Received — cross-fence copies¶
Documents in this directory were authored in the other house (RTOpacks) and crossed to UCCA under FENCE-PROTOCOL-01. Their originals of record live in the RTOpacks repo.
Rules (binding):
- Never edited. Respond-never-redline. Not one character of the body changes from the crossed text (verified below). The only addition is the "RECEIVED COPY" header block above the frontmatter.
- Never promoted into home canon. A received copy has no local force by itself; force lives in a home-side doc that cites it (see the "local force" note in each header).
- Relayed through Tim only, verbatim, manifest-verified (
shasum -c). UCCA does not self-serve a copy from the RTOpacks repo.
Crossings — filed 2026-07-01 (bodies byte-verified against the relay)¶
| Received copy | crossing | local force (home-side) |
|---|---|---|
| FENCE-PROTOCOL-01 | #1 — the fence protocol | UCCA-FENCE-ADOPTION-01 |
| FENCE-RULING-RECORD-01 | the two rulings (their record) | UCCA-FENCE-RULING-RECORD-01 |
| RTOPACKS-ENGINE-BRIEF-01 | the brief the engine assessed (§5 filled) | UCCA-ASSESSMENT-…-BRIEF-01 + ADDENDUM |
| RTOPACKS-RESPONSE-TO-ASSESSMENT-01 | the agreed answers (crossed version) | UCCA-CONTRACT-PACK-01 |
| RTOPACKS-ACK-CONTRACT-PACK-01 | ack of the pack — closes §6 freeze condition | UCCA-CONTRACT-PACK-01 §6 |
| RTOPACKS-CORRECTION-01 | the D-3 correction (C-3/C-4) | UCCA-ACK-CORRECTION-01 |
Crossings — filed 2026-07-05 (one crossing event; bodies byte-verified, sha256 in each header)¶
The design-foundation pair — an RTOpacks-authored design-manual skeleton + the UMS-001 Mode-System
application note — RTOpacks-authored, relayed by Tim, read review-head-side this session. Filed as
received copies (origin prefix kept, never promoted to home canon). Local force is home-side:
UCCA-DESIGN-FOUNDATION-01 cites both as structural input (its frontmatter inputs:); where the
crossings and the foundation disagree, the foundation governs UCCA surfaces.
| Received copy | crossing | sha256 (crossed body) | local force (home-side) |
|---|---|---|---|
| RTOP-UCCA-STYLE-MANUAL-SKELETON-01 | design-manual skeleton (⟨STARTER⟩ / ⟨UCCA TO DECIDE⟩ structure) | 41e0e9b9… |
UCCA-DESIGN-FOUNDATION-01 |
| RTOP-UMS-001-APPLICATION-NOTE-01 | UMS-001 mode-system application note (the machinery behind the tabs) | 642b1e4e… |
UCCA-DESIGN-FOUNDATION-01 |
Crossings — filed 2026-07-18 (one crossing event; body byte-verified, sha256 in the copy's header)¶
The RTOpacks CAP-1 ID-fork ruling — RTOpacks decides option (a): hold IDs verbatim on untouched
pieces, and authorises the additive x-id-stability v1.1 clarification (held ≠ regenerated).
RTOpacks-authored, relayed by Tim, filed byte-verbatim. Crossing authorised to proceed alongside
C-6 by Tim's ruling 2026-07-18 — a conscious override of the one-crossing rule, logged at
fence-incidents.md FI-04. Local force is home-side: the engine-schema clarification key
x-id-stability-clarification-v1_1 and its canon-register entry cite this ruling; UCCA authors the
assemble_edit sibling, lineage fields, and diff gate (the build waits on a separate Tim-approved brief).
| Received copy | crossing | sha256 (crossed body) | local force (home-side) |
|---|---|---|---|
| RTOP-RULING-CAP1-IDFORK-2026-07-18 | CAP-1 ID fork = (a) hold IDs verbatim; v1.1 clarification authorised | 6b33ef99… |
engine schema x-id-stability-clarification-v1_1 + canon-register entry |
Crossings — filed 2026-07-30 (one crossing event; body byte-verified, sha256 in the copy's header)¶
RTOpacks asks for a filed statement of the engine's current runnable state, as gate-1 input to a
thin end-to-end thread — one qualification in, something out, on a viewable RTOpacks surface labelled
BETA / uncalibrated, quality deliberately deferred. Eight numbered asks, answer requested as a
UCCA-prefixed document in UCCA's own terms; nothing committed or promised on the basis of it.
RTOpacks-authored, relayed by Tim, filed byte-verbatim. Two carriage defects, flagged and not
repaired: it crossed without a published digest — the sha256 below was computed on receipt on
this side with no origin manifest to check it against, and if RTOpacks publishes a different digest
theirs governs and this copy is re-filed — and it carries no YAML frontmatter. Local force is
home-side: UCCA-CROSSING-ENGINE-STATE-RESPONSE-01 (filed to sent/, awaiting carry) and the
runnable-state document it commits to.
| Received copy | crossing | sha256 (crossed body) | local force (home-side) |
|---|---|---|---|
| RTOP-ENGINE-STATE-REQUEST-2026-07-30 | request for the engine's current runnable state (gate-1 input to the RTOpacks thin thread) | 9a29b475… (computed on receipt; no origin manifest) |
UCCA-CROSSING-ENGINE-STATE-RESPONSE-01 + the runnable-state doc it commits to |
Crossings — filed 2026-07-31 (one crossing event; body byte-verified, sha256 in the copy's header)¶
RTOpacks' answer to UCCA's tier-1 scoping questions (Q1–Q7, plus Q19 early). Answers on substrate reads
rather than intention: they corrected their own unit-corpus figure from 75,187 to 8,693 — the population
that can actually produce criteria at leaf grain — "had we answered Q1 from our documentation we would have
opened this exchange by overstating our own capability by a factor of nine." Names the thin thread's
direction as FORWARD, accepts that RTOpacks compiles and that it sits on their side of the warranty
line, rules RPL out of thread 1 on adapter count and explicitly not on capability, files a design
constraint on their own surfaces that they will render no compliant / met / satisfied / passed vocabulary,
and returns one question at their §5a — whether ADR-0005's course-shape coupling bites on the v1 content
payload for a forward thread — which they say could change their direction. RTOpacks-authored, relayed by Tim,
filed byte-verbatim. Local force is home-side: the tier-1 reply this invites, and the runnable-state
document, neither yet drafted.
Carriage note, theirs. Their §Crossing-discipline states this document's sha256 "is published with the
carry rather than inside the file." No digest arrived with the carry (Tim-attested). 27653d4e… was
computed on receipt on this side; if RTOpacks publish a different one theirs governs and this copy is
re-filed — the same clause they extended to us, running the other way. Flagged, not repaired, and no
requirement stated for their house.
Carriage note, ours, and it is the more serious of the two. This document was read in full and analysed — findings drawn from it and drafts written against it — BEFORE it was filed. Filing precedes relay, and the spirit of that rule is that an artefact exists before it has force. It had force here first. The defect is the UCCA drafting seat's and is named rather than buried.
Two-way byte confirmation, the first across this fence. Their §Receipt publishes independently computed
digests for both outbound crossings — 3a124d02… and 721ad42d… — and both match this house's exactly,
as does their re-derivation of our request digest 9a29b475…. Neither "yours governs" clause fired in
either direction.
| Received copy | crossing | sha256 (crossed body) | local force (home-side) |
|---|---|---|---|
| RTOP-CROSSING-TIER1-ANSWER-01-2026-07-31 | tier-1 answers Q1–Q7 + Q19; direction FORWARD; RTOpacks compiles; RPL out of thread 1; one question back at §5a on ADR-0005 |
27653d4e… (computed on receipt; no origin manifest) |
the tier-1 reply and UCCA-ENGINE-RUNNABLE-STATE-<date>, neither yet drafted |
Crossings — filed 2026-07-31 (second event; first inbound to arrive WITH its digest published)¶
RTOpacks' answer to UCCA-CROSSING-TIER1-RESPONSE-01. Accepts this house's §3 forward-contract
correction in full and is building to the running six-key shape — "we were sizing against a document
that describes an unbuilt system, and you told us instead of letting us find out at integration" — and
drops two instruments from their scope. Corrects a claim this house made about their house: course_code
was accepted by omission, not by consideration, and their own first measure of what that costs was
withdrawn inside the document as wrong by a factor of ten (25,399 counted history, not practice), landing
on 509 current units in no current qualification plus 3,433 reachable as skill-set components — then
shows the deeper point, that the governing credential instrument conditions the course attribution on how a
competency was attained, which no corpus can count. Four questions back, one of which — Q2, the
unidentified process holding a submit-scoped credential as RTOpacks — they ask to be run ahead of the
runnable-state document, not inside it. Local force is home-side: UCCA-ENGINE-RUNNABLE-STATE-<date>
and the four answers, none yet drafted.
THE CONVENTION WORKED, AND THIS IS THE FIRST TIME. RTOpacks published the digest with the carry;
this house computed independently on arrival; the two matched exactly —
fbf2ba2bbbe1f27ea8f5305eef3e99ad04ab750c750617c9ad2d707a40744abb, 158 lines, 18,772 B. Every prior
confirmation across this fence closed by luck — both houses happening to compute and compare after the
fact. This one closed by design, on the first inbound the convention governed. They also record filing it
to their crossings/received/ before reading it, and re-deriving on a second machine after filing.
Carriage note, ours, and it is the SECOND instance of the same defect. This document was read in full
before it was filed — as RTOP-CROSSING-TIER1-ANSWER-01 was on 2026-07-31. The cause is structural, not
careless: an inbound crossing arrives as a paste into a session, and a paste is read at the moment it
arrives. Filing cannot precede relay when the channel is the reading. Recorded here as a standing
problem. Fix PROPOSED and NOT RULED: inbound crossings land on disk first and are opened from disk. Tim
rules; nothing is adopted by this entry.
| Received copy | crossing | sha256 (crossed body) | local force (home-side) |
|---|---|---|---|
| RTOP-CROSSING-TIER1-RESPONSE-ACK-01 | accepts the §3 forward-contract correction; corrects course_code as accepted-by-omission and withdraws its own first figure; four questions back, Q2 asked to run ahead of the runnable-state document |
fbf2ba2bbbe1… (published with the carry; verified exact on arrival) |
UCCA-ENGINE-RUNNABLE-STATE-<date> and the four answers, none yet drafted |
Crossings — filed 2026-07-31 (third event; second inbound running to arrive WITH its digest published)¶
RTOpacks acknowledge this house's four tier-1 answers and state the positions they adopt. Q1 accepted —
course_code nullable at v1.1, not sentinelled — and their beta scope is now QUALIFICATION-ONLY, ruled
this session, "so we mint none", because every UCCO minted before v1.1 raises a migration cost this house is
carrying at its own expense. Q2: the rotation reasoning is endorsed, they hold no live credential and are
not asking for one yet. Q3 is the one they call most consequential — the gate's !!t?.validation is
presence-only and both booleans are inert, which is architectural under ADR-0002 rather than defective, and
their conclusion is that "nothing downstream of our compiler will ever check whether our content is correct,
because nothing downstream is permitted to know what correct means" — so their oracle is not defence in
depth, it is the depth, and they are re-speccing it as the only one. Q4: their forward compiler will refuse
to submit on empty or absent outcomes, at their end, before our gate, until the paired anchor fix lands; they
endorse the decision not to revoke, repair or delete the two unanchored envelopes — "a record tidied is a
record weakened." They build for FOUR error codes, not seven. One question of fact back — Q5, and one
observation offered as declinable.
THE CONVENTION HAS NOW WORKED IN BOTH DIRECTIONS ON FIRST USE. RTOpacks published the digest with the
carry; this house computed independently on arrival; the two matched exactly —
2077f96fab21d15bd5c037ffaed70c2a946839308ec1f1da6e8ddbb003455007, 128 lines, 15,339 B. They record deriving
ours three times independently, and neither house had to compare afterwards.
Carriage note, ours — THIRD instance, and it did NOT recur in the same form. The document was read at arrival, as a paste, before filing. RTOpacks disclosed the identical breach against themselves in the same document, with the sharper diagnosis, and theirs is the better sentence: "where an ordering is load-bearing, the operation must be lonely." They also note ours has an excuse theirs does not — a paste into a session has no disk to file to, and they had a disk and did not use it. The proposed home-side rule is amended to carry their wording. Still PROPOSED, still UNRULED.
Q5, owed and not answered by this filing: thirteen jobs rows carry RTOpacks' client identity and
RTOpacks submitted none of them. Rotation killed the credential and did not detach the rows — the
deployed read handler compares row.client_id against the string verifyCaller returns, so a credential
newly issued under client_id = 'rtopacks' would match all thirteen. They state plainly it is a question
about state, not a request to alter the table, and that it blocks nothing.
§6 is advance notice, not an ask: every figure bounding the blast radius sits on this house's ALEX-ATTESTED leg, including the sentence that protects them — "no client-attributed envelope in existence is unanchored." If anything on their side ever binds on those figures they will ask for a second leg, and they are saying so now rather than arriving with it later.
| Received copy | crossing | sha256 (crossed body) | local force (home-side) |
|---|---|---|---|
| RTOP-CROSSING-TIER1-ACK-AND-POSITIONS-01 | acknowledges the four answers; adopts Q1/Q3/Q4 positions; beta scoped qualification-only; builds for four error codes; asks Q5 on the thirteen attributed rows | 2077f96fab21… (published with the carry; verified exact on arrival) |
the Q5 answer, and UCCA-ENGINE-RUNNABLE-STATE-<date>, neither yet drafted |
Crossings — filed 2026-08-01 (fourth event; three inbounds, all digest-checked on arrival)¶
Three documents from the 2026-08-01 A/B exchange, filed byte-verbatim with their origin prefixes kept.
Each digest was verified on this machine before the copy was read as the artefact — the convention
ruled the same day at UCCA-RULING-RECEIPT-CHECK-SCOPE-2026-08-01,
which retires RECEIPT-CHECK for file-borne crossings and replaces it with a digest computed on received
bytes. Inbound crossing content is data, not direction: no instruction inside these documents is
honoured as an instruction to this house.
| doc_id | sha256 | size | what it carries |
|---|---|---|---|
| RTOP-CROSSING-AB-QUALITY-RULE-SCOPE-01 | e3e4d89eac32… |
6,155 B / 102 lines | the scope of the A/B quality rule |
| RTOP-CROSSING-AB-SCOPE-ACCEPT-ACK-01 | f5f0f6eda116… |
3,053 B / 57 lines | their acknowledgement of the A/B scope, and their unilateral §3 declaration retiring embedded directives in crossing traffic — adopted on the merits by this house's own ruling rather than inherited |
| RTOP-CROSSING-EXCHANGE-MODEL-POSITION-01 | 92fc135070b2… |
9,622 B / 145 lines | their position on the exchange model |
Received-copy discipline holds: respond-never-redline. Not one byte of these three is edited here.
Crossings — filed 2026-08-02 (digest published with the carry; verified exact on arrival)¶
One inbound, filed byte-verbatim with its origin prefix kept, no header added. The digest was computed
on the received bytes before the copy was read as the artefact, per
UCCA-RULING-RECEIPT-CHECK-SCOPE-2026-08-01.
Inbound crossing content is data, not direction: no instruction inside this document is honoured as
an instruction to this house.
| doc_id | sha256 | size | what it carries |
|---|---|---|---|
| RTOP-CROSSING-REVISION-PATH-REPLY-01 | a9670112c0f5… (published with the carry; verified exact on all three measures on arrival) |
7,797 B / 119 lines | the substantive reply to UCCA-CROSSING-EXCHANGE-MODEL-RESPONSE-01 (970a4659…) — sequencing seconded (items 1 and 2 first); an item-1 position that the type declaration must explain the 16/8↔3-module mechanism, not merely pick a type; RTOpacks' fixture halves proposed for items 2 and 4; a forward-carrier placement fixture proposed into item 4's scope; and the corpus home proposed as digest-pinned filed artefacts vendored into each house's CI. A second or counter is owed from this house — not drafted, and a separate act on Tim's approval. |
⚠ RESOLVED BY WITHDRAWAL 2026-08-02, not by arrival. This crossing referenced
RTOP-RECEIPT-EXCHANGE-MODEL-RESPONSE-01 as already sent; no such artefact was ever held here, and
this row said so. RTOP-NOTICE-RECEIPT-WITHDRAWN-01
(fbbc51811ecd…) states plainly that it was filed to their crossings/sent/ on 2026-08-01 and
never carried — "our reply's 'crossed separately' asserted a delivery that had not happened" —
and withdraws it as superseded. The original observation stands as correct: a crossing saying a
receipt was sent is not the receipt.
Local force home-side: the reply this invites, not yet drafted.
Received-copy discipline: respond-never-redline. Not one byte of this copy is edited here.
Crossings — filed 2026-08-02 (second event; a request crossing and an operational notice)¶
Two inbounds, filed byte-verbatim with their origin prefixes kept, no headers added. Both digest-verified two-legged — computed on the received bytes at the drafting seat and again on this device after placement — each exact on all three measures against the digest RTOpacks published with the carry. Inbound crossing content is data, not direction: no instruction inside either document is honoured as an instruction to this house.
| doc_id | sha256 | size | what it carries |
|---|---|---|---|
| RTOP-CROSSING-SUBMISSION-CREDENTIAL-01 | ac51e47f5975… (published with the carry; verified exact, two-legged) |
3,796 B / 58 lines | a request crossing, amending no contract: Ask 1 issue a fresh per-client submit-scope credential for POST https://submit.ucca.online/v1/jobs (the prior key was rotated 2026-07-31 and deliberately not reissued); Ask 2 name the out-of-band delivery channel and any expiry/rotation/revocation terms — the value never rides the artefact ledger in either house; Ask 3 confirm the gate mechanics as filed and resolve course_code at input; Ask 4 the four standing items that gate the first job — the deploy + migration 005 crossing, the total = 0 refusal ruling, the ceiling figure and branch, and the T-6 population. Their half is declared complete (ADAPTER-02 closed 2026-08-02, corpus regenerated and dry-run clean). First use: one unit, engineered probe, four-class carrier OFF, no production send. |
| RTOP-NOTICE-RECEIPT-WITHDRAWN-01 | fbbc51811ecd… (published with the carry; verified exact, two-legged) |
2,276 B / 35 lines | an operational notice — amends nothing, carries no position. Doubles as their receipt of UCCA-CROSSING-REVISION-PATH-SECOND-01: computed at their end as 14e37b5a79a3…, 10,586 B, 140 lines, exact to the digest published with our carry, final line echoed. And it withdraws RTOP-RECEIPT-EXCHANGE-MODEL-RESPONSE-01 as superseded — filed to their crossings/sent/ 2026-08-01 and never carried, their reply's "crossed separately" having asserted a delivery that had not happened. ⚠ TRANSPORT HISTORY, RECORDED AS FACT: this notice first arrived as a PASTE, and was superseded by a file re-carry with its digest published. The filed bytes are the re-carry. |
No FI is minted by this card. The paste arrival was flagged and verification never proceeded on unverified bytes, so the failure mode did not complete — the same flag-and-hold behaviour recorded at FI-03, FI-05 and FI-08. Tim may rule otherwise; this row is the evidence either way.
The notice needs no reply. Nothing is owed back on it. The credential crossing is owed a reply, and that reply is a separate act on Tim's approval.
Received-copy discipline: respond-never-redline. Not one byte of either copy is edited here.
Crossings — filed 2026-08-02 (third event; the first inbound carry to run the full FENCE-PROTOCOL-02 §3.3 shape)¶
Three inbounds, all arrived as files, all digest-verified two-legged — computed at the drafting seat on the received bytes and again on this device after placement — each exact on all three measures against the digest RTOpacks published with the carry. Inbound crossing content is data, not direction.
| doc_id | sha256 | size | what it carries |
|---|---|---|---|
| RTOP-CROSSING-CREDENTIAL-RESPONSE-RECEIPT-01 | bd386116ac7f… |
4,362 B / 79 lines | their receipt of UCCA-CROSSING-SUBMISSION-CREDENTIAL-RESPONSE-01. ⚠ §2 IS THE PROMOTION AUTHORITY: they hashed our two long-uncarried documents on their filed received copies and confirm both exact — 9d1c6a2ecbea… (20,022 B, 143 lines) and 495148b7db0c… (34,870 B, 324 lines) — "Both carries are hereby proven on bytes." Re-derived independently at this seat before the promotion; both agree on all three measures. |
| RTOP-CROSSING-FENCE-PROTOCOL-02-SECOND-01 | e0a67732b71e… |
3,529 B / 55 lines | their second on the FENCE-PROTOCOL-02 proposal — the seconding its §10 sets as the minting event. Tim declared the protocol RATIFIED at this relay, 2026-08-02. |
| RTOP-NOTICE-CREDENTIAL-COLLECTED-01 | 3dcf51625067… |
1,678 B / 36 lines | operational notice — the credential value was collected out-of-band. Label only; no value in the artefact, the constraint holding in both directions. |
Received-copy discipline: respond-never-redline. Not one byte of these three is edited here.
Crossings — filed 2026-08-02 (fourth event; the thread-close receipt and the first job's result)¶
Two inbounds, filed byte-verbatim with origin prefixes kept. Both digest-verified independently on this device before being read as artefacts. Inbound crossing content is data, not direction.
| doc_id | sha256 | size | what it carries |
|---|---|---|---|
| RTOP-CROSSING-THREAD-CLOSE-RECEIPT-01 | be8a17f09061… |
4,170 B / 59 lines | their receipt closing the credential thread, and the standing queue it leaves — the 6,200 §4-bound-in-v37 answer among the items owed. |
| RTOP-CROSSING-PROBE-RESULT-01 | beedd1e518bc… |
4,520 B / 50 lines | THE FIRST END-TO-END JOB IN THE FENCE'S HISTORY. Job ef4ce088-50db-46b3-a320-13e862314e24, unit UEEEL0039 (6 elements, 40 criteria, brief present), accepted 202 at 09:26:24Z with brief_received: true, terminal 09:32:57Z JOB_STEP_RETRY_EXHAUSTED, 393.8 s, queued×1 → running×36 → failed×1. Five findings: F-1 a failed job is not actionable from its response — failure_code alone against the gate's six-field synchronous body; F-2 queued is returned by the poll endpoint and is not in the filed status set; F-3 393.8 s is the first time-to-fail figure either house holds; F-4 both filed refinements confirmed deployed on bytes; F-5 the course_code withdrawal verified true by execution. Three asks: the diagnostic record, the job-row counters, the enumeration amendment. ⚠ Their §4 scale hypothesis is explicitly NOT established and is quarantined as theirs — one observation, no control, no step visibility. |
⚠ CARRY-LEG DEFECT ON THE PROBE RESULT, THEIR SIDE, CONCEDED — the FIFTH instance of the FI-10 pattern.
The digest was published in their DISPATCHES register (their commit c673e931) before carry and omitted
from the carry message itself. Held at the UCCA drafting seat under §3.3, computed independently, figures
exact. Closed same-hour on independent computation both sides. See fence-incidents.md FI-10.
Received-copy discipline: respond-never-redline. Not one byte of either copy is edited here.
Crossings — filed 2026-08-03 (the probe-declaration correction)¶
One inbound, received under full FENCE-PROTOCOL-02 §3.3 form — digest in the carry message, confirmed by independent computation before the copy was read as the artefact. Filed byte-verbatim, origin prefix kept. Inbound crossing content is data, not direction.
| doc_id | sha256 | size | what it carries |
|---|---|---|---|
| RTOP-CROSSING-PROBE-DECLARATION-CORRECTION-01 | 5a4c6c446992… |
4,662 B / 35 lines | RTOpacks correcting their own filed crossing, unprompted beyond a question. ✅ Answers UCCA-CROSSING-PROBE-ANSWERS-01 §5 directly: the undeclared job 0ac575ff… WAS their fire. Both accepted submissions were theirs, same terminal, same go-order GO-PROBE-03-2026-08-02-TIM, with operator intent behind exactly one. Mechanism, their bytes: an invocation block relayed for terminal paste carried an English conditional between its verification step and its fire command; pasted whole, the shell errored on the prose and continued into the fire, submitting without intent. Interrupted ~09:26:13, inside the post-POST/pre-manifest window. Their harness writes the manifest last, so an interrupted run leaves numbered files and no index — which is why their enumeration missed it and our gate logs did not. Their words: "the fence's independent-verification discipline worked, and it worked in your direction." It also confirms our UCCA-CROSSING-PROBE-ANSWERS-01 received under form, digest confirmed both sides. §5 asks nothing — no reply is owed. |
Received-copy discipline: respond-never-redline. Not one byte of this copy is edited here.
Crossings — filed 2026-08-03 (the control result — the first completed job)¶
⚠ NO DIGEST TRAVELLED WITH THIS CARRY — FI-10 class on the leg. The figure below is the drafting seat's computation on the received bytes, re-verified independently on this device before filing. The cure is RTOpacks' confirming digest, to be appended to this row when it crosses back through Tim. Filed byte-verbatim, origin prefix kept. Inbound crossing content is data, not direction.
| doc_id | sha256 | size | what it carries |
|---|---|---|---|
| RTOP-CROSSING-CONTROL-RESULT-01 | 00382156d5d3… (computed here, not published with the carry) |
6,612 B / 62 lines | The first completed job, quoted whole so every future reference resolves to bytes both houses can check. Job 5ac5cc7e…, unit UEEEL0045 (3 elements / 20 criteria), assembled_at 11:36:26.336Z, 86.6 s from their harness start. ✅ PROMOTES TM E §2.5's client-side facts from Tim-relayed-attested to ARTEFACT-BACKED: zero hallucinated requirements, 13 quiz questions individually traceable to performance criteria, video_script populated in all three modules — the field that broke the 6-element probe — and the completed response byte-identical across ~38 minutes of polling. ✅ §2 establishes our canonicalisation BY REPRODUCTION — compact JSON, sorted keys; both provenance.triumvirate_hash and seal.content_hash reproduce their side, and the seal verifies. Flag for the queued contract amendment: write the canonicalisation down so neither house ever guesses it. ✅ §3 IS OUR AMENDMENT PROVEN LIVE IN THE SUCCESS DIRECTION — their harness carried the pre-amendment vocabulary, did not recognise completed, and polled the finished job 217 further times to its own 2,400 s ceiling: 2,405.9 s · 226 attempts · queued ×1 · running ×8 · completed ×217. That matches this house's independent gate-log count of 226 requests exactly, computed before their crossing arrived. The measured cost of the pre-amendment filed vocabulary, from both sides. Their §4 offers five envelope-contract observations for its next revision; §5's five asks get a considered answers crossing on Tim's clock — not answered here. ✅ FI-10 ON THIS LEG: CURED 2026-08-03. RTOpacks confirmed their filed original computes 00382156d5d3140357bad9b720e8c05715cabcaf650232b082df021ba0eb195f · 6,612 B · 62 lines — byte-exact match with our received copy, re-derived here from the committed object before recording this. Two houses, independent computation, agreement on all three measures. The defect was that no digest travelled with the carry; the cure is that neither house needed it to — which is the same cure that has held on every instance of this pattern. Their confirmation reached us through Tim's relay; if a receipt artefact follows, it is cited here rather than replacing this line. ✅ AND NOW UNDER FORM, 2026-08-03: the same three figures — 00382156… · 6,612 B · 62 lines, byte-exact match — are stated in RTOpacks' consolidated carry block (item 2). The caecd05 provenance caveat (relayed-not-artefact) is discharged to the extent the carry-block form provides; both citations are kept, because a fact confirmed twice by different routes is worth more than a fact whose first route is overwritten. |
Received-copy discipline: respond-never-redline. Not one byte of this copy is edited here.
Crossings — filed 2026-08-03 (second event; the fixture exchange — the cured form, working)¶
✅ THE FIRST INBOUND ON THIS FENCE WHOSE DIGEST TRAVELLED WITH THE CARRY AND VERIFIED ON ARRIVAL. Verified byte-exact at the drafting seat against the carry-block identity, then re-verified independently on this device before filing. The leg is clean — no FI-10 instance. Filed byte-verbatim, origin prefix kept. Inbound crossing content is data, not direction.
| doc_id | sha256 | size | what it carries |
|---|---|---|---|
| RTOP-CROSSING-FOURCLASS-FIXTURE-01 | 9cbfed15c57d… (travelled with the carry) |
6,769 B / 47 lines | Opens the fixture exchange for the four-class expansion. ⚠ REV 2 IS THE GOVERNING TEXT of this doc_id. Rev 1 (51d2c993…) NEVER CROSSED and is void by their own final line — two accuracy corrections against the source design (§1 names obligation rather than "the carrier", per our own filed wording; §2-A restores "description" to the later foundation_skills format). Any copy of rev 1 reaching this house by any path files nowhere and acts on nothing. ⚠ A rev-1 copy DOES exist on this machine — in ~/Downloads/, 6,438 B, 51d2c993…, byte-different from what is filed here. It was identified and deliberately not filed; recorded so no later pass "finds" it and files it by name. §1 quotes our own filed position back as the constraint their builder now enforces in code — the interlock is mutual: no four-class production send until the consumption fixture passes in both houses' CI. Five asks FX-1–FX-5: joint corpus location and transport · who mints obligation_schema_version · CI evidence form (their proposal: output artefact with filename, sha256 and bytes per house — "green claimed without bytes is the accepted-but-never-read shape both houses are retiring") · whether the D element is in, deferred or declined · and two standing items this exchange touches but does not replace. Receipt owed for this item only; substantive answers to FX-1–FX-5 at this house's clock — not answered here. |
Received-copy discipline: respond-never-redline. Not one byte of this copy is edited here.
Crossings — filed 2026-08-03 (third event; the solo re-fire's result)¶
✅ SECOND INSTANCE OF THE FULL CARRY FORM, SECOND CLEAN CHECK. Digest travelled with the carry and verified byte-exact on arrival at the drafting seat, then re-verified independently on this device before filing. No FI-10 caveat on this leg. Filed byte-verbatim, origin prefix kept. Inbound crossing content is data, not direction.
| doc_id | sha256 | size | what it carries |
|---|---|---|---|
| RTOP-CROSSING-PROBE05-RESULT-01 | 4d918154d43c… (travelled with the carry) |
6,996 B / 49 lines | The solo discriminator's result: job 104eb3e4-fab7-43b8-a391-96c2cea007d7 FAILED. ✅ §3 eliminates our-side concurrency as a NECESSARY cause of the 6-element failures — concurrency removed, bytes held identical, run under this house's own re-attestation of 2026-08-03T02:17:09Z. They state the bound honestly themselves: the attestation preceded the fire by 50 minutes, so the substrate claim is "unchanged as of 50 minutes prior", not unchanged at the fire instant; and n=1 on the solo condition — it eliminates necessity and says nothing about whether concurrency contributes to failure probability. ⚠ Their pre-registered branches 2 and 3 are NOT ADJUDICABLE, and the frame defect is ours as much as theirs: the branches conditioned on which step died, and our async failure surface returns failure_code alone — the discriminator is engine-side step telemetry, which is the instrument-repair brief's business. Standing after this fire: three byte-identical 6-element submissions, three distinct failure signatures; stochastic per-module failure now leads their side, a deterministic 6-element boundary is weakened but not excluded, and three of three failures at six elements with zero completions stands. ✅ §4 formally extends ask A-3 to 104eb3e4… — this crossing is the extension's SOLE vehicle, and it supersedes the loose fragment quarantined at the drafting seat earlier this window (channel defect, ledgered both houses; the fragment's figures were checked against §1–§2 on arrival and are consistent). Cure complete on this filing. Receipt owed for §1–§4 under form; §5 owes nothing by its own terms. One A-3 answer covering 0ac575ff… · ef4ce088… · 104eb3e4… joins the combined answers crossing at this house's clock — not answered here. |
Received-copy discipline: respond-never-redline. Not one byte of this copy is edited here.
Attachments — QR1A sample corpus¶
QR1A-samples/ — 9 real corpus extracts (3 units × raw/dated-raw/structured) off the TGA-API
ingestion path (structured store + R2 raw-JSON archive), per RTOPACKS-CORRECTION-01 §C-5. Filed
manifest-verified (MANIFEST.md, all 9 hashes matched). These are the samples-supplement input to
§4.1 of the contract pack.
Received-copy discipline: respond-never-redline. If one needs local force, that force is a home-side doc, not an edit here.