG3e8/1j5H/P48wVJMyemixsmLCOeZpFRA56
lBVRenG3EpNrS4kHhEmS7YbKJlSU5VygEJtgLAbAvGQ5JLxmFv5Ks1Vx0mlH
QkEsFbo2WNYRQWdMchFDnsPv8Ycy/P3vw8E+IqGW6FWUL/+fsPexp0F+XMfC
+f6YJkdbJblmQ3qbUq5zrT6TKFzumpZArjwbPS6KuYtkAL/7Xdg6Dv85bOEQ
O+Feuz0WfoBfgJCyB6IoGv/keRg+P48wFXgEnoeV2rS5MaGORezdEAyHjmAo
RSKbpEJQLrYHQ/IZ6h/x+pA3BH42n+1B9S92knmOGvLStoX9B4Qbu8Pk4PAQ
7yQQJiqg9RQuoS8lkBqi6YJfMCe4JvzcmVxXWNKCfv7s4qv0FaqMyCnHn74R
HrWZ4104m1QG1CSmZBkFfoQjsaUohvxrIDZAy0edSCdock75UuYbyUVDzTh6
9Po04BflDQp8OpXGrETiMP2mghWBW7BCqyf4tTQ8fBbl3TFZnjB6pKTyt0g8
mElL8+x7uam8eB8Gx3JGLildn3Lx5sAmsERF2ya6NFnitThGPcmlDIBbHTrN
X+aU3kuaQpw/P+lEtCHnG9ZydtVfE4egSFr0m49aMzYD+o3kREMdbmpPzDfj
RuS1vYpzjfhoN80BS9Go1xoIBbxnUVAs7RnsvH2illoBT2YdcdLUTySmF1sa
aRiMTTiPj9UX5y58XeWNkBYbDy9Wn1NrBSuff0PNqqcMbw/0NfV0SA6uMdfH
csO46hHPLvIw8PfQQx76gW5ob6BITEdPZ2NKQIgZim0mRChVwuFpDMXAPq2c
Adqa5B+niIqNg43R+ufNnBD2wMJlD9L4lHwwugrflkEtcg21Fapfr8tLSWLv
yTojdzaeBaJmSY7eQM9+/eYYqyWkteGh5iCkTntDBjxKDT4i3aRpV1XA9PJ3
AMHx5nKiPNCrk4Jsaeg/Ium0IZTAaQyzB5eYccgXWaVOCKkzXv6RlPJSobre
dliaZea2CMSDZrKx/mGXnTEFYrqfgJgX572jtlkXquXtb/uZMNm0Vk/L1Rxx
mJjsNn60oc8TA0lzJovEaZ0tKbK7A4QjgV/r29b1QwSPRWhkMprTyEvPGAj+
YJJYcCGDEPx6MF4dk5wKueAxpEysAdwNbZGhpJgGzVQlfunQE6iRYsmLz/BG
TvBXqxFCJ9eraZBK4iq+wWeRKWTBlpBNZDNlI/KRzQ0FQiJm985x+GTnEjVN
INze0rryCMOQApt5HHqgNFqgei7ij+1ah9sh0k3lTXIjh36ytU1abhWHsdEz
2tTK3z1/C/Pbko+sYclUz/NEi7R0wn1RliQY7pIcICVX56FjroTlODSkrgBp
IKF15zRPULije4eTypwxasWjMnNMSVxtzOWZlqa+hhPcwS2yPEpKwqJeCaW2
sPzwxiZ/zejdTd+yB1tadwtPSjKHxnAsH2ZfpwgTbILN4+2EgN3RHNOztjx4
rmUwUhCFSdAdPy2li6FmUX1HYygriVFHnpI/sPJPbgzD9Zg4DCg5ljVgEyAb
KoQVGhaFNWISY1RRkEcdQdyQa5yXjH9Wa+WIkqePJgk2PcKAbkIbX5neCOGm
Z1NrFdnyuSZQiaqTDDmMhakmqNfSKdPf0Hy8TLk4JHzq0Ceut+KUYyEJR4lH
yrXArN3KJm7Jomt1NUvCA1sLKfAqwmyUQuqG77hYT2myavtiOedjrrMtrfdE
YinWE+SLmSU9P/VjaAQbaaXtVw+lyDAJgZf08rY20U2SrdHMY0qFQoeSfpZn
p0cW98gybk6G5VTUSpFNzNTnKkceBHD0ZmYIV9MYJAJsSP9UB36T5kQeSeTq
D/wxClLNcDdJJc5XI6YQl0syCN5A0+NpOgu/t/nAXFQ2gsBnN+6vTRnaglZa
KUaLvc8YND4FCcF5U5USKapjy4EypJvKI/E93A2vEIcZK8LpiyYzm/wx4JSE
NflFWTCLMdq8YdcNUg1G3qinrcRyIC0DmVW9oWAgpCQTYKwO41Mxlm7uzpx1
Os0b7kyBsFskj/qyR00wwdTPC3Tl1kK0G3U+H74uDNp7ErcbRkdRxMD7RXai
y4ETHWDvcpEvQWFihvs1vXlcmbZiTUU5pF5UmS4oc+oz7ck215VQaC7N0IWF
rCWrtjYQ/hNMDP7EmaGRB05jw7OwAP9kFmCzXaD4LwRQjTd1aWyRXmxo0V0V
Xm33spJzaW4rV5Yow5tcg08JD4UzivweGL9NbwXuzSBsuGWrFXFqKimUxFvd
NWyhgLuInM7BhtMZt9eWyyPH8yljPRaxAYKRYE1hRgHnf3UbJ8+PG8On5LMg
xCOB6eQ4Yu4Njj8Mk7hYpMjhKMrRF6v84JG0NNvN+o5UPRRB3ysKKAGUeAu+
oVvw6RvnEiTTtjVf98ns5i7rZl00z6jdYNEWuXW7RdvusNgbA/bYcnFaThzU
xH+WSSGTM7926FeY5TDAofdrNfZU3PGYVy0mNXgmJvWcvLznh4iA2ohJvTBD
fEFDfPqmPqwguH7INfC3SUzWY+wPEDg2hjIRSiQOtCrZQ0oyCsfpSmy1pte4
x7I84UszdyxMecViwwtUVdBLICIRWz9oDG4FQh9yw0qCJQvMGzwUySiZBTyW
FsFRYuBxs5hcHfBxwhmj2nxdsfeNVQbPesBwhWsO9eNRS+pzqUciAoTxbpPF
tRv+gVKSW32oDBrK9nEaGIoVb1BFvOqbFAYYB5zpnIbrzjmxKdm87SLL7xkO
/FHLDBgFZgcuPkndxL90nVQCtXwBgbjHjGeJnEmn5k3XoOREmuGjAhzg1MAc
nEDdnArKAG1JDR5p2mHNXU7NEy7P8WufGk2VtjJcJDcMzkBjA37TgW+ItXJI
tRiXL+Y+TpfWprY06CtNViakwp4HLAiSEO65xbOI/PVmPo63Kr9MdpBu+Equ
WSs6BqG5BTx13pTYHsnVPG7LoClsjgF3OrglWnPEmqIGFMO5HCiZxDgiJJWm
Q2Vjvq4kqFPfM/AD1rF+ZTqjOlFiJNskMFsUguGIJE8KUsAsqgPsomQwnPLl
qDHlC3ORM89iByyjZsIDrcF/Ii1NMTN7+oYWGLypvPAm4ZHVnAQLCWvRFMUk
F2oRDXGJoBxJpb5I9X10pUKNvq+V9Ag/eTYlDqLi1OR+qnHLpszPgs3eMRgB
BWtzptFPoaut54V4gx4ZhS6RMIIpxz85ZtODrbNPHQ24UslGZuAVH6HULDIk
N2X6Jzng8o3bfVNDXKrEbcyd/rOz8aO0xTkFvJV3B8U3j6WbzJ4bqfzOPQyt
D84IeFyR4yqgirBcaVuWLaptghOER23oFYNNRRob5OvFhh75KFJUhKTzZ39A
pn3ShWJXcMiKTOnXp1XIDKnY3gZFovWIrl1jRPIthiES3hENKTd0YAObscjx
2rHn35ooqG6CKRgLJ1Kgd1W+6pCfVSpy0g1lnVdGZYx8/xjpuxtuLYRMOB6q
qObjijzfHabn5lqs9KR4IKkJhzoiGI67ps+Mx1nm59sD0s7yB5DNb8j++cVG
xf4mojhPcrrZk+8SdXwk7PFwDUZacxlufFzvtJJnmDo05SeVYiGoUgfBDbIn
7aaV9zEwBOOWAIv6u2bNHcVtS4YijYGjcBSxFaSaFd4JD94czqmeJXmDwz+C
UN8a7+R3XDuJ0vjwnwgGg9FpwimqIEEHozYFWC02GTnYGZCEzSg0nLD+Xrzx
iGG7tcBG4y0VTKJvi/j6bGrIboLQNldL/2x69SIyjM+qG7bOiM9DE4lohz6U
ilW3GzTQIn6GpcdNZztSrL1nu23S3FwftYZZKVkagtXUZ777zXjdlGXFhHS7
SQRTTmlQiWth2RAme0xVwceBrE+k57mlWemgYI9ptmbHs1PBmpONxlNB6LLu
zKGrTu5+z6tgJ0Pe3QZfTeD4aqQUtFMJWeU30NCTWI20aLvm4fKLakWIQswe
i+gD2JIHR2/VB0lxzyQ1i+ReRP3MM1OLexlzBnAo7NdyUnLN+5yLkczieKdM
K2b7EeBjsTy6xnRh/EKlFzOtwHZhH+gg3IQSg6mLpIwfKT0HZdIz+BYiPrQJ
wlouEQxFxcRepMBz0kzym2ps6jB82jFxVD9E4Y7VWuCLnd8d7qNB6/c78Aup
IfDln/NVFGKZ4s/UUpM7nq5AogrkP2qkNkYTxyadVpIvhDPGMAZWqo2rmE1C
ziJOlwpBdmqaayKVgPD1tICMI7erxtP7rHhjheTGDonUlT1GrmkBPpfzUM6w
8XalsI6Tc24fID1+Twp6uxFFL3n6ZsxkSBUwIOf+sMXNNerz6UkxYJ8jCgV6
3Khzjsm4fCUy5Arnji7oG29dwLFWZeLtYLQTzASkrqzUVItq8uC81+4orbIK
rKZMTCWNum6BKZdFbN20e8le4qBf4JTsCvEMv2CH3uISVyqwKrbQyGe2fDRp
8NtNvc/7I9H4yZl52R8pXf+uGTz/pdbofIzcIBXeMPZoS+EvCnXThMwmVEg9
lKY8mKNkItOg5cW5Kp+8xXJnlR5tKffKBzSt3Ln8HjHv5hp3YgVopTVYwDFz
fMU0Ea0+yieYYRO4LWGINRXj5t780/cNlpVah54gnxcrYD9w6QwR0USp4SM2
dtIw5GKcSNVMaUMEBNQzNHaJXJJwWa1L8nY7GpdDuDTSHVtj7R2Zdwwh88EE
dUwmRp+3K2Ru6D7iMw3kqOwGVxRHYIwIWgjBrRXq2c8kLjOcrguKmGMMDrE4
5icczqw9kHafr2AxOYFC3cHztJOv8NJgGX3nM5KzNv29hnbIF+Kl0VfgJC2w
6F3tHRwcIUTF+ucGH5kTiIGi8obXJjGqGd5c8BR88cda4/rXd9x464+1Ma1Q
7HiuAfPad84AnNzXPiSM0qON/yjRD1raY84YsPBvwYAFigFzaWG/d3IYOiAq
MzgB4XHJSW6CTsYfuR4jVdcCiYLGmang5qZriDxxV0spRCz4Bo7gW20KAgLf
cAVfR1ZEQcIm2EJwr6Yi1OFEQv8itJPVgDiNR7IBArQ0G8vhwcHeIaG6Mlyz
ZBVxsi5rWpWoZ9oeZlScsD6gYy2kZoO+3LAhL1wsjCvnDvy2DOy14gPCCYmf
8oVaDrcFWwrvo3GZVKcijxoqIvc6KozLCRnfasn/LLWZCCcsaHptQnxgK96+
O7+8vvjp/Or67Kd3V58/n7LNr/Y6x4PFdP1zyD5usq3t/IL0rjfJDQdfYKVv
FSlQ6ybFvhAdUc2HxrQvlZFt5XJ6mQmXxxkYGzYCvHmv8AqS4qMFGV4kgwSb
utkOKIxEUPsfA2F7RPGwkQlMFPlCjEht0MBNWnnqNZlZwBNnPVuQBwG0wKt8
SMGrZHRGJmiKZOPI0tL2LP41XBQxslNk6QXnSTKPucmUWNzKzR3KURieLyOy
EZVhSxshNJJUTUSb+B665CZrjjKX5RtSDIzsbmnqephO0aHJaP+K4wf5+7HB
P5D9z+TBYabK1gFhkBLYIaPDJlpj6hqGduNkF+VjUDlBJtpJ9DXvXhunBrzc
Qsim9K+8gsN1KAoXy/9xFpsMk6LidfOdc5ls3B/VrX9LfkcjoRKwq1oXaaYx
hcG+OygmHF2nD9evO8dUAl5F9qLkLL+CVMBQDOoEpXjawM6vJSJtTylY5TK+
tEn8yRDAsRFulSkBZfFyUfeU1AJBXOM2KiX5ZgieU3iULUPUEKuxHUoDjMa4
Dp/l+z6i27DWXdvouh0x5dpHIrcJFDGTzgr+kt8xfubcUofZXqRcnLAmk5Ov
eXyjdYbGwWzMgcECr6zXMdWkJS2ivm0NSGDeeGoS5pq8N5Xrllsk2U3FuVtM
d9SyMzpqA5NTk3UIR4cFSdz6qPLqn9O/MsZ6TOgOTZMid8dHITP0E2XJULLA
GQQp+8AQh0e9wgL+lOP1qs/I7+TFHVPsMYeM0U7poXfvJfXILOOVFF7FoCQ+
bHXfyxgfGJtvxqeO4dfkh6nZbiVpI9ZSS7OZpsxgdJ3W/SBLvzNRfNLsUhQ2
FoCxi9nFx8dOPDEdKgqtkt13m3U2iELgnHYc7BW00kwJ9WWzVKBuTaYxlw4k
TlrkK0lzWesaKUH5gwapyS1SxBxtifxCZHI2vDjRq+TcLJMF1bJV62VSTmO0
VJkx8npHlBYc4xWx+opNnGjredvKuZxmCCHHLE1xAoE5ChUe3VZ5BR/sMeHb
z+fAm+NwuHmJXlzGeMdM5XJtaXgm3ZmSNWCLlGQ1RNOsprkhQELolm5mYKGp
0BzOkgoEMO4PeTt5Pn2DCds8XesIChyRQxckf2t1a989pF9z6l0iiHPKc3wl
eY5fq9r49A3hCTtlIrW32RikX6v1B1uwmZJTzXIonatUVRJ6CYWrZv97QPeH
f+cxoWLwXIGJqhZiMcGQ+VJSeouZ6scU91VHhucL7ZTsFKShsZ0afxi6jCOq
peArPDcOcBC4OIw7UoAw6E5VMw9CXIyRB3E4Ol0F3eLDbLTQMHWagyZSJAAr
3cmShycg8mJBkNIDgeLz9EQOXthmsqTe5g+h7VdARNi6ZGhmEjMJbZ6e3O36
LAXv0o8m5wVRuLHASA4sRKxhmXatcF4mxjdhc1+sl4yUckKh2qQKxJXNpcW+
JjGjzSk5l6+CKf4msIYuMiELtM5AkSkWM/MyLVu0GHBzEB5I12Nlia1JBNpw
DLmO/dkHRxmbfb7mGtfGPyKRV6KT/UKhLrDRhoZor/Fale3IQUKvEBtfBFzt
a2MBTf1DNCpmBpHHA+JTjHOgUPssqEyYPSwNike8MU70HElgjA6sF2dEozWs
yYz0W9/jImWtFYiTcy1GdPhpRnTJQET16qcd3SZr9cQcnQTiuOW/sEgCWXXS
VXjtBjHUK63cSUIbZTTyPsU0X4ed8A7k4P5YizKg4YrKMZAFK2BMjVjVcWKk
1/h8GKPQ8RsEOaJWOEYkjjV5c/ZGgg9L0iWfW31rrGkmhBjdmmKO23gopQJv
QCc2N4bM6DSUm2ZM0RilZirldaLSzdKx4BhrffA68cxpmjzwbmBQq9ZtpGAA
0k4Nul35kWxpzCYxbpH4/1vCscL6CqBV4ABAbq/Jy09wgCLpsM+ftGbJVlMr
3fHTh6trpMMin62n5FZAHDaLnU4gAwUj8+5RUMe3GFdPU+iipA1EsZDsf100
5LGb+5MIKzVT3gb09VPoBTSb6DnCQTD0wqa2uuUqbnqXnzLa2kQbuyUwmYdR
YLpoKrUMZF7afK2fy4Y6odKxVJ0C5jLiIuqRloGUi0jglqMqHzGQmO+jGmWT
3YabZGsuNMvbOTblh6TslyE40445D3WTsFbZEA7AK4nfdOQbuDjoKTPiEfEP
RffI7S42njqToWiZWcJgyHMyE5pldg4xufvtknCH/vRHLGBt6bdRGGvsfp0t
Eqmf+wwXkZpUeH3ilminqtrVblbJJUgguPHTDsoaaJUVOzPZZ8mD+XksmUX5
5rFCm5DxxhKRaWSWSqnBdYZWIxE0qZZKvR5epJmuNsGBkXdpbsiVLcvBOHUW
eW/ajFXXAn2UiTclRzcMrO2EXdOwxesydgOScOWslg8a3orQ8d4NDDItgzox
mI3CXH1nne90J5S8qSBaC37yDNzEy5h7kW17g4LpotpGZhIrJPFFppRKgJE5
RAcxC/E+jDlsacVOxm20JTfWcr2AyxEUDE5WGjhyP992r9waI0/fePNiL+IM
Ax6kglKdHwpcH4+/G82AnyWYQQ62BvRqHMdmKGlzuGtjeOSXQg8/+YKajVjE
5O5zFrk3AlOdTBbu922yTlMs2SJXjd1K2q5nzJTZswVOmC43cLo0QhvOp4ui
WYDc8fNJcPFx8riC7ujpelx65MV3qpXLxfoZsJMHw7SORNdBG33BAWl/9430
0TNqKQ/7eSUz+k8s9laHcEOHz7ZPrRmHaNNwuT9VbilkiNqyqqnLFdyDhmyB
kQEbge0NoS4kwRCaykNUo9HJxvOohSm2/CuuXDpi8qJcMRw1yy6GxQNoeIEp
DmIHKV6HnLltzNh3adQhXUp5GpgwRu83CkZqKFJ05k9bgh0CZ7C2rJPDgQ1c
PTPRVBFIwNMYTeQNyb2b0vtGdTB7UAez03lz42klVipSsIoL03Jjbk5wZoG7
+nLGMHzOy2oQbcTStaVxN8aWL12O9TDStUwX7zwbUs0CaKJwmqAWhcPb6I/8
VNBxGp5kAk3N9RCM7/uj/mHX2JF8LIsUmSYfq8SUmIFzoCpK7G9Fv3v6xhPt
yNRkhOmuKfjmxyGCSpXtxmOmF84WZco+4H29DOj6q+fFM1nbnRT/UhaOy0z4
bKNeJa9VF27a1lIhsxH8XVCTcK2EqcmCGLHtZwV1fBc4q6C1MX4pk6mjbosf
iRdDE6E9JKA6iFMY26nrn1ps3AqbQV3YPJV2VwzgHNkKA1qoXNXAwBPu9T1M
Rj4mwclXMxjyt6lcSNQlvDBPUmKZc041pPK3BKA5pkmjETqEaRVDIkinfkLl
CkyclaSsGTsdPHUgxavR3g7rgrcVUbzKIhLpP1KG5iGwudqmnOgRxteOqYw0
NAXX64gS5o/Q8EVQcsmlPkLeg++ZdOH8hbm19CNnBURGRZ1SpkAQ4/BVU+1V
FH7SvChR+YjupdKXjhi+brGaZsw8LrndSHtLZjH6ykydsMjKGJEjnUATVqDh
cSOWG59xKpiTrgkXB4lS5ntadBhtMkK0qCzoPcY/woaZ78zq6BcuRJ4BCag3
b+hORF2Gmsqx9YSxHkQRF7dA3uhPKKOA7d5OtbK0cMF+romcSY5L45TAEFlr
DzhUidAnnmmidMwPXHqPA4ocMY5MIX+gMJpz5BcansSGV7ZxkxpGl6jJDmRl
i1ZNGWyTnbUbvkmqYOwF+/Op5MP13gqFmCtPP7iQSsc0ZZ61iLHbiAQFUypG
vz+Hg0bxArWKyrbayxCe+T687TY949d64ScrJ4gOF++W3M9eSRVlzLK/+FTN
vFZzybEkHIRNsRMuTtD+yqWB681IaQx3idyAGr/hwHWneu4G17euSXSdrCyM
eqhd5wTRJWGM3ojdsB4jdnChSDd7RMAIB1c+62q054ZAt2IhS+BDBhkZOD1t
psAeyqUa3ve7+5S4hA6fbIgG2WKHDCEsU+TS9p3+gWtCl1BnN5iX/aZo65sF
EzZQU1cme0lbE+GrkQxphUavmtcIu4ej0LClZoTSdFq5ZCCiw7mNZkfR47Zh
GVDhxOxhGJh4sC0w8Sovqtp2EwyodV4PEuWoMBMM4RzEQOj0jmtDImzHHTAK
aOfh7/DEURCkU2ZJo2TD84BgayVdjDHV6mryIxGaW5QfvOo8si8DGYT623VM
ZtCRl1t4S1ROYJPuOzEa0tbYhP44xZIypX51P0q9sBb6f6mwVhTaEm+RW10M
YyJszS+EWthidjZhDp0B40LJyMzCGGLyQzkeWg6HFP3A9+9JXI7HKmxoTKCh
MX7iNC8oImw9a+0Q38g5zRgdkvSHiNg32yu/1xJieFskNdtAO9qMxFHyC38f
em1rSlriM1JaXFKCBpsJiFmIl+gGLnqD0MQpB9JRr9AAhsibHQxkVG7uZ68z
FgDsC+ONsxNogkF/5Cb7IHUpccNaDQ/afGWJ4+kbIdimGxPJqoH65/qcVopC
oEXWlHrTwqiiUPm17Odng7NzahjeNNUwHLoHyFuNS5B7GKfCN9tAMEipabSq
xZsRU4FuvAp6hDYdMhSaywI67ljuRbvAKDG3qp9soaEHDICWOkkTIvuGqvfj
bhsRfLWRanzfTVPxPmcJAq4CZB9CQYg1WZsCsiFmUYTJWhZUA1DGcxt4cRmc
y7pISw7n4EhoeVrLRZFYaGB+lCpzLkFAQCMfspSsrRmKiIv0N5U/rk0cr1O3
Zzzjmk/+EKVAJ6xKUwXPVmLN/JLhySMW+w5HB6q+27KBmW2TqmxKRSz40uBC
j1WF7fNJptDJxGXLWLnRhjCXazgzyczmAbc04p68M4MAgINniEw3cIt72wh/
9ui39Fhy/jxEfwfWoN8NyTs6Xt2NBb5j49AlET4nwhqHXhTJt2VQcwjyIb3c
SHGF0BETCQ1NDkkNvhG/GP9JRNR6i7b+xmJlgpqN/CGKBU/gJuuS8pFh1D42
4tdGIhzGjhc/u0NQBTrJcMoYjjsm4+2IogZH2yO6Io6gTOrAfex3b7C12AD6
cvgqFWtHU8Qo1+soEtd51CZFEKENI6A86dWrvrzhnZWSqfXYpKG6adlAMrB0
xMCY8hZTUCoimCFxHPN/MXci+k1OT1XOloK1BjIaCrDSmIWkqm090ZVoBkhb
auWhm3FEticukWNQBBGxv83+Ta88Ekk5zXgOZGtECixijNkwwRZxQlhyBu37
fm0RxDTuJtvCJa5ly0JnIJW28vzXEVOaMYexR13ueor+TlbhwKkUy03U1wY9
8pHAYGWUvdooCWvsoiz4TGmwKqWNwxWgbmu7zGk5MFLzfoBRMYO4o0907gd/
yXYwfEVLnrWeeIzhTm2UO8Pwd5RDy//695E+zxeKPsaf7K+bRw2e7HbN73RS
/K/sGaDvP7eD4L6/MYH+lyeQkrPY31H8Rkb83BgjHVnkjYfG0tsYS+/vHgs0
jBtlnSl6L/hZkHYILKJxMyabhemYhP24lOrqSCI26XY4NhkwNJM4GmqBvKgY
iDl18JvG1vKtpyNhH5sxVQWp1BGRotySzIPyqXB8RummNo8YbMP0G7h3QuTU
EKOrU4p85WuYHXstuv65Zmbo8pEako6R0kHLsXC5SeWIc9UsWNasRskjGmQ6
lmr90H1HW9Vk0OzPuI0DssxKAT2rwJp0aTLgSColQsvW9EUSMvpeTPjMBjDC
Ttgrk6VVE9RL1o6CEqZVEh9l7KskvXDqkguTpvrZHfWbWInDcR/y5OjGp/Bd
TdRWUXkoXAigfBPoy9CHW7wRVYnd0L2drEhrTMgJA7hFpZ4G2HBR0wjFElW7
LbFmJhvv9OK+laQOE8yNaeCVj4kmk2dItJdwIvSXC/PxEqKSyjxdaHJM7jkg
pDN2uKE7Cyw8Uaso90yPEV0AgRSk8RV5yT3FjlSInx00pJMJ6oA9enWDi4Qn
mZlsTsLVGDYRjR5y3RSgN6gtcxQ0HVCg7EPqgzSdF879FG9QPAvEr+n4e/Kv
cIRN4dcCU4xQSfmPPVtKIL+ZQqjjyPqV7dXkJvEycVgy4ocYEzKq0K4ZhE2I
vqB1WJ6nfaOfzC8d2kzNj8mpBEyiCDeQpZ5p6UxDtYxLNfi66lP49CKZowcW
hTdgxB+1PAtOMRC3G0N8J49i8nONrbW0ekbjN9TXlKlPDSsmhcbksQ7hiHhc
JGMGnvm1RUdlZdC5ageMXBSAST7CfkhNRBBPcgnNFqwz+RYV4GnZsjEQE4kG
DSbMGtaQ7Au4CSNtbWQR7TC0RZ7fBWtbWkrM7P5aOlZw8zLeu06yEDqoc7zE
jE8mwPbwMZQVko+Vn3ZCkS1NGWjd/EGBnz/IBHCg4O0d/QQNa+iMYrMxX6LG
JSbhilvHz6byIHYIFvONKA2SDs8MFwUUJuKtca+03hvlEpzEN9NbBmfHVAcb
N4lnCBrRvZRipdWjrk3W3ikJDrccX2pmDidaqyTxsJPCppIRzIhCRky+cs3h
QjaNFnLg4HLwTC6+98w7Z+G5rLLlDcJWkb0R4F35LMVRdXVXuEbVjllxfgcz
NHl2uMAxWTJ91MvyNGWsMAbowDGJGHoQqdIpcsRVhmfGUGcwskzBQc0pvunq
xqq55Jt/+kbfJQda3a+ZMu6hOfDngSoqrykAreXigyKT7LcOwXTBLZHLTTgX
k6IuJRu271JFyxaxZglJFnxA2IB+J4Odi66VedjVNWOXkTv8kUYiQSiRNyjT
cgN6WXALXmwQJ+Z2UvY+UJnrgsrTES6GnQp2MCIjWsM+5gszac+6GiBA047L
0OKahYQwJb8C/jUgrQazd6bhUQnWevPJZsp1qk0oej3MqUtZcWvJYxSFxNQv
lwFH+28icOdcO0Yw9mSKTwtoIn9oO4BdzZND8Cg3EMxiaqC3hYesqVOyg2ja
rI1kM98wc0HJl07JLxIT+CK5jbG+Z0G091JSBhMe+VycclfsTIRzVQqeVQIK
OxN9WZJJbbQJj046KG+w771MTOxADVbCNyeNzlWRDFIAGXMO+t99z4C1nOAD
rD9OtiLO0Ym8mKIOSHz9pPeuj7I903ikSeyWJqCZa4ogt6iwiSOoKY78JJUU
l5WGgcJ+UC53zhmL3b33A1s3o1oFeEYpB7zwWkqr25J417bb5jP+1rDZ38qO
SuO3jFUS9RqxETF+X7XTxvZ4zs9h4q6AbrVeuYMh9Sk2QmVIc6SnS7rqqDsm
AifzjbBCgvHGi9Fqrz9SLLEjSJh0RjhMzZbRnD5PIs6+lHpX6kJpODGxCRA3
BY7f2ihi2j7V+H9bWtVtYPLInrVPDdVFR0BVBIUeZckDjlQnQoTV8Qt6EMU1
ZF5yw9+MVcQ1M8Ji1qyM5OxyUuV7BaJ0BbxM8YKEtTcNsGS06poMFoTBMzRj
7UrKFwy7kPxEym1g3a4sYOEQQ5X4F439pHPMij1eKA/19sj+aVugWpCEfTjd
UFtJPWcdjnBKYkkiRAf2wpkPSrp0MpeJyIHC6T00pBUGEe0TiZJaAoZ0NZv9
T1Qkhug6bkxiHEYk5HoFWnsqw8jkmNPasYAJGqhJvUgPC+xPYI8wBKJVE6xb
uwM2dCgGf8Ti43mU8dJ8+iESoRJn/QyRVcVmxYrcknSUucIUtVXFnLE92azm
+8fTLXon25Rbrli1NQCGg50u96Lw8jD8VHfYcdSBrZtIcjfp3cBWVcA+dRSe
NKs3gT51ZMH+NeXFENAI9uG711igqNTNi+cGI0ukbTJG4gT4KdQfSh6VkTz8
8IpnI388zV/GcUjb5YLLJG9hbNNK37K8Zh5hY9i2mVrdhDo4FvomOquQgxBr
I1htypmXMZfIn/86pjMKPdGTsADvHuFu4OITsfnyX+P7+GpapFidSoLWbAIW
7vEEfrmy1Y5ouVQ5AnrFVENXV2+k1JuhIqc+kl8hkKm6R+tEAC1xlf26zu7c
tyRZNHn+RHwjoJTNgy+LpSDgFkZh06SI5+AVvR1upUaJjeJlNDo6dCbhZlON
dhqYosn01gXVGEjsNBxjivessqp1KBVZGrU+PhXE4bkYJurgpIZYUR97q8nO
z4WDnG4EsZ1aejXaKHzZpGd40V72anAY++GOYDhV7mMJDM3YeE1ocLikhvek
v3qBIOLXWoiM9Vha5BGlajXFwzWDbbRRlYy3eKQC14gE/rHlnVZk4fAdYY2b
13301Y96NdDc1ywjbmxeYXsMdQJG/NxTChMm7vzMg18WYSx35jdcBqvBtZaJ
ykNc7K1WFs5dfWCvI2Gvzz22rXRc1GB+q1fxcwftgMTSgj9MFok5C8oZNT0S
8jE6Q1wtPvJ+4NCDETv1HSYnz9hQvXKEBahHC/s+D895ROZESN64Snx6MJIf
6DYjA2qwb48dZui+tXrkpE7IEp2+iCOOHRa1IY4zJ7Bcyg5cOBIt8ZptOxs/
6g9AmwUoyOXGAw/IstwXa7aqkTBCTeEuDz3PYn4WvjCXVKLT7SwHZdVFjP5J
VeHgofcSpFSGvS6pEr3useKNr0AohUv0GkFbIM1670J7l04ZtsI0w+yeh1G7
jiUxmmRgZrUtsJe1lmPD4GJOModvU4KfWxTu0BPFpkoE002BYqicyEORU1mg
9zbcqu+m5C0T9AFUmMbOpBDyEPnoUsVvOlxmhaWTi6urD+ed/RNO+YmXJWe6
kMxjdKVg4GsVtNi91N9rtm4axwBb6bny5IxdqZh0Lr5LAjNPWP8lYVYajAtc
s0Vh7XwfcF0ex9NshG7jXXYLRrNSWo8eDh24iufQ5gKr5JB2unCd0WEt4Qqu
nINY+lkTr9XZsj2Vd8kjx8J0qcLQRhltNiFhIi5hEd6EmH9wHYJiiVH0XejN
m56CS+huRcQWq+TYDRUosfVtTiX3CKWhh2b8iuEOmm/rtBbxx03mhgx7pPPB
WbZeImPLF922XwPEm5kffeqWdKaMNW52fhuq99Urjo2OqLWxf3IGEduOsOfB
QT2rBqyIzSiLbQS1B/Qs7DeeBZJnsA1gvobAie4zyiL0WDaaudTURTy7sZ7g
35IqAwVWm0GECu4+kzaDrRh7nDHA1rdG2IGbQHV/r9fT9LOoiaixzQSoUqpk
x5a0kcPQN4nh1T3Ll6D7+IhPIoQVu6YliSVzBFVwgUnBFq2wbnyLtC/ZiRpg
0Jk1d6MpMUdVPJFJHzkaKdnoKesZghsw/lBwo6c2gm2sBULHT5/H0jvczV/q
3nQgEocsvLn/TICc9zMPka5+vl1DvnXZdxZL2CrFgsFS8yMd84g3Armd8dnR
P8PABwdtaR0Ow1n/AJ/E3JIwootsnmYYBQl/d8wHa38g/BBZXiJHZ2YzVOYl
vdzY/yisVRuzcv1mgWErkDUlspTR72uto5oNp/RxOL6jn1U2yQHfqivp7mF5
1u8rQyAJtG5mlJxfFA9Cic/V3r8R7DoW8wtXzxOS2mJH8QR6ypeXbZCSI+xz
8GHM8i2PdtBTtrTDCJ5yhy7j0sKrqQb0PCxBxuSSVxgQjB+8IZD7DDoAgnI0
khl5CVF1we+lyz7V7QL53pS6XYJIAQ/GuCC/JUWOedQobU1FYdoV2syqIRti
K00aZISClKp4UNZQVxcB2RbbGmnj2v2Auiczgirjk1rIVz3Uq2aKJzo2kYLu
KmyJr5Ke98nUkazCPsIFfQD0jmQZtGKFuhU4wzqhILZioJtMtyL2MlNwwvIH
KBY7t9+e3H7AVyiFysZ9V78Q9b47CP7r77sXA55Kh9hMYmy65HsSQfYLjAkB
3gpwJMOkDPhQuLL+5m1cKq9vZToZMez7ZOuTtMcvjtSO6wekC6yP81w7cLlT
ujewnhJVVerB/9C9ToIR2uOErtOsYe/tORSj45cvHh1HucEm7C+i4zb8gsdJ
5okXBPcpd+3mbSN4gNIKKXJprdLMAB0Lm/9ueov5qBmClyUbV5UdBj1IB6zk
0bzc+4rRSBVFkYnNhrDJMHZAsYr35uy0Novj2JbvkmzkWp2ONYxTE1cgB7eQ
ZFR6OFsuJ+Xr9gUaIDQGi+HZ5QroBgPI0dKGFxvRwneWvk3IGKtU8Fg/2e/1
bEgCP/iMtNXyxTLEyvDitL+KgezVGMi+4QjNdSeN5ZeUyXp6CU0R5vqEYT3W
IGw9hpLKruvV5UUaCXgB8MZ2ZHnU5w9EnyctET/q4I7bQ8rVSJISO24ZH1bL
kpAQq0H1mnI3hpQPjw8usnZNhMXxxiYV9kiQxyusPZjNONYno0gudeqBgHoX
YymjDC6BeKYsVvLIuGMAhpsXXNSBlLJZPl2zj1isBTeYR5ne4SpFWWKD9oFw
yYJBPLeMUwTC8G6PMJ3DAQ7RxpYSg5/aBF0G2BFrjVUGpZLCTLsTiVEgdeou
otCUVx2b72AmyJRe91A3YNCDDYi3lAwVQWCzHBprpQ7YqmTGGlP5afIT0+VM
2B9XYz3VPFVBU+yQ0TSFJ6kdi4GiyMEpZaRYwJSNjDjzzbghJ0o3fIHSCdxv
1soTWoKWfTOJFmlpjszS9NtDsxuSVHZ79lY2QDh5byW1Ty1BaIesWtrDQKif
kzs2k5varxiFGbpnfp9ke0nYoiQitjakHvSiU6v9wbFj4sc10KxZIae3tMYz
071DieKqJYcTjqLhDIIUYaUyrlwWqjELGnXAdZYpnFCLvW6/32zmM5a8Tu+k
0+/7/O3ELOI2a9eZlLXGjGNwgRRrChU+iHAxpCx46yDCkus3SYaVM8ShW1uw
dqgFK2o2QnmUMs36eTV4QJ4hT0BblWb5dNqhPF5joBK8gSzyubRJyCiLx+Zx
UTSEH9UEfAArejRklQr07reBT1tlVzk5tYlhHKgUPRX5sWbBY3waue9aDns7
7GI6BYbbCrJiBe+kHw11k7A7TqZ7e4NkOtk7PkpOJmhTwXz28VHcPxz0B4P9
+HAvmQ8wmPztJmEL5blkzVU5DSrN4agYyRI8T+ASLrHOnL0La2nCAi7I5udZ
FHvqxKriwNmRKZIP10kpFruJtOoboig2TZ2G4IpT1rZIZscLU0iSrS8bR9g3
0JZDwZ1IrpbMRVAEmldX88y7MxKrAmdSrpm3+3wMSziHB/Ty5eCQc6ctML8/
rl6ayd0EdH1TxEv2PAeqOBgjH+fP98as55Rd0k9P5+/fn12fdX7s9fqw9rIt
cP5Wi3gqDJRaYSjzKsl4I4/r0Z3AB8bL+OOIddjRBGbMthNsAQtvpSaASaEx
6U0HIVAx2egbkrhtgApIesO7iPoQjiSoaT7zpOBuHlPHbMe55GJJc2CRDyBq
1QjRXGYoW5w2t2tUvTLHNN7JYo5BwCyj14kn0tsbNkH2Qo4qZ8dHvII6fBAY
TH9pkuSsIkHOfN+Zylc1bzOu+Gotsola0OvCEo2MCr2zuBWBrKpcg5OzyHUz
mx3uz3rJIHTT3e0z9IrzsnKpzSqQF5L9/lH/YHZkL4+fz19ev31/hRso0No5
BagGvJodstfe93ZlArvs2Ah+d1tVq3K4u3sDAsx60oX2d/ObuFwXc6B5VOk6
v65udiUXVqe8RRa0W4GQuqtjODjq7Scng/2DA9AajnpHs/nxyUEynx0PJvP9
ZDrffciLOyDd3duyw9WZdtkgtrtcl+tJiqPyB/l7yndvY+NMCmpTC+qa16Pw
2OgwHAxw71nd7fcOQ1L4kiqd0mFKCyMvaQZ13sgaqgEtVXvHXi5n+m7Qj5qy
rtFvx005Z/mnw/5G7wxmlBz/imSmZw/qeZ6pY9u4AYAH0jST5dGhoUw5Duy2
KxI4SAK1NbQXcU22ec6FCroeVzSGGoeO+czvRZ7VR789jgJXl5OvOSuiS83y
wwE5D3wvKJULtfoYshuUyOl598DZ7ZTwOQ7B5uoAhK5ZT+Q7hF463j2bCpFq
/0gGOPOBQC7lbnJ5yV9J9Tv6NuKgS04Xj/WXTEmi3cw87+S+5pLJa7VVILwr
srK/+3xa2ediVRyMNERZvNw86uNhXRLY9MPIwqQUiaVHXf7lZOxUGM1JdsE/
chRhOqN/gzmwbpA7WnSWcDa67m3JXp9o+poNUoE5ISAMizmGs6QkVBd5bvhO
4OkoXCpSVC/xuZw1AOueDWxCLjYcz6WOzAKh4RcVAya50hszvLUzVmbP0wDr
ldPND3NRYlqbXIYoXevVpEvP45GyxYpexIuXtH1OaRgm2U0seG0J5SdoJsOO
aLgt7oHEJdBDYZcDMeTIUwrZmsAwDFSahqR9S+AIbKsD8IHzzMAa0XP4pPM1
YP12Ok7B+gWMTTcSizB3QU3JRVbT44IWzDCrOojyRAmqmxfsKLtbw/iXaQdk
ja7cG2kemVGzvnkC18Jxrx0FNALcuL2DE4VkS9HkK7NZ7s5QhXk2jZjd7Bjj
dhTYL22sVuQ8KpXCOILGfAv6Na56gWkJi/ReBXeW+EuMFY0El2hJSEZL+SFD
SoV1g9QdoHaWTtcLtIptSmPjQe9ojv8Ba3HQeoJsPwa1rddvc0kXplGesWLV
cObokaOfyNEsyEX0cdegaw3rQ5ic4o8vXs0Plq8H+yfV4cNs+dP9j9Xd9MUf
/23588/z6fxPr1+8+PVidVVV59+zybNpTbGl28Xex+XH344Pf3j9pz9eT3/L
dg9++LeH/X+9m+ylvezNUZre7A7+8Md/K6blRku6D9DMzU8/Xvzp3buLX/9w
WJwsq+PHSbE3u/jDH5aHb/708S67uXpcniyq+x8uqz9tNKMbh+38afnb5Xr3
/vDHk4cPb6vpMivLH85e/Pzmxav1mxf//tP037KfepdvvtsbFB++9yE+ltRw
sdHiCLQ9Uta4QqC06tKaUk94uZXyqRgQYgVEtAflwzswmvNXr2mswwOiB2gS
Iwvfwa7Wmab7U4WHbcdut8QTBXVB35PJgmsMawkTDoQh7t8N9QId8fC6lHdQ
pMPzP56/f3lxdQ6iNJVpJNTSMlnmBamXgczIucEW8SPyJuVY5LBu8Y3ITEOF
M7brRUEfyVqZqU3pKy2rVIIsGnFdHEbeIOlT29IrcDaU5mW1HPUwlxCfgC8O
McQjj0HOKAlbdM7d6iPGi3U6tJqazg1PH9YR6FAGXKyN/Kvcak/frMwHIBO/
lgnLGtbCDrxeajQ7Fk/2mTj3clBPrlezPXxbspfdMaIrGr1UVg1iY55xTvZl
SKF0aM8offFGAR2RLVIS2WIiJEX6RUG47tMqZRPif6KSAgsAWrIDbavGDD37
3719a3MbR5bl9/oVFfRGiIwpwHjwzZ2NpSW6m9225BDl7oiZnSUKQIGsFghw
UYBotlv/ffOec28+CqBk9+xOf3CLQKIqnzfv85zKk2bFI03KHXdpzRnRgBoh
FX10Rza25B8EzIIIp4HD42NVPcY6TauEzzphkRc4cFtwFszZAttf4elB2Tut
KeL8JAwYRmbJpYKc2EmAEQOa51R0thkuVIvAVcAJbVfzoZgNbtRJJZVwIo+8
n9Dmm4kJAg4WuPJsNxNqtRV91bJQuFVKO/0GqbRjdwszry/vTg+H25XIEoL/
oXkJ5BeZgtto05IiqCJFT/TUgBcp1+ji/HlR+b/fV3ZmfyrX9yCLMyOiFVDa
oalaEImCdHeJuBVOJMoro0RepJnmEsesdiJkMIPercbDIzELXKvrzpvufdWB
nuW+WS87mkW46lhyvuxOJiHSFVVqyY6Ikkc37EyL5MomDzMT3KQSwZuKH1kD
b+Ixn0qoUipt3X5wUkecRF/g96GvwwzwMUoFxapEmCjj8Lv5O86Gs8Z3x+MK
dXoOY0PV7MNMDEf/ij6NaaPHMfqpfk9TX40nrWKB8xoIDaLSLWJ/d5wx28oe
Rdopy82DsaE7X1rqNGqJ3fIpx9PlgqG0hZ8gSaI3Qg6UQfNJF1/IrqA5GWA4
GCJnX3beiub+lnFBP3wEk+BCEjVTRN6X+wYBtl4igQXJ3aKcVU3a5+Ew99/k
xqU5HODU516wdr0DTd+647E5OVu1gf/8djhgTwLMtC/9/9rs4cKTZ8Zo1iNv
R8XI1dtJnPFvfJK4tZYOeeD3x6VTvJ9/f68CvL/vEdnx2qjJ/k3WDW0mvcAJ
ZaaX3waj8/xlzSuiqVCPL6vmWbijxvJoWzkcKc3kFRLWr4bR22U152LRSAai
KnTneaLgKTzWuNKP8Z4LH6HbLJiMpsgNlECTct5BLoSAmH6pR4dELL06fqFP
apK7PgWOpoGYVT0PHWZme6A6JgCtCkjuwC1cU7/mBXyCsmNhgHm+nRawue4U
WS2ZFzi1x2QYkHKLFBr4XFcK0UdRbK16/aXwmFLmanh2SygcfHES32BZ3xyj
LDfLIxVRemvTNAJU9lY8ykTuYGDC9kW5LsJ1s8os60KzOJ3FLDCZVHufSNOs
D7/YKRKy/dQJe8CsnjVyyd2bSwn30Mv+ssjpGmHrrr5oKI33w1fsnQz2DuXx
bzSr9IqAdl4vMtpW3fx1csZY0t8+yxGSqU/WgvtL0Yz3/v0/9hSgG+qt4kDv
lBR8oxwh1rerWTYFJexkncfZtMlzwvm6YM361VH8iN7uR/iEjTvQwazp9ELu
ppx/fzz5jAN79LFV37N3aYmkRRjbj88k7gHozXB0vGAJmJPRM6BptkWPUtCu
AhQe6d1S2BZ3uN6sEPoUE7/xdmxNsl6MUlSgUIiJkN5Fm9pIlZhM8jgEA4vR
2tgvTjXZvIzjlfj8kB7WXhcZTStU91XxdxCgqKQXQUHUIu9tpQf8N8FI362O
CKYtq0UjbtB6ZYG3Wog7kLp2/s9J2d8iY9Wa3RKzVBIFXUUd0ONK89geN+rv
JObDmQz7hRTHzGKSpgJ9VTNjkqaHG3shM+F1KpslxrxGopJldbijsguh26d3
w4oxuSBdpGRQtOmdvxVsfHPFSbo1HhUlRA/7lqEoL0+JQ2Qf7tUPD9W0RvGX
vm3UAr13zTp9/X36zc5+iwu039sbHbRkWvbSJUVviexIObdAhEM2JdRzGtRa
spcpj2JX4SKAVeC6lziDPN8GxFER/FRFmmEspzapFtsdtWh7p1bVerMSHM/R
I8rdh7vuEHNH5bapPXn3orJH06+VrT3ycTKKNIgnU9T4n68CL7aYo5pgFxl/
yjhNN1ALitSH3BZT3tyLZcsLJtUgwUeGsNOVmtFBxIhxCsm49D6gTYPqumkl
rg2zdHUdQfStU7c7kIr6NU2AplTI75YVn7/0kNOlxXZZ0Eb4gbKpUYFgWbga
EoV7BK6Dp+XK9VZ4nrh5kO8lQQ8UBwmJp8/MJImn/xOQuPEHQLAVwiEnJmD2
s/5BUaNtbxWtWItWnracM3XD1FFuQk0EnwQcfGElRP4cl7gLdPT3m8WuZ3mK
I3N6B7RFCwLqliOcdfPlUKOiYqchSlAeyPv91m87v0fBnx6Fbz1QEBw08n6d
oyX8U/XqoXXGeKeFtNMQWoTvKScvgc+SeiiFAr4C3Hdr1FYnEy9hQW4zp388
Lc6TpJ7phT03TM7Fjh4E/3eyCK+8jZQGUaMkkdwS1QSEYQpI8seluy/dbjLu
8OpRHhCqe2MOR0YDpzVIv6ZkY5VcHNcLwEEbBJwVQNKdAD/KiqWhTJjXpm5h
9g6AAn6DTNTKb2KbkcbyelFTsA0H9llha4kVX2hGa3hOQO5xlsoGOOYiybvZ
kfHdJIdruWhd5j6j1TZxuV7L6urw2JJZqvsh1dBi+6JQFYzSH1iGflT1Wgqy
kDwGKZiMOnSz4y4QR9CRptlUQcLBocugylZcNXY+7pdr20fr+gE3lMycHIPf
k4hzgIfUCPSOouxZKRqIZ+08/+9+A/+PUcTKWo6X4CmO3cKVX4koVl1u+Y2l
yhq1eDbz2bhymtAn8I3QG7GUWkp1hbhlR1nhUkQ/C9D19/eEVs2mlWgedHoz
YW9LrLrbcDn/RPUVHaPcEP1CxEUajTN6r7AxfGEIhlQbFq1FyUt3CT3J7dHs
OpdIHru0ArT8fWQFRYhVZ/n+O81mfs2FOBBmDfuZguPdPxPk8rrJW63ZuPN0
LwRD9hlXlpQmgeu+XGvVT9OCdfFvg8lgpXXd/Gd4kVmKHmXdiqIwf6Y7WXIE
9ktZ/0coepc/XXNWQ3iZ6XcHucfEXVvpN8roipCRaojHmR13+egOZQd0Xwe8
LT8sqX69W+blPdAjEcGv7xbCf4v36FunJNPyi6GajrfLL9/8eH1zc/3urUgf
p9tEeKWBfHwR6YHhUTV6B6tQkUhVUGfBmkh/4dWs3LYO0tupYdkKMdobTfMr
qUqho4yEcygtrWUGplWIoSH50Bb7ocr8yZUVrwRW0nY1jZoIruKkwP/Rmc69
ydK/9TJr974Jc/f+6vvOX/qCkwTlRcOMNgvOZD+R892XCtI48auF3HBz9eHD
D1ed1+/eX7kHCcwFg5/3RG4ivIaKKqr34n/iB1HGV5SzJiIcSW/yD03V2BJK
ADvnoaGZYw4EMZXCqGHHbxYlbB3Q02A7LHcx2xhaWTRnVCA54YQ5VX2Cyx+0
kqRCwUsWgnvULQ3e3YzgTaXJAtq4KXHc4z1o43HCsl5UySaxHSTH89xvCZa6
2BHXNCCE1GtLsjFMw1KZEqM9xa1tm9jZAfqyEENvS0QVYPzlZ3I7to+K4e7p
gEdgs7BGnU89gYEWDABCrQbPzmiv3fJ/LfZGXmEyroBW5R2ynuMiTZ8wzEIh
89UnJKhg2A3lxhFOAjruE9NovZMrJSVDUxeUslO0BLeB3Edzo5Mcv9VdZef5
v/9qwPHt7hT6tmr6+T/MRuZL68YDjVkpWCzh9bYTZwBp5iWzeKRtbtGmKw6Z
zWpOex9JwvnP73/IyRkYNyT0AZcsG4l8bghGLIQOrsdrEte6uReyWaA006li
esjI5xMQcDWbzJdiPyKtRFHL5Kh19fyMIrmdXNY2/G6WXMgs1/9kkLUeoedQ
K/r60f5uyeT4ZLEAasCCrlH98LABBFbkGLBe8DuVD97qMoHN6IyC3gcfQnzo
6L7Wd9yKfeSjafoZQ8Aj9mmofdLueuIEt0seK54u/Yo+Bzlh+X7Ug0w0UVD5
zMu7Rku9DplW4YfKrprCsfNdYI/dflt/VFDmRYqnegrtObfL2ex2vXysJyPz
3NjSirAwBWMX4zFdHpa7Z06FzEIoZR6nUFj/rYRSX/JKE4fi1CAu/XesnZsZ
5DEmhiJOu66FAUx6BL1CjJSCAUfpQgJOUKEEHbl+/LcUsMNoTJP7uO2Bsre3
a1r3mCmnjmJWBkjrLeD1QPdOwbG/EragAxBU4D42udaSnNtqiEIlaAYKnfQr
jlpqT+Q+a8p50pIOnfUyPgQiEeeG0zIi2J6J2C/1H49N2Hzx21fcHbrQYep+
8zN9hoRPmvFXp43EvUN9TpQ+eAt063+y5+uy+aggGH4gsMMBN8vnB2kqe0CK
a1DWvyfRk72RPZ1tH8rVR2wWTcjjyZQUJdJd+ssKPl1/3vbUkb4nIxb0npkb
5HpPPbdUGPF8Yd4sBY/TOOHOPet2vEeeyhgAOANCZCNVU+lPw5C5Q7CXcjaN
dkXrV6YmA7OdpXx2+ybJnh+iDo3tBEfauaR/to5NQDgOv9V9ZVqRugf0yyzE
3e82FXBwQ3Y103qjSJakDcsYNaGoXl+wKLiE+yvsO+u1r72W0yRuCO8mZt5O
PEK7MNIxUmMy+SqS/Y7k8vYRN6+tuHLFt4Qx9SD7XLRxLkqUmqaF1phlZLIj
vUXH49Xkusm88zbxqPPUmbRGhk1f9fGwBcR5W87FVZ8qGCQdsa1vXASFUpJy
vxIA2O9BmRV9VvvCVZ0bv94xZHFtbveLpE/pZiraMo13bFsqmb8dbGnimeTT
Q1//M8O23740yl1DFI8j6Gt5DMl4aW+GE7Cl34lEUtoJNSPq2Ven1LYiPHyj
ROx43l0b2u97cmtHw1E3ijbzf+rhsYZyYtr2V6e0/ZZofeC74+7ZXnLoyVv7
v5udysnTedfgBoZjibR8OBpfhL/lNRYrIaWwUlq6ewRnWZ7hCbPSa62bncGy
UWU+UpYZWdshYr6wJrnnEW1vvZrOH2RevDjxwl1DVxJKTQvEhFYCbiDXB4fL
a8kkiri/1fhoVM3ZcsTwYt8yexVi07Ap6qm50OQZqgyMnDXmO3iOzhCJEXaT
/4a93GLpzlQfQsV/gK/SZ3sjUf1+wchHdKJhqMsX7kX2bPwLSOau+h3dlZ7Y
Ru9hG5nJ7v79WckHgzcO8YFtPwOquWUCZS8w9gFi5MQTOdusNvXoPAYlwwCt
3kXTrhOxmO23tajCFB3Gi3V7kqtqv9yRRrlcgY5K6ggXSkWmxtBOFo99ZBHE
Hb9lx7V26YWoNBip1dfuSQQxN1C7aBFtV1xbjR7vRQPjglvuN7gJNFQLd1bi
htR8osj/16fl23KV0uWD9I9s29v1So12738YeUxcBe/3TN4FXph51NPgzG0f
L/EzSCEdXXZTryJufCJKVim/jLGaXwXyQcRpX6SobCXyHxSex1Ak/Q7acSN/
18SDyKtX05mHdKBuRD5PhzlHjyHYLwpf0lE3Wq1Ke4qLkHERfCZJiaj4yHw2
t0gWudWoeJhnxCsDK0S2680RJy/1pVgdFFztNkOw31bkT4MusnVQu8k57AYJ
zM/duxGwCdWu0BXi+QviaChnFlqoojLYw9KX6DVoPIiB9zthHbZ+yPHzd2Lc
CdNKIqeTF+ZbHrYtcSpP2I8lKbo+OIDckb4rOYJcB+FO5jW9D/cL1TuU7Mu9
Ha5z3+gQGAq7RxIdXHkMfPe2XRThMVbrw/YLX/B+ypldmguFD/hueW7G5eSj
XW7uQ9kfksKyYBWYDBAgFFBwZSPZfQkRVdCDFbnPMF50pV4g2U60iC72VGtv
22iD989vyGTF/PeFBUVNs8LlU7YsnXxf+BXCVtupXRYvqyBkvN6thCzSuZYC
4x2KCMGxrLPB5xs7EhNSV8WM3jVJ6tH7T87RlrMw3uVsGUod3AN+R5+HkSSg
6nzuhdXOjWlBEWlAvEWvcUKgQhey48kGqo+qvuo10gDKXNjeTikVY9TuiCzK
xx59ob06uHwY81WTtQjJzVMbUSm9anLztOMYSnpQBL646xmuWfLRsNUfSKQY
vLFhEeNLdMO4WkCQh1B9ut88O5ebsCxhI25PCumBDTWxUaXAGukVlVlCrHbZ
LmUvD1KuZDozsn3PruPuMoA+kT1E8gDJXESpbI30i9GBhbYeqwUiFuZ2U6+L
0U6Jugyy4ihrwZP2otQqwjjtGrecBueQtRFfhFGFpadUefABKXXktygCUyQQ
EfDbICCqasKMrFeB4LKJ6PssTBA9LdgcIMPRfDwLurqxyccd/VjIUjAQlqxF
5FuppI/fZTXKyeOZ+qa+QB+E/avNUQhspmHPln9JN4ltjqUlnUXBDzh50aEs
CW42jBOExejAHEHpJXELbqFORqEee4AFT8PMaZDVhyasR7d3EKeEYdCSY1gt
vsuTuSSRNwdFW8Ax8+oF7UkiVHJhRL+SN91+6h2OIjuL+/A95uHNMn/r9vGV
BS3U4iLGMThJP8DbunhuWVkroTFYaoqk1w7VCCQNGH01FOsNc7meeTRTtdAH
KHap6HLgNNyaB0DLi2g7aGqgBa4KxoqwCJizQhP2VU5j90RWI8q4zX+LIQWH
ZssTy6KtVBe1GWp7mt2VnNPABuJWbpd8hPwvl79dKJiiiDdj5zxwdqU564R0
aSJxiSlXzJa14ECUQUh2t5gkEDePFOVXTexRLrd5JzSBxsdQ7NFTYH9ONfcR
l4XJd781yiammSIinI2B2AYPlSSVedWS9qAPd4j+wErg+BKUOWQZEpABN0Tp
wV7/S5Qs/N4nznF/B2ytyOC1BO4SbhrvMo9TPyR28QJuFTD5Tr7N9/u+kPTA
2LT55al8SeysLP78TD4/O9id6IKIYwInZH3KmDpIuPQ2JtjnIjUxGr/vahGO
AG/xIniHWdxctCuop/7VvLcV+kYEDBFDEpjLCKfwDDiFEyH4kBTObJR8F//x
YhZkGMa/drvdwtxkUpuBTw7czb6FZaEQFfT1+9Qd1P2JTIqXtZv/UCK7UH0U
ZlJYCBmYhSGFEiJE9v0q8001SrKVUWIlPUyNYZ6WBwv7Jr9Os4SUX/jXb7Cb
lVwYOUqVBSrff/86v5pKR86ddQQMLYIPqHNHk3d5VDPGe0wpjTN2fH4VRsj3
ukGIhdZOXrLRZSgEnyznmuSjSRxUlSYeNssYWy1/1M3b2mMpOrP6WjI5F9W6
82bllqMwfUrgg/SEu/e4H5XzaMaQw+vGfnJ2OBBt6gPMEA/jBAimVqctzGgD
hpPdHY8pEkEzp2zWWkF/ffXhe7M4vW3jOjFhtXG9IMxiheKgfCrdlsBx5vrj
9s5PXAUgwfvrcF77QcvlKaSln+qpgBO02Ingl/H6q3z5nANWuVFl8Nl3sZt/
v1mJ80Gqq0A5Us3cKq8z0RLJn/SoMNesxKLeY4gGGFMl1TgG22c09ZkHBXZv
w2TgJNbjDU+H8Z6jhzaFpd6s4A2Ru3WMG7ARl75m2oojwO0Jp2xxIj6V9Rwl
j1v7a6Xq6axiolMXDmDDwy+nn+BzAEmCzTK9p+0nPZTPWfWLm3yxwSayw5GJ
u4x3T5HvYWM8yW2FKmnUvVVPBtIuuICSk3a3Wm4ekSIgm+VukU83HGM9FTwe
TTa2PFx1dN4bYvW4WrhjshbuACecpHwDCTWWgCkubCQ5sSaamApQZbVEA6Ut
qzpIB+masJuKMh29C3mZ1HAtUXqa21FtMpTiPWBakTcpCYaPJk2ifamDzsOg
CUbmTmy8hWiQSNelyGC9p05jn4VJNdhYubbWx2N/qvu72aLnOlBgnS3o1e2f
rCppiZ+kTZlr2AJLDm4GcBeukToh0m7TRFidFIqkqPDJ0Z/bGoYzazWUvmYN
r/5caxPl36x3WD8tC0trcybGR48FA7Bc6huCgHv99sPV+3c/dT5cfvcDE1O/
ad8ODQhFMQN2Q+icsbp4h9BuZd6LSxtleI2eHlRoQrJrQcBqe9qR554rVhrk
szs1UU12oQZWto3Iso0CmL/VaqOYItlUrSyBdhQL7Rw6o9IPu0sXoCXdhyhQ
z3HcU6f3xWBqevE0z5ks86Ap2m3yuRRY2K1X5uwUxYipRAo60bz3vpr7855M
z55teBSPSKWTM6javFkin4VOHDL6t2D0GuyVNxa2Bibf3FfzqXdjo9aQmHFu
mQWNK1N1PVRlXej9pNTaAcx3K8VZq+vWWYoKm8BmyuoyX7qDbzu9AQEzW5/2
A2j8ljSQlVA8Y1m/LXDxcXVfp3CdZ0pKk6CYRzsrUxKkdh55bDJunW4rJmyt
m5c4sno4kt/kZeXu6sYQYmNbgOaZnkw2cyfz3equXCi/13l2rr/P9y9Rb/CT
IhjmN8+NM23FdEkPvfzk5Xf9enn17urmpvP6j1ev/ywXW7y5nF3EZPdbSXa/
/dRzBwfbxU6ReUGUkD2/FJRGko6hcwTiZcUd5QYsxFWlab/MirWaJ09XbEwT
HlQ2t13rRvejXERC2sH8UqenuHfJIHHcthHLFgn23Nb05Lwj5STKQ35+nKJS
bL1Mt+r4ORqEVFsBq/hoOJmVs/FBgdoKEXMS8+FspDLOx67RLUJhbxZAChC9
xBCWzgzDawfGUuO9lrjd9k+Oi7zfP8E8ROBDBz7t1fiJcGTn6aCiZOGhoklo
vRC02OpBBiKcla7J2/4xUEZfEGAybX/VXbNr9Wztiny8Uas47B9PUaE9jgQ4
x6wz7iTvvGReXBk5yPwVOkfuRwuLvO3Vrdc+lCtCTH70qqGDYKaeKPp3miXg
h6TaM1l4oOXnkhS+LsHMo1nn65A+B2cahQ4RVbawTfd3Wdtybn+oJ9VCDASZ
0ctHd8VXnUG315VqKffCyTqcflFXCPcRM5HKMwS5YcNNLM0HvcFxp9/rDAYU
PhG+qUJVdwRRVxzV8FVq1prXD7T9DjEUI6XukjkfvEvaHB168sUAdvLzS2/H
Uv7558u3f/jxOogmBG5w7sbj6vDoaFIE3vWu34IquKLTe7QT+DneQUA69+Bd
izkSzmwvG5J/qtS4/fKSVO2KFS64LlPA/Y1XgvACbIYFTNoLZ34D9TSPMb2U
P96Ybyar58e1IPM/3svOhsNWjS/tu8K9aP64WPWnveEgJB4uLcr8/11oXkLy
7ZR7hWaQAR1h+zAU/oyAQVB5yJUXDgNosbzLTA2ODFHavC/uKZGsoLZXLtor
FvlitGgaj7xz8kF6DMnr/kGF5gv67Ep4yyHx/U70DwrOaoqv91eXb368ipCc
dTc1X1cic3F3Ny29rrEywYt4hx/Tsd0AN9Gyod0DlIFBvVCsO0cAVXVoz2Jj
Jc+ef+ile+h33gO/++xoDUl015pBFXdt1wX5YafCQAUx+A1cp2seIEORThYl
1HvXAPzk8sGbXUf12Os42V1WL773vEvwWhmZYGUstkALlLkkvbHUObp9TYlI
Q86CQiWydDF+X+sC+fH6Q3pzRAL7n7w+5ELcYUOqEW3xSXNWzuhq2u0KjKuP
PURHZmxJO813hFashfpmA/0GJ6vMoigo9PjC1tRwleB0FuNxiizPhd8AlVrz
mQkHqCoREwYq4eQAxKAEqKYS+L0o/ioz99PyY9VZbJr7OoSDpu04rSGQDfJ3
bg3HVO3MV5rnk8vBJY2vBXxv9Vz4w9I6+fwaZfrfHPULRDkGB8EazuPeuy0l
sGXhDj096g3OplVUueRf9cr876K87+O9ov2TLshYzps88nXIzz8CsSNZWyFk
g6e3gfJnQe/UqWBl16zCGpSdleQJOCktEFK9QkNDEVNGem9rkJTxNCPBUo+7
YpwTDhWT7yshvWAVvIMP7nVOhD48AiBO0LpmenDF8xiQzgvtiiBN87RHqO+c
eJRVNYI48aCa+EP18LhczrukXTY4LLrQDFDedf3Q4DiUviNkPXmgAk22HjH+
otiKOYCD1Zg6QwXzoTgrf37b+endn+EKyjUf4V5khtL+aIFLmScnplk3W2eC
PRYBJN0QlyyhwXwZTU20pXhZTbF3yq7uV94e30kdFWuS1pZU4CT9FAI/7gpU
ADlI13EIZSOpPqoB604so+8lRtS0VSPp+PIpiOtmLeGAF8znIiA0eE/8JWb+
O3Yz90Ez3hCI/a5YqRNhtuFy1DQEp3W660sLa0eTwdmgd3hy0u12of70iyOt
L1N3+OiwN54MysNjtHDf93sKY3Wgy0qbHQL7EnmupTp+13EKDWZvXI6fx/ON
UxWcwtYPgxNwD7mYgVpehofYkdYsKH+SZOJtHiLSCsyo4BK6jWVWmZu1SwPB
K8HrupHNM5HE7lVd6kPAnJWjNrMgWobpCd5uvNSO8RfQ6ZFozUn47ru/KASC
Qq8bemeV8Be5zTw4NPHqniE3mnuC+7UU8kveSSxNVjswayrXN+NWg1rWEntK
TyF7TEVKvbIyasKAcCco7bhIW8qWh3x/5LbCWf9o2pO1xqRfFj7BRebwIh+d
VL2jyfGk75t8J9vGGsk8intAol6IKNoVAsiPkN4YMJ21ijTX80stS7KbzAeu
6B9NNIsg0URYzSlRD3QVghHUPcZAA1U8rywt2rQpO6ko0iifSp+y5N7Ohkgh
gE67SwCa+PPL3nl/5f7F7S+q7N296wnw1zRkylgDbuF6suSwLq/VJ4X3BhOe
YuBqtSrXZRFfMkgedUIyyzmbPqynL6l+0X7tXx17QYaH3Yl9tcBxEqi/sSRR
iA3n0T4fyscD2ZgzJ2/v84n40JpzZ3rIY9x/iXIaec3DsqRX+nO6VV95nVml
//XrdzwiaERXSJYHEOiElJL6bUtZqpHKHpIwwvs7bq93QqILYp3TDSU/AOAD
ithnAx0z/b9lTHo1crbDzSDn+4/v3l//27u3Tf562S3yH9bTnQ6HFrMhOISh
GDC7akdW1nHgcnhwqz6H965eG8ln7HHYZiV7f/W9SI9WnoLsPdUcXvAUv4RG
AqZrTXFweqz3HPjEMR1Z/MdL+Q6yTLCQke2grhY1rjXfgeapjA4RDtF3hUhC
c/5gl3c6eCg4wzvyXP6LT+O/+ZbRgcaqjNRDqg9ydEU0SZLTgl4c2pMo9+6v
j675sp5otaJmg4sa5eFNFNoYyLKS3irm3D19GFomlLAz8zJoESGFiXSb5k2I
99veahllUKIor9JrIAo3YbngRIHkclfcdqDM9tUdzHKzWkSmsC46Ri8K7hnp
1PuA7S4rFDDETJFO7kjDhEsA+2H3BR/P5a4eagZdi/29BSciGbrzRu68ODLs
rOM/OYv6BlOZ7/tfIEDwN0laUoAgkpYrOjF8q41na9hhq37Zx7R1FJl3BwGp
vrNht3/YPfFzn3rUoIIEn1p+1OuKC5z8Vpv13RJcoNX6CTtzuVj4eLSo6O4a
o7LXorupSAe81x+cfhu7+7ekt0X7CkqHRd4fftsfOvXu8Ns+S0n6Z9/2z6L0
c08lAPnhM84+W4L+DoTsGHpTO5bgajJEscd8t4SVbtSWYyIK5jPRlkemzMnu
PbIzrn0IFVVMGKGZmPxaftZvebE/LJ31+rFcN5v83V2JG1rNhO5isfyEDfA/
a3czbqYCF/eSY6Lf6Z1l2Q0ii0iqyTu9QbF1E5ibawkvwgxOpRAOzHaGA1Wb
ZVScOjM1f3uY53j2XovogJ3nuycmtZ/Xcb5XkgmWXBkKALJNn97N320R7mZm
yLd4pNtz8iUm6YzZB7/EPBa/idz5Q+qvKZmm7m/RLBxd5MA0nmqkqQzHkBdJ
/xgYE9P6ro5SBaXJq0a58w6AMgLmqH/kKX1xi0jMRi7D/kc+6lVH/eFRbzjs
T07P3H+UlCvMj7YrB8Ozfnk0HffPzsaTadluN9B2x5Pjk8PZ9OiwdK3dI9vt
htrODb1/Nh5PTk7d3j08mbXbHWq7ajY8OTmb9o5mk7PT6rDXbnek7WZn46Eb
yPHxSX84mU4O2+2OrX8nR4Py5PC0mp6Mj6vhuN3uRNudnpyWw3J4fHZycnI0
HW6N99TeWx33Z8cnk/7R2XBycnLcbnem7U7K08NhbzCu+tVkeFr2t+a5pw2H
h+OqqsZl2R+enfSOtiawbysyG/eGx6cnk3E5mQ7cqLca2pIc9srT2Xh8OihP
q5Pj060p7NuaVGPXwWF5fHR4OO6PZ9t9tEU5mfZOeoezsnfcK6flbLrV8Mjv
mnF1NK1Ox+Vx7/jouNpqaMuyxZ/dbmjr0huWPWcjVoPZZHI6O95a574tzGQ4
cduh1++Xbh+W060N1reVqapy3Ds9OzvrV8OT09MTNnxBlLv2g9OTI9f00G3J
02N3cLT9tkDCTE2ODsfHh4dlb+IWYajjUo+VCCZnC9sUHGMnDSc9d8oqmytf
TvPJr9Kpsw7PTtFoNp7pDLDE4bb6RM3DNmd1eDpxK3o8GZzOKl1QX+UchuRm
6cjtpdOh23FVP7QDuKyTXqHp8VE5cJ087c3K48FwdmRNV84YXjKJL2o8KKv+
1L38pDwbDAanbByCnPGDj3pnx4enp0PX+PTseKrDFy0dZC7JrM6mh5Oz4azX
P5odjw/PtBPNo9MI04bldNA7OZ4enrm5OnEbiw0B6R4PaTiYHk2O+243uzN3
aG9ul4mHX5wdj/vHZ1X/sOqfnvb7vZSBMJh2kOJ7n9Xg836SGGLKXZw/CzXI
Nz4dLo4cwNzUlNWXvTKkkPfZNVSJ5Odqv0a56qxE0SRDXyEALUYUibhnL6Vg
B2i8xyoQp/vaTXhoUVhhTxdKTeVaX6Ma4vuf3/98zU86vePCq2dwe6Lmgk7a
SkJ25H7iiWnVTKzKp3yfTNNILXvOB8O+pycVH4P6kw97JwOf1CikXjDu0x8O
8n3f3GkQvvLTv8mZhvWM1RyC1cPYMRm5VhVrUqCRD3q96dH07BgRw+9+uH77
xg3SnHP6egtQwuckfIDuniaiu0QtrH5vDpSWe5A+/oCFU7hH/BiuPT/JFs8V
MwqbTFlx89xpiSXpdmWXICGcVtdKzuDK4IxCIciFdw1j6p3SIlPZ2oQXmiln
PdBqBlb9hLCGAuq+W9Ad7b3RDKhDZTTb0z9asqJK4SUWrVlTs9yAgpLr5vXD
u3c3f7x2f/bOPn923QMpR+HthaW9b2mlkLXi1GtaOvmg8n3/0BP3kN7w/PDk
38Rb7T8+1Q9196RFW5WTj3PYr9jRbCMzThixrZOzf3P1Z21lftLIx2+G2tur
t++v34qlHm3QIlcUFjIOg8RZtM/+iSqaO2BqzamJiUiK9g+L4MBDFX3ROg3H
p7S35R3iJj7iBIuHIcYCS3y/9qaUcI0heEYH1Fiy7CI2l70iyZyGfZ7vS828
++Qg3fTSGFkAjWZoL9MNaRhe1QOS5mk1N+slUO737xtng32sO/K1WwHsh9qC
iOvambSWb04GVIHbcTO+FgclvAoX5sdKADat2jg9FX731cRVinqUedpM2Y2F
RtD1DHGmUMhieQXOMERkDDHm/K9ijf/6zapCldjngM0KT+ulRGSkb5eryb1b
uYlE9kLzDiI27kdkI/y4qe7nlZwDftEpo19JtphHQV4t5xogjEIL2fpZBA0c
VfZeOHwF7liyV3LihTbcSOiba7vnme0bjxbtk+2CdweRAg0zuOtRFrdO8NIy
4z/RwtiF9GRSxbUye0VIRFhtWD6ytsliZBTeDeT3iSRjmUxl5TfPqQO9Kejy
h43lFurm9fWHD7xp8/+zEWkmPVzNSqEHNswkxDWM3cIq0803PH62u3OR2fyA
xspuBpurkvmhxoQiZ0rbNzGpQqZhXp5JPClYta5D3vetTrzQLyKfZKFrsqwR
RZOTUhYYl1vBZzNpqdESWcZ01S4yhezXzgL2yDBU0EKue8nEUQ4EBZ2JNp8z
anUQXTFq46+wF/N/+FokP5bUyJUt41pxuo1OHNmKrNM0UY70RMv4Rg04VD8e
p38kzZjuKKvFR7Txf5uLiH2HtZZ6jYfn3mzGHXt2lCzp+vWAchHUsIylEGte
3SE8xswX8lB29WMUp6ObiMBSrxSoG3/+b6rVp3oi8wSPsJaOyNSd8/4ImlyK
HqxdglYChyr46CadzYLtxAsa3nwjwlWGsni2ZMKQgXLhA9grHr8mPRTymA9u
WA2c5pPn/IflHTrswZAI/k//7AXqY8QzRVHWTOq1sG75zixtve2diUK+VDW8
vZkUMNZJsq/vqTd+9g0AO90gr3QHneeBmqmQ035fj+u1Qqi6xlMQAUt1MT2q
RR7WtSCawrPmVRg1lsctC6uVoHJF2FN+hSZpT8ttxOqCtfn+Y6AFc0Jj0YpF
qltPizqnjwkfyHMKf6rKiZRAQrugqdJuXRKn+ALS1YN2RMAWKPdX/KgILymq
1Fg0m4fHgAJCfAyEPPYT0IiDC5tfKelgoidcv/KZzRzKKcvW9MnWvYiVNbos
Cb0iehXnKrUC/ebC3vugyQHIx2RhBqwnZDX6i4W1/qx5s2gO7yqT60iUFIsD
6MNq2qVxz1YcQ+4od5WUuFEt2Tak/hRR8mVyPMpGlbXX3717/+3rdzdXQl4k
r/32T3+96eY/6faD+ROoh8cCgfXkXyRFxWag4sKdL8tpSHqZuIPOAhUUKWq0
Hbbr4mME9R5J68x95tEJ1PItDTuEaSnQ039ZR3i/Uq4/DTJJy1pohgT28Bkw
fJ3dQWo04SURkPLNyqcaQveIZwlsOg1vNkgrpzhJbWNS11j4bMnpdIUih8xq
lr9JxOC56hVBa6OoyzL9XMouz84OhyAPvRP7fuWluMdzsOzj5NGZXguFYvMB
21lIihvmLIjQD4/IUfWpYgtWfrnOIky1xxUh2lCvov5r+dttReS0w6YcV6CC
uRfFfL0M5UiaTdW0NRJfoSBnN+29XWrd7Fox4ULGnVt/tfl9Lf6vv7776ert
h+sfr24+XP740w2KebCAyDvISJkdshc0NwNpeVE//CgNvRf59CgWU73LJ+Ru
3ZTd6MzTUmctMep6H8ZS8K2W8mbsVKT1BgQVdgSUkcjWmNrV1joHpbC09VRW
Cg4JBYduuJiwnGA3L0ysCOB0FlMrr8jCJw1/IqImSgZEIA/ABnH5LXIXQxCb
F3yj/E7P0DJ3rrOkTalQjBHd9Rz4zet9Q76i2orp4ZFhdBfHLLnWzvM/vL38
CTv5nXyRmEnr+7+7A4cGPG/HwyMLPLE1xe5SoHh85guanh6fydHUuykbIVV/
Mi/rB0jS+KpXZg85fKYMLsdaSRN3FbBRbBV+zgQJ3oxYUnRGk4cJzpK8QPlA
LDgX9FJl9ykClEM+QwqF5/IJ9pneim77PTsBprspypvSDncVejQcae4ZUX4U
Za7k27NEjeHujheeZQACeRT1WL1VXL+JYApkK7nKkRFtNXRcJn7NuTF4kyIw
ldJbLVdjPW9GGdd6eHasC6iLd++WRAo33HEqwf6Dd3j+AyB+QoEJVhQ6/5Xt
SIs+BW9oom1IWARgBtFgv6vn09JsdbEi9N/KwPLZp0SUqeGqSTjGyfSqtb28
J8hJ45uNNw7Mq4myAZ+DLUnNsRpUobgemeDWzfuKvUPFvC9i6phIiPoZS4kn
5qvnoWAn8Eva/rc47vy5m8cB+3ZRiPlY4psFsAcPS3UBPJRIvpwTTrGZbCSi
H4bwcbWs7uccRj3pwHzplGWVzLHmxXcqjy+QtybW3IOEbXCKv8TrqZKZ2csd
HbQS4ilFXXENPrk1vqs+PUZv5ySpcRA54TTHFyk55r1uz0JkD4biergIfDKw
lUEcHUnoPzOccM/dKI3Nrm+SSiI5DfFMjuvVR3dT/l33AQh1uXdjvbmJhhZG
I++FsZw0LXa4C3bo7JYRjSybVBgZaQHIqrxF2lTR2qaOEHhS4N9amLH5qskS
54wJJQDWehvYVA7zEornNC0I8C4YOP1unG2DurTXse4oQqHRb+jy+1mVDGjg
DRMyPSlTooVITmjCxSR5OI/4lVvltTsDd5LxQg/reDn1KRm10TKtMj4qVJmq
cuCJC0INAeko+IDFUh5nMj0T0LzIu0DopUAH7tZ5U3t/U1Ns6/zyAdgboRmJ
7NRqfKsG8BhMoUdaaCagLPV8XmRjz/vuf4ZBJwVLHDXTdjVmY3hWTvFfaYnY
Cu5wWUyQC11quTZTm9MZ0cGKKwa5f4WiyciOortdizTLBQuZBKZrqRcImCqE
1lPukHNNSwXTF+5x7M6/MfkNEa+wqFrbttQ8aCETlUrt+bx8FKIBIS9nNBDp
J8YWytOOuSw0EqFu+wNxwYifFMBntsNsd+m0Ou0PUtxN7+oO3XMPnlVSZAZk
wm6YK814FuYhSELdAWHxZLWgTf5ts/jooSPWkn8XU3gFkrHg/cRTZH6VUix5
hc9Z7+Z/ip+M++5uIVHGQp01MVu9+V5kv5A61yZzSgcM/0CylAzLpGUYrO1R
ALoJwe/ayrwfJNW1kWqmipqMVn5qbk80a4YtoIcwJkbVzNJkf9DMkWvV3mlx
VDwtlv4YWN1YWW032dN+V5BmHi+S9SH1p0i01fLxMTbXORtf60D0PjfG7wQ+
h88bP4c0BU1nIpGQm+oxm0FueV1HYkLuNXSYGjQAAv0RRCCqYuyxNRygIkTE
XKcrW3JsSQ6DUzQFEQHw2RHLYEBH6tvg5pWAQeVkmoVxmnJWrZ/pajeQzcw2
to2XoE/oj1KRl+nSijltCWsrznCmk2+3DBZgbtI1TIEBkLo1m8elddkWzTbK
7lFZALeYGSAiW52RIcCSvqZCsUQ85CTgUBaByVkdrko3akcr2T0jRNZvBUVl
hAheNa0VV8Jy48MqIesZFi0AWdRP59ZmtFw9ujdXinvsDC8wJ3rpDWKIeJgK
XiApjU/jjshiHax3TBP0SbNyPWHq43PE4FAxcboUxlraNSuLFLpNLhU4vkpI
VueB2QklEbPVLe72+2ZS0dpZ1c1Hp2pvrYjhfvmqZ3iUgj8GOcMMzMumRpUV
lJgP3kS3k6KpxtLYy2lc+r5019xijY63gF1O9OKYYLAlSPV3PONo+m4h0H+L
RugtWDFl8SuP2Xbuz15HKZpW5EFyBlf41KmADZJAxs/paDwdU5s/SYRz5o4F
QnZKcz+DuyV6V6j/glSZIzDD7PrgJs7MRyzbltelxcoV0o1oZMaxpV6c0s6f
GMwZPbcyBr5f923kozIdy/jB1KQkilBmXKlJQjvTRdoZ7XpWpNJMn06A9gxj
btVno7C3bXze+Fv2R7l6ynE9NzyEG0p3c99Soaz4l9Mn7fN2HbN6tJFQAG0d
9lW4p52ozoTAI/imb1LO8pv8X/IfJOuG16HtklX+ABAm/vleQGMV5sTpZ3ZP
zxCgA2rFqtk0/pO5SOnEWHL7lB12m+/qzZuby86Hyx+v3/5B8nP+KJVH61h9
U9g/eInhJDcRYLOZWQZQpLcEHGrV1jihEmjQWZSL66YIpzxzky+F/5WTqrgA
1tWd+Flrg3/7gQN930Z6+aiV4mFS3SFcsngfSJG1Yos+rgAeufJ19FaWLqLN
fPHu0jLMCv2lJSJFv242YyDaca/4/mU88FJaUvvwtYn2pSSH/5JUG8vnqFhL
vjBXpocJQcVjwEqkdqas4vNnUttHwUyv95RNpCOChIGrMDL6SNMCsta7U10q
3KiREZkK7mx/FBSVXW8kOQ1TQtx0blbirTNVNLzdzfcSRYu/4OLY53EId5us
teKNRZIiSwW/qGRGsUFrzlfeRFpIIfxUqnYVGUGGjXl830TMqHlerO+jITW3
j/NNcztXqjj9GmOMWo0rmf9bKLTixhoBggt58me70uRTwAL02yeBqNfJp9A7
m0yq029uwE4bKv9MJEmMUcsFjXmGbHzZuHpemrmKHkN9E7uOl13rQ9kFT7zh
vWGUMWdVVAWK0LfkM5AJfm3Hzzw/b2pLnFX5SfID8ebFrlH/RFyvHbCHesJY
j5KAaJcGTjIFA3AKBMJ/GkfRBudO8gs0NOuEoSuI59Mt5tvyra7b9WJWL9xh
HxVsDCE0OjvpHZ72ur0RtY7ZfFkyMULx2tSCUsmUsXlhfzfKWTD430eEz+JI
nPRgfySYLfYyvaCapJpNN4/A0o1YotyeXM6FbtQPeK57WCQdBRAsg2lFEaUh
JdW9fIOfP3zf6R+z3WYh1if69/O/fO/+BwdzuRa/qyDDCjCpUAj9t5Hv8Kq6
I9T3L48AqYUv746+IYnemlgqxd/lbk6nH0nyZiZgosplL49BzUYlnZAQ3cf6
UbQUtyMfwcESQWXjXCjqSkfgocV6T4rP8NBg2+eEKM7Ub6WTNamaUDDFSjgt
75XBnis1nS0Z3feZHGresgv3j30sYSfvH9BXqA4LLFDdwshnYDBTUakvlF9d
3ry+vi4UPFHOjyUYNVbGQj5i4qhKVDqDweKpfUGR+zXTVvps4jl7UTzHfntG
Iw1RKCcs5eUioBQqFeIaSY6NUZYQsHbhNzXmD7XRkKCM5mQeBz8+ckUezhtD
Cp3ob2TEdOxM2ctW5dpDiYqltFqJON5vb5ADghNWzKYxG+2xms/t9kd+6DE3
V7yVNPQZKSyMyGDas93TrqAhik2QLK0xnXq7MzN2NiaxCA9btPFBEFb94vb/
jZyRaZCUZmP7RFBfSRbvsDgZ4jbQs6fsz/FGKbyJL4X9C4R/qZa5ey8LEBBC
Ol03ACUX/TBKYWmlU9BGFtFtSK2G0CKHLz54pmRxAFtUqskgsr1AM+tFP7PD
91o+Ha8SRP0VRiMFn7fiA9gr9860qcRrHqnahfG+2F5eL5f5fAn8C1a7aa2o
jUty1LDXlhvFfDFabWshiWuWcOhsGfgVacOHmL/T2e7W9xKbEG3gcKc2oNC0
UAm4FXHNjSEmNKexCnBvpeLTiFNImnTmlbiFMPE4wtNMby7wDGqJ4jaGGY+7
nCQrzl9GhfdN/qZy670ScgXKwO/rRQmD6Y1gzugdv5If+Rve1z6vIaREp4vB
cYz+mdXTzIuWjBcn3ovMQHeYFGgHDNN7H0Ez5Yvv6rX6yhFIekVFQIgq1s+Z
u8Rbbnqr9IjAe4zYrpoTncVqJ2DTL5wthQ56b4hnYw/6GoajEQ7XO/cnSt+F
lsSp9wEJfzSTSRshD0LPB7NvPDkh9A118YQUxQxwnVLwqmmQM538rgL+wOht
AstPvnY3rRnMrnMXzsBaJQvqji/WU+GakWkr6ygnm+qwDw/gZdH10s1fi8cB
LpFq/QSQeBmGEw7yrgY2rdXyo2R8iYrTbLvQHZpnWK88Xa+wUPTSMwEn9tO5
F92hXfkoWhLc5VyNGH6+DAY5XYxZtcDZYe5z46c/YrSR9/9FvX9vZOnbnDZu
K3T8VkCYKSK20ZUL+eNP9LKJIGJ4P1w94Mcy743oeBZsIjaH32Ry89oT5Ox4
IgPPebJY4vei+Ddrdwst8CL8yz/Yu967+Z+FW3O5tpx8uhEeXHceai29SuuY
Az/MBJms8OP4xAjDZmEXu3lKKxX54+RVtt19pNOpZeHxdlsVdlcVSX5V4eM9
RTgriO87dYRgvcEeJpVw/s4gB8QhYZaRAf4xDa2Tfy/88pb/408/Jwy+ATzF
RFS+U0QJmsrvE1L5LiFllSLc/eWqMvUjYQf6gtzxdK1B9qrQdY9WsSt38Rdk
QpWbmIlkg0kFqTeSb5Map6hrcMuszakuVKvuk9vNaj6yDitTaqxPS8mbXh0M
QRBqqUmeLBHQfFKrWTjyUCYzutVtntzFyq/clapE0Err1Mp9rxVaLKI/gqvi
b5uphDnntXkGJfhQuhWLSapqi4dUd+Ucr9CRKPKIhg9gUG4Wxm7qXSEBLieC
6GMo6nndGjdD1kst3m+LUriGGXJISXUFi02yQTRDQepDlTU1qm0t2nR/sIdY
KEmv/LjyaQRTTTaNAwgevuDB91nBD1Tf0JpFgOyo6ytO1Do3RkJAFyAHRusr
2hXVTO0LKlVCDhBrr8ANddILwEpHfdqk6+Vj46NM1S8ihgCFJVAhaSKmDNcX
h7DbLSkVCkN2RaK/KHrk4WkAi7zPmpYBYFJrYL7sltRCVeQqEsovsKk9WWeR
rYJGZY1LKDa+ETaPQYlgy73EfxYupaQUTEZFFWufCCXK+Ub6rPynVf2pnOzI
7HjkFykdN+nnkuQ48QKhoIrZtzWrWOpV5ApWhkm7V1jIplkxeGJcauBko9v+
Mr7H8hkp2TpgfYb3KIjyL3kX3o0pZTS1qq+aOcPoENQnLFCRgfBsZ/MH0eZW
QlZwv/QjtiJllMIA0ciiXUwBAif7JYKg5H15wOmER6O0eIpWBNolBr0w0/S9
8cY45eSXcBTUjc0dZrQSP19lerF6F5H1Yx5mRobkCkS5GPy1jDLR5cNPxB4K
ufHNvYSFMiSJh5hvs4U1KToNc80ZFqVqHMRM9Zzd6y3pROUySbFUdha9L7xK
HwlIOUYZpS3i63LSPlXCq2mN0XXD+1mOQaSzIqXW5dvL7W1bl4vyszJg+fSy
e5xS/kKrW9wjOp0O+CrlYa+XTnNzPxenmHhDf94Vr22icNV5FtzrCSl7YWyC
cLVjMkRyjIrsK/5490PJRBEXNSnD5SdE7mIzVZ6loX3sTtLySRjSBF+QX4lC
kTw2Gxlcd/IuhXrY+SFMlluYprfQ9eQpctjazX1BkdGb32qRrfwAWpV+Aazw
23oBtU1+6kx7N9LF862WplUN3sE/hPAtbix/evZk+XFd3bL0iOzlnKtQu2Nf
Sm/tNuciSO7sx1sWWcd9NaXKfum/cotSTkkZ+2yUuIWkLWALu3FbFU/oiVB0
u7N9i5Wwu0d+xH0hI5+LjIsmPKRWSDu/ChIcuNVNP73FU6MfQdUN0xR+Jx6H
WydPfZek0F2TCb+yavnuJfhNk0tH4vZ05d6BlJmfJJwj1ifYKekgh4Be5hj4
SjDhjGsWRzQ9idK5j5x6UlqzAzgpvs5NCbmtzI2rGOaCJLu+kNH9BTs2rLD8
YLFcaOhFDh3PlTS1jOHdkPjuh76BzbQbklqyzXnLLCtSft4dXL/FVt6MzFCk
Zp0T3NH/DXDHdwbKHWY+Ui4013KazjOFlsga0qG65vC0xuzmMZW8DPVrFOzY
Jl/hIKfGkJgE2MzOtsEO2wZBAfNt/lpDkKQx6vQGItPb5EX/j4kWLSVDOZNj
LK5+CsFFmp6O0PSIT3GdtbDyLtBjPqdK2JBOI5DL6cnsqDzra/ndFqVTiseq
AII6Le7UiJsL2sOqOmcOf4ST7gM66eY517QzRXw32glWLFnWIlQjGIhx1ac+
T2Sh+8OJ2WnwX7cNg1WVe7f3foxo1sYsQ/bociEuYknElMTLLwhkBhZCpzoC
gwFtemaoqbGRqHOStC10CEwXf/bDgvHqWlFndRMlXXkof7nlB7dO8DTethb8
6FrDBYEfr76z9Ao/ZEH86Pb7B+zKtnQ8Z4ES0+pNNaEVxSSd0qd/w4jwpWDx
zrTHt0Z6LkQf4Wx1ebYs8AIDJMnZeStuPT0iLTi8ZSpvKF92wYHEnN5R8Wwg
jQ7pAj4uaRSuzgY8kdxb0HeHEYI4qVfIfwf4sZxFuUdY3aLnQUWe27aKL017
W8+ZgJpoGaYPXSwZzrLsOaOKVWQaKbB9+cBfuM+f7nXzZF4JNph2SqNFmM/m
Iv4r3wF/eJGBLjUmkVVC4EhMebykbfjwepG1GBZbBIuvCA1o2V6A2pASIqSP
q+Hg+qhyzFhL4j4HMsWL3MjX5fO2p/ZzpvWcsiNFB//M0UcG3w5IXmGh0w8v
3/x4fXNz/e4tT2vANyraqDxFDCyua9HCrL4IRQVpmXiz85bpb98y/f/CW+b0
i5eMoTratXGBDu+6ZA7jS2Z6fDjtVQOhOWTuSTvrxdL4lZPGADct+rDWElM1
iZN7KNt5D1kSgWJJLZ/yvS1/DNzbIXS5J3LAJLgxMVuilVKWyRWhvEZtM0gF
LP4IwlX9eigHlhLFqfrCRba/ZHhd5C+Yg8hBJxeQRWXDXcPUce1FnHHK2yWt
230hXVXrccU17Wz0BcFblriHYx5xSUpNgNtD7Ex2IfuwdSj5TstaT52HaYI0
nHF0aR71TeQkULvq6aNbjxeH55v9VCVijihCn89bjoikzEkr/BRkJ3hkAuhP
VIoVsad40CDznhAv6EEiJNMLrXK3Omj8rNCrD+Aw4mctYneOpsC7yW0V7lr1
hwcg0drSC80/wCBbJNlNF3dEKYqT1hUzXVuSQBAQlrtC/SJAZPrn7jO7t8Jp
/qduMCH78ymsjfb5OY8v4SOyF55etO8piaCuKlWDPHUui3HbcN8IFzKwT4PN
MwMb8d3nA++dY6in9QQoLurupts2869eKV2SppS9RDocKH8Iftt6g0XDt5jV
dlyoWbX4VK+WCw+vJhrProtVl659FWY7aPdaFFXFNtdwkdDNFFnKUmLfgrDD
NwU3QRGjO9gfWnoeFZcXcaFykcfFmqo5uGPeritGUeKclBvL1ceL7HddvL3t
i7f3X3jxHv62i1ev0gt02F+8jOXJTnITgH6F2/ewL1iNg9KSiuzKbXFt/Ja7
VpS87btWdc9znhiiJS4UNS2KVi3NWb9IYlxScC1Zv1JsrM7KfOSjVCM1cSL7
Kgrs+bQ3oVwqP1q1DsG54pfwSkqTy85/S/raytepgJnAZ7JZ2tyFkkmHpDSf
lqNZbe4271eHvR5ub2Bd1vr4izQ9SXAWECM7HPZ6hoOt3CRt2gNaWyHf7VwT
4QrFUNjuKdJ9lcxgt60Gxeww39cITMhmO7jgd/2egAzGji0nSiqnFYy0olPm
EoqEOKUM52v062f3OfQnri4w5P36fsEwJRAfOe9KX9TqtrkSzkmZC39NTarj
NSn59UKSYlT18l8YI4+i4BR6E0t6LKsUcSey+KUta9UHLUAmlal7iR2vU3jm
CbJIg4C8yuUYPEaI0PN6kmlQYrARRHW3pbhKUov6H14x0/92LFDQ2KPe6889
imtK/WYACl3F6d7UFesVyuY53fXfKwbv0O0IDAG1HpWiwkfsVlpkiprQBwkq
69RHKWl+CzFWrPnYEhnnFrR9F5G7n0eqLlME6R7YcLck8oM5j0oQ4qfUTrZ5
lDFzWt4WyjXj0tXI4cMQpLJ6+UfwkRGk1jkLi/vQj2Vxn+Ol9SBecVYpDjp7
eKEMbn4BayvXsTlUXlwgSurK+7SXAGhm6Zxa198J1i1VNRyXzQJuccSXzZsT
HZaXVLzi96p4har/NNe0Pmgx9QbB2L3841RSY3ZzHxdZ7JzId6h/NFuC474I
l9wOyvVUpcuDK4OK2NecEbHzYiaNd7kKjFeD2slvgwHhKNqM66L4MVelbpZK
aAI235PTI/VbBNmuYBBToB2ulxEvy6qhzyYuvvhcfNndcNn60Ck9rl3zOfv1
nI+opv+6h8QGw3PzwXkwXNMy6jwqiV53ubpj7ULEDe1urvvNuFsvs1C/El/c
OiIrkjZbW7sdZ5C6mz2oNoom7CWCJasYzzqzBi3PRq5OyFwk61pHmnQbXOif
ktlotiWq8oXr0Z704HSgp2UWl/yJ3lY97UiyJYLw1P09WfssCie1fPjYTX6U
NSjpTAiHu1fUD2W+T6bDA6mIZtq77ngDNJGqO/X2pocgeMIDaInmvlaNUsFn
O6ngt5myBaAr0kslL0WSBh+ZkfVb93433hM6oFJmYzVtH2C/ZglPt18RcWyE
8YlyilFlLYJ7YUXuvsDCukXYGvKxWnSv3E5G2WlUgGz9ZOyd1S+Pc7d3BdwN
fHNRFseXyeY0745A8WAHv0fqb7I8cKl320SO6Vna4YFVizKllu1m/xfNrs0D
gh0CAA==

-->

</rfc>

