Skip to content

T089: phase 3's checkpoint, quoted; T092's phase-3 notes (plan 034) - #1238

Merged
brettheap merged 1 commit into
mainfrom
bookkeeping/034-t089-phase3-checkpoint
Oct 5, 2026
Merged

brettheap merged 1 commit into
mainfrom
bookkeeping/034-t089-phase3-checkpoint

Conversation

@brettheap

@brettheap brettheap commented Oct 5, 2026 •

Copy link
Copy Markdown
Contributor

Lane: openxfactory-4 (openXfactory-4-openDox_extraction)
Authored by lane openXfactory-3, delegated by lane openxfactory-4 (#656 5984809456)

Plan 034 T089, phase 3's checkpoint, and T092's phase-3 notes. This PR runs every check T089 names, at the commits phase 3 pinned, and quotes each one. It also carries T092's phase-3 notes in docs/opendox-carve-manifest.yaml, on the holder's decision below.

Ruled, as T089 cites: F5.2's close, 5962785556, item 1. The amended forms run: batch H (F10.1, F13.1), batches M and P (F16.1), batches C, F, G and K (F5.2), batch A (F11.1's guard).

No Rule 6 window. This PR touches nothing under openspec/changes/: the checkpoint finds no record in #1144 that is wrong, so #1144's tasks.md is untouched.

The record is specs/034-opendox-standalone-operation/evidence/checkpoint-phase3.md.

  • Each falsifier in it was extracted by anchor, byte for byte, from File add-neutral-product-standalone-operability: open the BUILD arc, recording ten rulings #1144's tasks.md at 36908480 (sha256 48db295594a6…), and every extraction reproduces the sha256 the earlier records quote (F4.1 8900b985e72f; F10.1 d94265449831 → 2358bd8f76ab; F13.1 1e1f7da5564a → 0a47fa31f5e0; F16.1 fe22ec4873d1 → 619b32a955e2; F5.2 808619e066fa → 4b1512c26814; F11.1 60beede1244b).
  • Each ran from a fresh detached worktree at the pinned commit, with an empty git status --porcelain --ignored, in a scrubbed env -i environment, in the foreground of its wrapper.

Verdict

# T089's check where result, quoted
1 F4.1 openDox-code dede32b4 PASS, exit 0: 1 passed twice, then no deferred reach names the consumer or the publisher
2 F10.1, as batch H amends it openDox-code dede32b4, pip install ".[local]" (opendox-0.1.0) PASS, exit 0: the bundled database, migrations ['0001', '0002']; validation: opendox-snapshot: 0 violations; serving http://127.0.0.1:8080/index.html
3 F13.1, as batch H amends it openDox-code dede32b4 PASS, exit 0: the bundled database, migrations ['0001', '0002']; every assertion holds, the last grep -q OPENDOX_OIDC_ISSUER …/default.err
4 F16.1, as batches M and P amend it, with T100's named test openDox-code dede32b4 PASS, exit 0: 24 passed (16.6), the dialect line, no model configured: the catalog offers nothing, 83 passed (tests/test_chat_model_configuration.py), 532 passed (tests/test_model_binding_trust.py)
5 F5.2 whole, as batches C, F, G and K amend it, after T086's repair openXdox-code 56e1c238, openDox-code dede32b4 installed over it, OPENXFACTORY at 36908480 PASS, exit 0: the seven suites pass whole (45, 23, 18, 6, 40, 9 and 12 passed), test -s holds, and the --chains step prints ok: 5 protected edit(s), each entered and holding
6 T098's interim F11.1 openxFactory, ARC_TIP at T094's landing 36908480 PASS: quoted from f11.1-phase3.txt (#1237 → 20ce593e), and re-run here with the same line, requirement 1 holds: 0 note(s) annotated, every other path a declared surface (11.1)

T092's phase-3 notes, carried here

Where they ride, and which reaches they cover.

3 notes on 3 rows (2 added, 1 extended), for the eleven reaches phase 3 closed. They are #1144's 4.3 reaches into openXdox that F4.1's scan still named at phase 2's pin 047bb4fa, and the three rows of openXdox-code's ratchet OPENDOX_BACK_IMPORTS. All eleven drop at T084 (opensoft/openDox-code#77 → e49b17c3), which also retires consumer_reach.py.

entry carve lines reaches at 1e4a57fb note
branch_session.py edits[0] 1574, 1992 :1587, :2005 extended (phase 2's note left these two open for phase 3)
serve_project.py edits[0] 246, 247 :246, :247 added
serve_workbench.py edits[0] 346, 406, 407, 543, 1214, 1664, 2606 :347, :407, :408, :544, :1215, :1665, :2607 added

Each note cites the landings that close its reaches, each re-verified at this PR's run (closers-p3.py, quoted in the record): T084 (opensoft/openDox-code#77 → e49b17c3), T086 (opensoft/openXdox-code#37 → 56e1c238: openxdox.column_contributions.register(), and OPENDOX_BACK_IMPORTS at (0, 0) with its pin at T087's commit dede32b4), and T094 (#1236 → 36908480: the host's call).

So the stand-ins consumer_reach.py still held at 047bb4fa (gate_console, serve_gate, serve_projection and the two late bases) take no note, on ruling (A).

F11.1's content rule, T049's script byte for byte, over main's manifest and this PR's:

  extended: scripts/ideation_dashboard/branch_session.py edits[0]
  added: scripts/ideation_dashboard/serve_project.py edits[0]
  added: scripts/ideation_dashboard/serve_workbench.py edits[0]
F11.1's manifest content rule holds: 3 note(s) annotated, nothing else in the manifest moved

With every edits[].note removed, the two manifests are equal, and no other line moved. The note check, the reach mapping, scripts/validate-carve-manifest.py (OK) and F4.1's scan walk are quoted in the record's last section.

The pins

From pins.sh over openxFactory at 36908480, with openDox and openXdox initialized recursively: openDox root e1e3a3c3 (code → openDox-code dede32b4, T087's pin), openXdox root 9564d5d9 (code → openXdox-code 56e1c238, T086's landing; its contracts/opendox-pin.yaml → e1e3a3c3), and openXdox-code's pyproject.toml pins openDox-code dede32b4. All six T090 equalities hold, and the runs used exactly the pinned commits.

What changed (5 files)

  • specs/034-opendox-standalone-operation/evidence/checkpoint-phase3.md: new, Status: record. It is linked from the feature 034 entry in README.md, beside T098's F11.1 record, and from plan.md's evidence tree.
  • specs/034-opendox-standalone-operation/tasks.md: T089 ticked, with its Run bullet (T063's form), and T092's entry gains phase 3's sentence after phase 2's (T063: phase 2's checkpoint, quoted; T092's phase-2 notes (plan 034) #1218's form; T092 stays unticked, since 11.1 is ticked at ARC close). The box map already names T089 for F4.1 and F5.2, and nothing moves, so it is unchanged.
  • specs/034-opendox-standalone-operation/plan.md: the evidence tree and its prose.
  • README.md: the checkpoint link.
  • docs/opendox-carve-manifest.yaml: T092's phase-3 notes, note: lines only (2 added, 1 extended). With every note removed the manifest is unchanged.

Validation, at the head against main 20ce593e

  • python3 scripts/validate-openspec-cli-pin.py --all --strict (pinned CLI 1.12.0): rc 0, Totals: 110 passed, 1 failed (111 items), 0 UNDISPOSITIONED failures; the one failure is the accepted add-chain-attestation exception, as on main.
  • doc-health --single-repo . --as-of 2026-10-05, run at main 20ce593e and at the head in two worktrees of one clone, each named openxFactory, so like is compared with like: rc 0 at both, Findings: 31 critical, 26 error, 69 warning, 20 info. New regressions vs previous report: 0., and the two reports are byte-identical, a delta of zero.
  • Every 8-character commit sha in the record and in this diff resolves with git rev-parse --verify in its own repository (openxFactory, openDox-code, openXdox-code and the two roots). The two spec shas in the record's pin table, f7ee3c76 and f088b097, are the roots' gitlinks, read with git ls-tree.

Rehearsal (2026-10-04, at that day's mains, before T086, T087, T094 and T098)

At openDox-code 38d3350e, openXdox-code 6a3b93b9 and openxFactory c8d80020 (ARC_TIP fcb45380): F4.1, F10.1, F13.1, F16.1 (24, 83 and 135 passed) and F11.1 passed; F5.2 was red as expected until T086, on tests/test_session_snapshot.py (17 failed, 6 passed: 16 SeamNotRegistered at openDox's kickoff seam, 1 IndexError) and tests/test_snapshot_validation_launch.py (9 failed: T070's install-mode precondition, hosted by default and refused for no OPENDOX_OIDC_ISSUER), with the other five suites whole and --chains at ok: 4 protected edit(s). opensoft/openXdox-code#37's scope covers all 26.

Host paths in the quotes are written <workdir>, and the user <user>. No closing keyword appears in this body or in any commit.

🤖 Generated with Claude Code

Summary by Sourcery

Complete the phase 3 standalone-operation checkpoint and record its passing validation and carve-manifest bookkeeping.

Enhancements:

  • Record and link the phase 3 checkpoint evidence for the openDox standalone-operation plan.
  • Close the phase 3 manifest bookkeeping by documenting the eleven reaches retired through the completed seam work.
  • Mark T089 complete and document the corresponding F4.1 and F5.2 checkpoint closures.

Documentation:

  • Add the quoted phase 3 checkpoint, pin verification, test verdicts, and evidence procedures to the feature documentation.
  • Carry T092 phase 3 carve notes in the openDox manifest and update the feature evidence tree and task notes.

Tests:

  • Verify F4.1, F10.1, F13.1, F16.1, F5.2, and interim F11.1 at the phase 3 pinned commits, with all checks passing.

F4.1, F10.1 and F13.1 (batch H), F16.1 (batches M and P, with T100's
named test) and F5.2 whole (batches C, F, G and K, after T086's repair)
run at the commits phase 3 pinned, and T098's interim F11.1 is quoted
and re-run; recorded in
specs/034-opendox-standalone-operation/evidence/checkpoint-phase3.md.
T089 ticked; README and plan.md link the record.

T092's phase-3 notes ride here on the holder's decision of 2026-10-05:
3 notes on 3 rows (2 added, 1 extended) in
docs/opendox-carve-manifest.yaml, for the eleven reaches T084 closed
(openDox-code#77), on the holder's ruling (A) (#656 5994463071).
F11.1's content rule, the note check, the reach mapping and
validate-carve-manifest.py run over them, quoted in the record.
Bookkeeping: no Arc trailer.

Lane: openxfactory-3

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Copilot AI balanced review requested due to automatic review settings October 5, 2026 14:57
@sourcery-ai

sourcery-ai Bot commented Oct 5, 2026

Copy link
Copy Markdown

Sorry @brettheap, you've used your own review budget of 250,000 diff characters for the last 7 days.

You can request another review in 4 days and 6 hours by commenting @sourcery-ai review. Upgrade to get a review now.

@sourcery-ai

sourcery-ai Bot commented Oct 5, 2026

Copy link
Copy Markdown

Reviewer's Guide

This bookkeeping PR completes T089’s phase-3 checkpoint by recording reproducible passes for all named checks at the pinned commits, closes the F4.1 and F5.2 checkpoint boxes, and carries T092’s three manifest notes for the eleven reaches closed in phase 3, with documentation links and integrity evidence.

Entity relationship diagram for phase 3 carve notes

erDiagram
    PHASE3_CHECKPOINT ||--o{ CARVE_NOTE : carries
    CARVE_NOTE }o--|| MANIFEST_ENTRY : annotates
    CARVE_NOTE }o--o{ CLOSED_REACH : documents
    PHASE3_CHECKPOINT {
        string task T089
        string result PASS
    }
    CARVE_NOTE {
        string status added_or_extended
        string closer T084
    }
    MANIFEST_ENTRY {
        string path
        string edit
    }
    CLOSED_REACH {
        string source_line
        string destination_line
    }
Loading

Flow diagram for the phase 3 checkpoint evidence

flowchart TD
    Pins[Verify pinned commits and clean detached worktrees] --> Checks[Run T089 checks F4.1 F10.1 F13.1 F16.1 F5.2 and F11.1]
    Checks --> Pass[All checks pass]
    Pass --> Close[F4.1 and F5.2 boxes close]
    Pass --> Record[Record quoted results in checkpoint-phase3.md]
    Record --> Links[Link evidence from README.md and plan.md]
    Pass --> Manifest[Carry three T092 notes in carve manifest]
Loading

File-Level Changes

Change Details Files
Adds the phase-3 checkpoint record and links it into the feature documentation.
  • Records pinned commits, hermetic execution conditions, quoted verdicts, and validation details for F4.1, F10.1, F13.1, F16.1, F5.2, and interim F11.1.
  • Ticks T089 and documents closure of F4.1 and F5.2 at the checkpoint.
  • Adds the checkpoint to the evidence tree and README index.
specs/034-opendox-standalone-operation/evidence/checkpoint-phase3.md
specs/034-opendox-standalone-operation/tasks.md
specs/034-opendox-standalone-operation/plan.md
README.md
Carries T092 phase-3 carve notes for the reaches closed by T084.
  • Extends the branch_session manifest note with two phase-3 closures.
  • Adds notes for the two serve_project and seven serve_workbench closed reaches.
  • Documents that notes are limited to declared carve lines and that retired stand-ins receive no notes.
docs/opendox-carve-manifest.yaml
specs/034-opendox-standalone-operation/evidence/checkpoint-phase3.md
specs/034-opendox-standalone-operation/tasks.md
Provides reproducible evidence for the manifest annotations and cross-repository pin/landing relationships.
  • Verifies note-only manifest changes and maps all eleven reaches to unique carve entries.
  • Re-verifies T084, T086, T087, and T094 landings and the relevant pin equalities.
  • Captures clean detached worktrees, scrubbed environments, and successful validation output.
specs/034-opendox-standalone-operation/evidence/checkpoint-phase3.md

Tips and commands

Interacting with Sourcery

  • Trigger a new review: Comment @sourcery-ai review on the pull request.
  • Continue discussions: Reply directly to Sourcery's review comments.
  • Generate a GitHub issue from a review comment: Ask Sourcery to create an
    issue from a review comment by replying to it. You can also reply to a
    review comment with @sourcery-ai issue to create an issue from it.
  • Generate a pull request title: Write @sourcery-ai anywhere in the pull
    request title to generate a title at any time. You can also comment
    @sourcery-ai title on the pull request to (re-)generate the title at any time.
  • Generate a pull request summary: Write @sourcery-ai summary anywhere in
    the pull request body to generate a PR summary at any time exactly where you
    want it. You can also comment @sourcery-ai summary on the pull request to
    (re-)generate the summary at any time.
  • Generate reviewer's guide: Comment @sourcery-ai guide on the pull
    request to (re-)generate the reviewer's guide at any time.
  • Resolve all Sourcery comments: Comment @sourcery-ai resolve on the
    pull request to resolve all Sourcery comments. Useful if you've already
    addressed all the comments and don't want to see them anymore.
  • Dismiss all Sourcery reviews: Comment @sourcery-ai dismiss on the pull
    request to dismiss all existing Sourcery reviews. Especially useful if you
    want to start fresh with a new review - don't forget to comment
    @sourcery-ai review to trigger a new review!

Customizing Your Experience

Access your dashboard to:

  • Enable or disable review features such as the Sourcery-generated pull request
    summary, the reviewer's guide, and others.
  • Change the review language.
  • Add, remove or edit custom review instructions.
  • Adjust other review settings.

Getting Help

Copilot AI left a comment

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Copilot review overview

🔵 Needs a closer look

The documentation is internally consistent, but its approval depends on extensive cross-repository execution evidence that cannot be independently reproduced from this snapshot.

Review effort: Balanced
Findings: None

What changed in this PR

Records phase 3’s T089 checkpoint and T092 carve-manifest notes for feature 034.

Changes:

  • Adds quoted checkpoint evidence for all required phase-3 checks.
  • Marks T089 complete and updates planning references.
  • Annotates the three manifest rows covering eleven closed reaches.
File Description
README.md Links the phase-3 checkpoint.
docs/​opendox-carve-manifest.yaml Adds phase-3 closure notes.
specs/​034-opendox-standalone-operation/​evidence/​checkpoint-phase3.md Records checkpoint commands, outputs, pins, and note validation.
specs/​034-opendox-standalone-operation/​plan.md Adds the checkpoint to the evidence tree.
specs/​034-opendox-standalone-operation/​tasks.md Completes T089 and records T092’s phase-3 notes.

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.

@sonarqubecloud

sonarqubecloud Bot commented Oct 5, 2026

Copy link
Copy Markdown

@brettheap

Copy link
Copy Markdown
Contributor Author

READY at 8bd4d12 — Lane: openxfactory-4 (openXfactory-4-openDox_extraction)

Holder: lane openXfactory-3's D3, the second half, T089: phase 3's checkpoint, quoted, with T092's phase-3 notes. It is bookkeeping and evidence, with no Arc: trailer (R1Q20 (a)).

  • The checkpoint, run C at openDox-code dede32b4, openXdox-code 56e1c238 and openxFactory 36908480. All six pass:
    • F4.1: no deferred reach names the consumer or the publisher;
    • F10.1: 0 violations;
    • F13.1: the migrations apply;
    • F16.1: passes, with batches M and P's cases;
    • F5.2: the seven suites pass, and its 5 protected edits each hold;
    • T098's interim F11.1: requirement 1 holds.
  • The closers, re-verified against the real landings: T084 e49b17c3, T086 56e1c238, T087 e1e3a3c3/dede32b4, and T094 T094: phase 3's consumer pins, openDox e1e3a3c3 and openXdox 9564d5d9, with phase 3's host wiring: openXdox's columns and the governed binding-trust policy (plan 034) #1236 36908480.
  • T092's phase-3 notes follow the default patch, ruling (A) 5994463071: 3 notes on 3 rows (2 added, 1 extended), recorded beside phase 2's decision (b). The content rule holds, and with the notes stripped the manifest is byte-identical. validate-carve-manifest is OK. T092 stays unticked until the arc closes.
  • Gates.
    • All 15 checks pass.
    • Copilot finds nothing ("Needs a closer look", no findings, 0 threads).
    • There is no closing keyword.
  • What follows. T095 (openDox-code#75) lands next. Its landing is X for lane openXfactory-3's T096.

@brettheap

Copy link
Copy Markdown
Contributor Author

Lane: openxfactory-4 (openXfactory-4-openDox_extraction)

LANDING — lane openxfactory-4, session faabad8b-8c31-4958-abd5-b35bdc9cd282@Eagle, 2026-10-05T15:26:31Z, PR #1238 into opensoft/openxFactory main

@brettheap
brettheap merged commit fab575a into main Oct 5, 2026
16 checks passed
@brettheap

Copy link
Copy Markdown
Contributor Author

Lane: openxfactory-4 (openXfactory-4-openDox_extraction)

LANDED — lane openxfactory-4, 2026-10-05T15:26:56Z, PR #1238 → fab575a (opensoft/openxFactory main; squash)

brettheap added a commit that referenced this pull request Oct 5, 2026
…h T095's acceptance job; tick T096 (plan 034) (#1241)

Lane: openxfactory-4 (openXfactory-4-openDox_extraction)
Authored by lane openXfactory-3, delegated by lane openxfactory-4 (#656 5987202449)

**Plan 034 T096: AT-R1's browser half on the host, at `RELEASE1_TIP`, with the HTTP half beside it and T095's `acceptance` job.**

- X, `RELEASE1_TIP`, is openDox-code `389e5a4a22e1778653dac9f17172de2c18f9f808`, T095's landing (openDox-code#75).
- P, T087's pin, is `dede32b4b6f3d0f147d599776f83628c5af8ff3d`.
- T096's falsifier is *"the oracle's verdict, which must pass: zero `pageerror`, and nothing undeclared"*. Its record is *"this run's oracle verdict, and the URL and verdict of T095's `acceptance` job at the same openDox-code commit"*, with X's full sha and P beside it.

Housekeeping:

- This PR is bookkeeping, so it carries no `Arc:` trailer (R1Q20 (a), `5817152735`; T091). Claim: #656 comment `5987202449` (slice D6, delegated by lane openxfactory-4).
- **The form is T083's (#1231):** an evidence record in `evidence/`, the task ticked with a `Landed` bullet, plan.md's evidence inventory, and the feature's README entry linking the record.
- **Based on `main` `a43845e6`**, which holds T089's landing (#1238 → `fab575ad`). T089 landed first, on the holder's ruling `5987619278` (Q3). Commits: `f632d70c` (the record and the tick), then review rounds 1 to 4 (`95271efd`, `1df862ac`, `6f0f5381`, `318d99c2`), and `5c663a7e`, a merge of `main` `f2acc88c` (Q-GRC-4 bookkeeping, #1240, which touches no file of this PR). If `main` moves, I merge it in, with no rebase and no force-push.

**No Rule 6 window.** This PR touches nothing under `openspec/changes/`.

## Verdict

| # | the run | where | result, quoted |
|---|---|---|---|
| 1 | AT-R1's browser half, (a) `plain-documents` and (b) three notes with no front matter | openDox-code X `389e5a4a`, a fresh clone, quickstart.md § 1's install under the commit's lock, Playwright 1.61.0 with Chromium 149.0.7827.55 | **PASS**, `AT-R1 browser half at 389e5a4a…: PASS ((a) PASS; (b) PASS)` |
| 2 | AT-R1's HTTP half, T095's harness on the host | the same commit, its own fresh clone and venv | **PASS**, `314 assertions held` |
| 3 | T095's `acceptance` job: a `workflow_dispatch` of `validate.yml` on `main` while `main` was X (`5987619278`, Q1) | [run 37333323932](https://github.com/opensoft/openDox-code/actions/runs/37333323932), `head_sha` `389e5a4a22e1778653dac9f17172de2c18f9f808` (= X), [job 111841708700](https://github.com/opensoft/openDox-code/actions/runs/37333323932/job/111841708700) | **success**, *"AT-R1 HTTP half: PASS (314 assertions held)"* |
| 4 | T099's P-against-X step | `git diff --quiet dede32b4 389e5a4a -- src/ pyproject.toml migrations/ README.md LICENSE` | exit **0**, so X's runs stand for P: nothing ran at P, and no temporary branch was made. Lane openxfactory-4's own run of the step (15:29:25Z) also exited 0. |

The dispatch's gate read `main` as X at 15:30:06Z (`gh api repos/opensoft/openDox-code/commits/main`), and the dispatch followed at 15:30:07Z. The `acceptance` job fetched `+389e5a4a…:refs/remotes/origin/main` and checked out `main`. The run's other job, `validate` (the whole suite), is not part of T096's record.

T096's list, item by item (both repositories unless named):

- **The wheel** renders a tile for every station the snapshot fills: (a) 8 / 2 / 1 / 1 / 1 / 1; (b) 3 / 2, plus 2 derived "Candidate:" tiles.
- **The lens** renders the radar with the documents as dots, 8 of 8 and 3 of 3, with no "nothing on the radar", and offers neither seed action (R1Q19 (a)).
- **The workbench** opens from a grouping tile's `workbench` verb. Before any turn, its chat rail shows *"No model configured. To configure one, declare a model binding with "opendox model-binding add" …"*.
- **A turn** is refused `model_capability_unavailable` (HTTP 403), and the rail withholds Send, naming why.
- **The editors:** both open and accept edits. Create is refused by name: no create control is offered, and the plane note names creating a document and Save. Save is refused by name with `GATELESS_SAVE_REFUSAL`, and nothing is written.
- **The oracle:** zero `pageerror`, nothing undeclared, and no 5xx. Each of the three declarations is satisfied exactly: `/views/intent-feed.js`, `/snapshot-index.json` and `/project-register.json`, *"4 observed / 4 declared across 3 declaration(s)"*.

## What this PR records, in plan 034 only

- **`evidence/at-r1/`** (new):
  - `oracle-verdict.md`, the record;
  - `run-at-r1.sh` and `at_r1_browser.py`, the runbook and the driver, each as run but for its review fixes (three in the runbook, one in the driver). The record names the as-run sha256 of each, which this PR's first commit holds, beside the current one;
  - `x/`, the run at X. It holds the verdict lines, the commit and the install, `acceptance-job.txt`, the HTTP half's log and the resolved dependency set. For each repository it also holds the driver's log (the oracle's printed verdict), every check with its evidence, and the screenshots of steps 1–4;
  - `SHA256SUMS`, covering every file of the run.

  Every host path in the run's files is replaced by a placeholder, so no committed file names a host path (Principle IV). Byte-identical screenshots are stored once, and `screenshots.txt` maps every screenshot name to its stored file. No file holds the console token's value: each pass scans its own record for it.
- **`tasks.md`**: T096 is ticked, with a `**Landed**` bullet. 86 of 96 boxes are ticked on `main`, and 87 here.
- **`plan.md`**: the evidence tree lists `at-r1/`, and the paragraph under it names T096.
- **`README.md`**: the feature entry's evidence list links the record, after T089's.

## Ruled before this PR (the holder, `5987619278`)

1. **Q1, T095's job "at X".** It is a `workflow_dispatch` of `validate.yml` on `main` while `main` was X. The run's URL, its `head_sha` (equal to X) and the `acceptance` verdict are recorded. #75's last PR run does not stand in.
2. **Q2, "a refusal declared to the oracle".** The record carries a non-normative note: the create and Save refusals are refused by name, with no browser signal. Save's click sends no request, and create offers no control, so nothing is undeclared. spec.md and the plan are not reworded, and T097 lists the note with research R16's corrections.
3. **Q3.** T089 lands first, and this PR is based on `main` after it.

Two observations, unchanged since t096-dry, move no verdict. In (b), the wheel adds 2 derived "Candidate:" tiles while the candidate station is unfilled. Save's refusal note reads "Outline: Save refused" even when another buffer is active.

## Review round 1 (Copilot at `f632d70c`): one finding, accepted and fixed in `95271efd`

- r4185918186 (High): `run-at-r1.sh` documented "a full or short sha, or a ref", but required the checked-out sha to start with the argument, so `run-at-r1.sh main` was refused after a good checkout.
  - The fix: the argument is resolved with `git rev-parse --verify "<arg>^{commit}"`, the checkout must land on that commit, and only an abbreviated sha keeps the prefix check.
  - The test, from a fresh clone: `main`, `389e5a4a` and the full sha all check out, and `nosuchref` is refused by name. The as-run file reproduces the finding.
  - T096's run passed X's full sha, so no recorded step changes. Answered and resolved.

## Review round 2 (Copilot at `95271efd`): Findings None; two previously-missed items

- **The runbook named a host-absolute path:** the GNU grep it called, against the constitution's Principle IV, which says "use repo-relative paths or runtime resolution". A comment also spelled the process table's path. Fixed in `1df862ac`.
  - GNU grep is now the first `grep` on the PATH that reports itself as GNU grep.
  - The comment now says what it means without the path.
  - The browser half, run again at X with that file, gives the same check results.
  - The absolute paths that remain are the `env` shebang, the null device, and two libpq socket locations in the harness's own log, quoted verbatim as its assertion ids. None of them is a path of this host.
- **T095's box is still open while T096 is ticked.** T095 is lane openxfactory-4's task, and its tick is that lane's bookkeeping, as #1235 ticked T076, T086, T087, T101 and T104. This PR is delegated for T096 only and leaves T095's box alone.

## Review round 3 (Copilot at `1df862ac`): one finding, accepted and fixed in `6f0f5381`

- r4186065384: where the commit supplies its own harness, the runbook recorded the checkout's status after the harness but never refused a modified one.
  - It now stops when that status is not empty.
  - The run's `x/tree-after-harness.txt` is empty, and the HTTP half, run again at X, passes with `314 assertions held`.
  - Answered and resolved.

## Review round 4 (Copilot at `6f0f5381`): Findings None; one previously-missed item, fixed in `318d99c2`

- The editors check computed whether the Outline buffer was typed into, but never required it. It now requires the Outline editor as well as a document editor.
- The run's recorded results meet the stricter check (`outline_typed` is true in both repositories), and the browser half, run again at X with that driver, passes 18 of 18 checks in both.

## Review round 5 (Copilot at `5c663a7e`): two findings, answered and resolved with no push

Neither finding touches the record's correctness, so this head stays as it is.

- r4186202361 (Medium): `git diff --quiet` exits above 1 on an error, and the runbook records that status rather than refusing it.
  - In T096's run, both commits were present and the step exited 0 (`x/commit.txt`).
  - The wrapper that made the run checks both commits before the diff. It treats any non-zero status, an error included, as "build inputs differ", which re-runs both halves at P.
  - Answered (4186321172) and resolved.
- r4186202500 (Low): the runbook leaves its scratch directory.
  - `OUT` defaults to a directory inside it, so removing it would remove the record.
  - The teardown the record asserts is R1Q16 (iv)'s, not the scratch directory's removal.
  - Answered (4186321635) and resolved.
- The review's overview also says this PR "commits unsanitized host paths", with no finding behind it. A scan of every added text line at this head finds no host path. The only system-rooted paths are the four that round 2 names: the `env` shebang, the null device, and the two libpq socket locations.

## Validation, at this head (`5c663a7e`) against `main` `f2acc88c`

- **`python3 scripts/validate-openspec-cli-pin.py --all --strict`** (pinned CLI 1.12.0): rc 0, `Totals: 110 passed, 1 failed (111 items)`, `0 UNDISPOSITIONED failures`. The one failure is the accepted `add-chain-attestation` exception, as on `main`.
- **doc-health `--single-repo . --as-of 2026-10-05`**, in one clone named `openxFactory`, at `main` `f2acc88c` and at this branch, and at each earlier head against `main` `a43845e6`: **delta zero**. Both read `31 critical, 22 error, 69 warning, 20 info`, and the report bodies are byte-identical.
- **Tests that read plan 034's files:** none. `git grep -nE 'specs/034|opendox-standalone-operation' -- tests scripts .github` finds nothing, and this PR touches no code.
- **Shas:** every sha the diff cites resolves, 18 of 18: 16 in openDox-code, and 2 (`504324de`, `e1e3a3c3`) in opensoft/openDox.
- **Scans:** a scan of the commit message and this body finds no closing keyword and no `Arc:` line, and a scan of the diff finds no host path and no token value.
- **The required checks** at this head, read from its check-runs: all eight succeed.
- **SonarCloud** (not a required check): its quality gate fails on new code, at Reliability D and Security C. Its 41 issues are on the runbook and the driver:
  - 36 are shell style ("use `[[`").
  - 3 are "path traversal via CLI arguments", on the driver's own file arguments, which the runbook supplies.
  - 1 is an `assert` inside a `try` whose handler exits 2 with no verdict. The runbook records that as "the driver reached no verdict", so a failed assertion can never pass.
  - 1 is `STEP[0]`, a step label, written twice in a row.
  - None changes a recorded result, so the two files stay as run but for their review fixes.

🤖 Generated with [Claude Code](https://claude.com/claude-code)


Lane: openxfactory-4
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
brettheap added a commit that referenced this pull request Oct 5, 2026
Lane: openxfactory-4 (openXfactory-4-openDox_extraction)
Authored by lane openXfactory-3, delegated by lane openxfactory-4 (#656 5984809190)

**Plan 034 T097, release 1's bookkeeping.** This PR ticks all 63 release-1 boxes of #1144's `tasks.md` (`add-neutral-product-standalone-operability`) that a realization task closes, each with its evidence note, and records research R16's non-normative corrections. Its After set has landed: batch P (#1230), T096 (#1241 → `8a37a087`) and T099 (the publish at the cut, and openDox#19 → `d77f8cbf`). On the holder's rulings it also ticks plan 034's T095, T099, T002, T007 and T097, each with its bullet, so plan 034 closes at 92 of its 96 tasks. T090 to T093 tick at the arc's close.

- Bookkeeping, so it carries no `Arc:` trailer (R1Q20 (a), `5817152735`; T091) and no closing keyword. It touches `openspec/changes/`, so lane openxfactory-4 lands it under a Rule 6 `LANDING`/`LANDED` window.
- **Based on `main` `8a37a087`** (#1241, T096), merged in by `0cde3961`. The PR was first based on `c8d80020` (#1232). The merges since are listed below. Its commits:
  - `76d8ff37`: the ticks, the notes and R16;
  - `b7f49673` and `e2efb3a1`: from Copilot's review (below), which edit note text only;
  - `bd32ea9c`: the merge of `main` after batch P landed;
  - `cf2c6bdc`: 16.3a's tick, on the T100 follow-on's landing (openDox-code#86 → `651c35fe`);
  - `5870cd00`: 4.3's and 9.2's ticks, on T086's landing (openXdox-code#37 → `56e1c238`);
  - `23665543`: 5.5's note reworded (Copilot at `5870cd00`), note text only;
  - `96c74eab`: the merge of `main` after T094 and T098 landed;
  - `935164af`: 3.3's tick, on T098's landing (openxFactory#1237 → `20ce593e`);
  - `e93ab9fb`: 5.4a's note reworded (Copilot at `935164af`), note text only;
  - `f6c6dc3c`: the merge of `main` after T089 landed;
  - `e873b25f`: F4.1's, F5.2's and F16.1's ticks, on T089's landing (openxFactory#1238 → `fab575ad`);
  - `9759313b`: the merge of `main` `f2acc88c`;
  - `10b6ba05`: the phase-1 and phase-2 checkpoint citations by path (below), note text only;
  - `0cde3961`: the merge of `main` after T096 landed;
  - `f94be2fd`: 10.3's tick, on T099's publish, and plan 034's ticks (below);
  - `9713f991`: plan 034's opening status, which still said batch P was to land (Copilot at `f94be2fd`, below), plan text only.

  If `main` moves before the landing, I merge it in. No rebase and no force-push.

## Measured, not copied

Lane 4's table (2026-10-03T20:05Z, at `de2ab703`: 52 CLOSED-LANDED, 11 PENDING, 5 LEFT-OPEN, 2 ALREADY-TICKED) was regenerated by its own generator, patched, at `c8d80020`, at `ba6bb870`, and again after each closing landing:

- every landing PR, now 115 across the six repositories, re-read with `gh pr view <n> -R opensoft/<repo> --json state,mergeCommit,mergedAt`, last at 2026-10-05T16:50Z. Since the first reading these have landed: #1230 (batch P, `ba6bb870`), openDox-code#84 (T104, which realizes no box, `32943cbf`), openDox-code#86 (the T100 follow-on, `651c35fe`), openDox-code#78 and openDox-code#79 (T101, `d59f3f26` and `dede32b4`), openDox#17 (T076, `504324de`), openDox#18 (T087, `e1e3a3c3`), openXdox-code#37 (T086, `56e1c238`), openXdox#22 (T094's openXdox root step, `9564d5d9`), #1234 (T094's ahead PR, `ca1c1486`), #1235 (phase 3's plan ticks, `70602a0a`), #1236 (T094, `36908480`), #1237 (T094/T098, `20ce593e`), #1238 (T089, `fab575ad`), openDox-code#75 (T095, which realizes no box, `389e5a4a`) and openDox#19 (T099's root README, `d77f8cbf`);
- every task's landing re-read from plan 034's own `Landed:` bullet (78 of 96 tasks ticked at `c8d80020` and `ba6bb870`, 85 at `20ce593e`, 86 at `a43845e6`, 87 at `8a37a087`), and each run record traced to the openxFactory PR that added it: `checkpoint-phase1.md` (#1204 → `9d2e5bc3`), `checkpoint-phase2.md` (#1218 → `a883bbf6`), `f10.1-run.md` (#1223 → `f261fefa`), `f13.1-run.md` (#1224 → `de2ab703`), `f16.1-run.md` (#1231 → `0e01ca85`), `f11.1-phase3.txt` (#1237 → `20ce593e`), `checkpoint-phase3.md` (#1238 → `fab575ad`) and `at-r1/oracle-verdict.md` (#1241 → `8a37a087`). The T100 follow-on, which is no task, is read from the forge;
- batch P's text read at #1230's heads `84113a44` and `d08249a2`, and as it landed (`ba6bb870`, the squash of its head `051f9945`).

**Result: 63 CLOSED-LANDED, 0 PENDING, 5 LEFT-OPEN and 2 ALREADY-TICKED, of 70.** Since lane 4's table:

- T082 (openDox-code#76 → `ca9e1bd5`), T100 (openDox-code#82 → `38d3350e`), T102 (openDox-code#81 → `0116293a`) and T083 (openxFactory#1231 → `0e01ca85`) have landed. That closes 16.4, 16.5 and 16.6.
- The T100 follow-on (openDox-code#86 → `651c35fe`) has landed. That closes 16.3a.
- T086 (openXdox-code#37 → `56e1c238`) has landed. That closes 4.3 and 9.2.
- T098 (openxFactory#1237 → `20ce593e`), phase 3's interim F11.1, has landed. That closes 3.3 (below).
- T089 (openxFactory#1238 → `fab575ad`), phase 3's checkpoint, has landed. That closes F4.1, F5.2 and F16.1 (below).
- T076 (openDox#17 → `504324de`) and T101 (openDox-code#78 → `d59f3f26`, openDox-code#79 → `dede32b4`) have landed.
- T095 (openDox-code#75 → `389e5a4a`) and T096 (#1241 → `8a37a087`), AT-R1's two halves, have landed. They close no box of #1144.
- T099 has published `opendox` 0.1.0, and its root README PR has landed (openDox#19 → `d77f8cbf`). That closes 10.3 (below), the last box.

Batch P closes and rewords no box. In #1144's `tasks.md` it only inserts addenda at the ends of 12.4a, 16.3a and F16.1: no box line moves and no box changes state. Since `ba6bb870`, no landing has touched #1144's `tasks.md`. In plan 034's `tasks.md`, T097's own text is unchanged.

## The census

`python3 box_census.py openspec/changes/add-neutral-product-standalone-operability/tasks.md`, with the tool written from research.md § Appendix (byte-identical to the Appendix's heredoc).

Before, at `main` `8a37a087` (identical to `c8d80020`'s, `ba6bb870`'s, `20ce593e`'s and `a43845e6`'s):

```
2 8 2.1[ ] 2.1a[ ] 2.2[ ] 2.3[ ] 2.4[ ] 2.5[ ] 2.6[ ] F2.1[ ]
3 5 3.0[x] 3.1[ ] 3.2[ ] 3.3[ ] F3.1[ ]
4 5 4.1[ ] 4.1a[ ] 4.2[ ] 4.3[ ] F4.1[ ]
5 12 5.0[ ] 5.1[ ] 5.2[ ] 5.3[ ] 5.3a[ ] F5.1[ ] 5.4[ ] 5.4a[ ] F5.2[ ] 5.5[ ] 5.6[x] F5.3[ ]
7 8 7.0[ ] 7.1[ ] 7.1b[ ] 7.1a[ ] 7.2[ ] 7.3[ ] F7.1[ ] F7.2[ ]
9 8 9.1[ ] 9.2[ ] 9.2a[ ] 9.3[ ] 9.4[ ] 9.5[ ] F9.1[ ] F9.2[ ]
10 5 10.1[ ] 10.2[ ] 10.2a[ ] 10.3[ ] F10.1[ ]
11 3 11.0[ ] 11.1[ ] F11.1[ ]
13 8 13.1[ ] 13.2[ ] 13.3[ ] 13.4[ ] 13.4a[ ] 13.5[ ] 13.6[ ] F13.1[ ]
16 8 16.1[ ] 16.2[ ] 16.3[ ] 16.3a[ ] 16.4[ ] 16.5[ ] 16.6[ ] F16.1[ ]
total 125 Counter({' ': 105, 'x': 11, '~': 9})
```

After, at this head `9713f991` (`9713f991` changes only plan 034, so the census is `f94be2fd`'s):

```
2 8 2.1[x] 2.1a[x] 2.2[x] 2.3[x] 2.4[x] 2.5[x] 2.6[x] F2.1[x]
3 5 3.0[x] 3.1[x] 3.2[x] 3.3[x] F3.1[x]
4 5 4.1[x] 4.1a[x] 4.2[x] 4.3[x] F4.1[x]
5 12 5.0[x] 5.1[x] 5.2[x] 5.3[x] 5.3a[x] F5.1[x] 5.4[x] 5.4a[x] F5.2[x] 5.5[x] 5.6[x] F5.3[x]
7 8 7.0[x] 7.1[x] 7.1b[x] 7.1a[x] 7.2[x] 7.3[x] F7.1[x] F7.2[x]
9 8 9.1[x] 9.2[x] 9.2a[x] 9.3[x] 9.4[x] 9.5[ ] F9.1[x] F9.2[ ]
10 5 10.1[x] 10.2[x] 10.2a[x] 10.3[x] F10.1[x]
11 3 11.0[ ] 11.1[ ] F11.1[ ]
13 8 13.1[x] 13.2[x] 13.3[x] 13.4[x] 13.4a[x] 13.5[x] 13.6[x] F13.1[x]
16 8 16.1[x] 16.2[x] 16.3[x] 16.3a[x] 16.4[x] 16.5[x] 16.6[x] F16.1[x]
total 125 Counter({'x': 74, ' ': 42, '~': 9})
```

The release-1 groups are shown. Groups 1, 6, 8, 12, 14, 15 and the follow-ons read the same before and after. Earlier heads read `'x': 66, ' ': 50` (`76d8ff37` to `bd32ea9c`), `'x': 67, ' ': 49` (`cf2c6bdc`), `'x': 69, ' ': 47` (`5870cd00` and `23665543`), `'x': 70, ' ': 46` (`935164af` to `f6c6dc3c`) and `'x': 73, ' ': 43` (`e873b25f` to `0cde3961`).

## The change: #1144's `tasks.md`, and plan 034's ticks

- **63 boxes ticked**, each with its evidence note as the item's last paragraph, after any addenda. The note is in 1.8's and 3.0's form: a bold dated lead, `**Landed <date>**`, where the date is the last landing of the box's realizing tasks; the tasks; the landings as `repo#n → sha8`; and the falsifier's result, with where it was quoted. For example, 2.4:

  > **Landed 2026-09-27** (T030, in T011's PR): openDox-code#46 → `0e88454a`. `tests/test_imports_standalone.py::test_every_module_imports_with_no_sibling` first asserts the four siblings absent, then imports every module, failing with the first that still needs one (`evidence/checkpoint-phase1.md` § 1).

- **No ratified line moves.** Against `main` `8a37a087`, #1144's diff is the 63 checkbox lines, `[ ]` → `[x]`, and 64 inserted paragraphs (482 lines): 63 notes, and 1 R16 note on an open box (9.5). This was proven at `f94be2fd` by setting every checkbox to `[ ]` in both files and diffing: the result is 64 insertions and nothing else, and the lines the two files share differ only by the 63 flips. The only other file the PR changes is plan 034's `tasks.md` (below). The file is byte-identical to a fresh re-apply of the regenerated decisions onto `main`'s file.
- **Research R16** (research.md § R16), each figure re-checked against research R2, R4, R5, R10 and R13 and plan 034's records:

| R16 item | box | where it is recorded |
|---|---|---|
| 1 | 2.5 | in 2.5's note: 22 modules, not 26, were reached by no CI step |
| 2 | 9.4 | in 9.4's note: openXdox-code's floors read `564/558/6` at `626f2c8d`, not `539/533/6` |
| 3 | 4.3 | in 4.3's note, since 4.3 closed on T086 (an open-box note until then): `run_scoped_doc_health` makes three of the four reaches, and the fourth is `session_documents` |
| 4 | 3.2 | in 3.2's note: the proxy's refusal was written for NOTHING REGISTERED |
| 5 | 10.1 | in 10.1's note: Q-R4 is recorded at `runtime/cli.py:39-42`, not `:40-46` |
| 6 | 9.5 | a note on the open box (open for the arc's close, T090): the openDox pin was 11 commits behind, not 9, and T086 has since moved it to `dede32b4` |
| 7 | Group 2's heading | in 2.1's note, the box the heading introduces: 1,305 and 1,239 at `1e4a57fb` |
| 8 | 3.0 | no new line: 3.0's own note already records it (*"Read at ratification, 2026-09-24"*) |

## 10.3, closed on T099

10.3's box says the openDox root documents the entry point and does not host it, and batch H's addendum makes the documented command `pip install "opendox[local]"`, then `opendox generate-and-open --local …`. Plan 034 maps it to T076, T101 and T099:
- **T076** (openDox#17 → `504324de`) documents the install and the start, and adds no `make` target.
- **T101** adds the release workflow and the version bump (openDox-code#78 → `d59f3f26`, openDox-code#79 → `dede32b4`).
- **T099** publishes that commit at the cut, on Brett Heap's word (`5997633384`). Read from the forge and from PyPI:
  - the annotated tag `v0.1.0` (`9153c254`) names `dede32b4`, which the openDox root's `contracts/code-pin.yaml` names;
  - openDox-code's `release` run 37339111713, dispatched on that tag, completed `success` in all four jobs: `build and verify`, `publish to TestPyPI (the dry run)`, `install opendox[local] from TestPyPI` and `publish to PyPI`, whose steps include `PyPI serves the files the build job verified`;
  - `https://pypi.org/pypi/opendox/0.1.0/json` lists exactly two files, neither yanked: `opendox-0.1.0-py3-none-any.whl` (sha256 `8ecea00db6f9…`) and `opendox-0.1.0.tar.gz` (sha256 `56869b6208a8…`);
  - openDox#19 → `d77f8cbf` replaces the README's stand-in paragraph. At that commit, the README's first line resolves from PyPI, and no paragraph says that no release is published.

## Plan 034: T095, T099, T002, T007 and T097

Lane openxfactory-4's delegation of 2026-10-05T16:37:31Z asks #1233 to tick these, each with a `Landed` or `Done` bullet in #1226's form, so that release 1 closes at 92 of 96. Its ruling `5999175092` settles three points: (a), (a) and (a). Every sha in the bullets was read from GitHub.
- **T095**: openDox-code#75 → `389e5a4a`, which is X. Its `acceptance` job at X is the dispatched run 37333323932: `head_sha` X, job 111841708700, `success`, and its log reads `AT-R1 HTTP half: PASS (314 assertions held)`.
- **T099**: P against X. The compare shows four files between them, none a build input, so the ruled check exits 0. Then the tag, the release run, PyPI's two files and openDox#19, as above.
- **T002** (ruling (a)): 45 of the 50 slices in the three phases' writer tables were claimed on #656, and the bullet cites each claim. The five that were not are named in T091's form: P1-C, whose claim `5850230046` was posted as a literal file path and never edited, and P1-L, P2-H, P2-L and P2-M. Nothing is back-filled.
- **T007** (ruling (a)): batches A to P have landed (P #1230 → `ba6bb870`), and its status sentence now says so.
- **T097**: the census and research R16's corrections, with two non-normative notes beside them:
  - D6's T096 refusals, *"refused by name, with no browser signal"*;
  - the reachability of #1241's first commit, `f632d70c`, which `oracle-verdict.md` names as holding the bytes that ran (ruling (a)'s wording). It is not on `main`: `refs/pull/1241/head` (`5c663a7e`) reaches it, and so does #1241's branch, which still stands. `main`'s `SHA256SUMS` fixes the reviewed scripts, and the record names the ran bytes by their sha256.

## The checkpoint citations, by path

Lane openxfactory-4 ruled at 2026-10-05T16:25:25Z that the notes' `CP1 § n` and `CP2 § n` citations name their records, as the phase-3 notes name `evidence/checkpoint-phase3.md`. `CP1` and `CP2` are the shorthand of lane 4's table, and #1144 never defines them. The change is non-normative and inside T097's R16 scope. `10b6ba05` makes it. The 26 citations, in 25 notes, now read `evidence/checkpoint-phase1.md` (T049, #1204) and `evidence/checkpoint-phase2.md` (T063, #1218). A word diff against `e873b25f` finds those 26 tokens and nothing else; only the 25 paragraphs' line breaks move.

## F4.1, F5.2 and F16.1, closed on T089

T089, phase 3's checkpoint, landed on 2026-10-05T15:26:43Z as #1238 → `fab575ad`. Its record, `specs/034-opendox-standalone-operation/evidence/checkpoint-phase3.md`, extracts each block from #1144 by anchor at `36908480` and runs it from a fresh worktree at the commits phase 3 pinned: openDox-code `dede32b4`, the `code` of the openDox root `e1e3a3c3` (T087), and openXdox-code `56e1c238` (T086). It says of F4.1 and F5.2 *"T097 ticks both in #1144; this record ticks nothing there"*, and plan 034's T089 `Run` bullet says the same. Each box closes on that run, by plan 034's Box accounting:

- **F4.1** (`F4.1 → T089`): as extracted (sha256 `8900b985e72f`) and run unchanged, it exits 0 with `1 passed`, `1 passed`, then `no deferred reach names the consumer or the publisher` (§ 1).
- **F5.2** (`F5.2 → T086, T089`; RULED `5962785556`, item 1, moves its close to phase 3): whole, as batches C, F, G and K amend it (`OPENXFACTORY` at `36908480`, last step `--chains`), it exits 0. The seven suites pass whole (45, 23, 18, 6, 40, 9 and 12 passed), `test -s` holds, and the last step prints five `admitted:` lines, then `ok: 5 protected edit(s), each entered and holding` (§ 5). One of the five admits T086's edits to `tests/test_session_snapshot.py`, by entries 16, 17 and 19 of `tests/protected_suite_respellings.yaml` at `56e1c238`: the three pre-arc reds that T063 quoted, each entry naming the test it repairs.
- **F16.1** (`F16.1 → T083 (at 38d3350e), T089 (batch P's cases, at T087's pin)`): with batch M's line, it exits 0 with `24 passed`, the dialect line, `no model configured: the catalog offers nothing`, `83 passed` and `532 passed` (§ 4). `dede32b4` carries the T100 follow-on (openDox-code#86 → `651c35fe` is its ancestor, and the only commit between T100's landing and `dede32b4` that touches `tests/test_model_binding_trust.py`). That file is the same at both: 160 tests, among them batch P's six `test_F16_1_batch_p_*` cases and A8's `test_A8_a_state_directory_that_is_a_link_trusts_nothing`, none of which exists at `38d3350e`. T083's run there is quoted beside it, as before.

## 3.3, closed on T098

3.3 asks that the carve manifest's `deleted_at_carve` row for `scripts/ideation_dashboard/profile_openxfactory.py` stay BYTE-IDENTICAL. T017, which realizes it, reads the row through the three interim F11.1 runs, *"and T097 ticks 3.3 on all three"*. Each was re-checked on `main` `20ce593e`:

- **Phase 1**, T018 (#1202 → `e81eed62`, `evidence/f11.1-phase1.txt`): the amended run at T047's landing prints `requirement 1 holds: 0 note(s) annotated, every other path a declared surface (11.1)`, and the record's T017 section reads the row.
- **Phase 2**, T065 (#1217 → `1f670bc3`, `evidence/f11.1-phase2.txt`): the same line, at T064's landing.
- **Phase 3**, T098 (#1237 → `20ce593e`, `evidence/f11.1-phase3.txt`), at ARC_TIP `36908480` (T094, #1236). It prints the same line (exit 0) over the arc's three landings, and it refuses a planted path. `docs/opendox-carve-manifest.yaml` is touched by none of the three. Its T017 section finds the row equal at the manifest's introducing commit `17167481` and at that tip, both raw and parsed: `not_moved`, `deleted_at_carve`, no `edits`.
- **Re-read for this tick at `main` `20ce593e`:** the row is still equal to `17167481`'s, both raw (4 lines) and parsed. `scripts/opendox_host.py` is still openxFactory's host: T094 modified it and deleted nothing, which F11.1 would have refused.

## 4.3 and 9.2, closed on T086

T086 landed on 2026-10-05T12:23:03Z as openXdox-code#37 → `56e1c238`. It is a squash whose tree (`78ef1ed1`) is that of its head `57bebd8b`. Its pin is openDox-code `dede32b4`, the commit T087 pins (openDox#18 → `e1e3a3c3`).

- **4.3** closes *"with the last of"* the nineteen reaches into openXdox, and *"openXdox-code's ratchet (`OPENDOX_BACK_IMPORTS`) is tightened to `(0, 0)` in the same landing"*. At `56e1c238`:
  - `tests/test_dependency_direction.py` holds `OPENDOX_BACK_IMPORTS = {}`, and `test_the_pinned_opendox_does_not_import_openxdox_back` asserts `(0, 0)`.
  - openXdox-code#37's body reads that file `17 passed` at the pin, with the census empty.
  - `consumer_reach.py`, absent at `dede32b4`, was retired by T084 (openDox-code#77 → `e49b17c3`).
- **9.2** closes in phase 3 (batch F's addendum), with batch B's note. At `56e1c238`:
  - `tests/declared_exclusion.yaml` holds 66 entries carrying 68 reasons: 60 `doc_health`, 3 `status-exemption-rail` and 5 `openxfactory-contracts`.
  - CI's `validate` at the PR's head (openXdox-code run 37307785866) printed `open extraction: the declared exclusion` and `triple: selected=1099 passed=1095 skipped=4 failures=0 errors=0`.
  - Requirement 9 stays open until T008.

## 16.3a, closed on the T100 follow-on

16.3a is *"Carried out by T100"*, and batch P's addendum adds *"Carried out by a T100 follow-on openDox-code PR (claim `5982447319`), which is no task of plan 034 and lands before T087"*:

- T100 landed as openDox-code#82 → `38d3350e`.
- The follow-on landed as openDox-code#86 → `651c35fe` (09:49:55Z). It is a squash whose tree is that of its head `aecac805`, and it landed before T087 (openDox#18, 11:58:27Z) and T101.
- It carries out Brett Heap's `5982436447`, item 2, and `5983805990`, with the holder's `5984069416`, `5985046107` (C1 to C5) and `5985553609`, and the holder's A8.
- Its body records the falsifier: F16.1's block as batch P amends it, run by T089's runner at `aecac805`, exits 0 with `532 passed` in `tests/test_model_binding_trust.py`. The six `test_F16_1_batch_p_*` tests are present at `651c35fe`.
- The open trust-time existence check, and the follow-ons of openDox-code#86's eighth Copilot review, move no box of release 1. 16.3a's own text says so for the first, and the holder's `5992038800` for the second.

## Left open, by T097's own text

- 9.5, 11.0, 11.1 and F11.1, for the arc's close (T090 to T093), and F9.2, until T008. Nothing else is held.

## T097 is ticked

T097's task is *"Tick #1144's release-1 boxes, each with its evidence note: the 63 in the table below."* All 63 are ticked, so `f94be2fd` ticks T097 in the same commit as 10.3, the last box, as lane 4's brief requires.

## The merges of `main`

- **`bd32ea9c`** merged batch P (`ba6bb870`). It gave the one conflict this PR expected, at 16.4's checkbox line, which follows batch P's insertion at the end of 16.3a. The resolution took `main`'s text for every hunk batch P changed and re-applied the ticks and notes with the same tool, byte-identical to a fresh re-apply onto `main`'s file.
- **`96c74eab`** merged `main` `20ce593e` (#1234 to #1237). It was clean, because none of the four touches #1144's `tasks.md`.
- **`f6c6dc3c`** merged `main` `a43845e6` (#1238 and #1239). It was clean too, because neither touches #1144's `tasks.md`.
- **`9759313b`** merged `main` `f2acc88c` (#1240), and **`0cde3961`** merged `main` `8a37a087` (#1241). Both were clean, for the same reason.
- Every tick since is a fresh re-apply onto `main`'s file.

## Gates, at `main` `8a37a087` and at this head `9713f991`, in one full clone named `openxFactory`

| gate | `main` `8a37a087` | this head `9713f991` |
|---|---|---|
| `python3 scripts/validate-openspec-cli-pin.py --all --no-cache` (CI's form; pinned 1.12.0) | rc 0, `110 passed, 1 failed (111 items)`, `0 UNDISPOSITIONED failures`, 1 accepted exception (`add-chain-attestation`) | identical output, but for the pinned CLI's temporary directory |
| `… --change add-neutral-product-standalone-operability` | — | rc 0, `1 passed, 0 failed`, `validated --strict clean` |
| `scripts/doc-health.py --single-repo . --as-of 2026-10-05` | `31 critical, 22 error, 69 warning, 20 info`, 0 new regressions | byte-identical report |
| `proposal-support.py . verify add-neutral-product-standalone-operability` | ok | identical |
| `validate-sequenced-after.py . --ledger-diff` | 229 rows, consistent | identical |
| `validate-code-surface.py`, `validate-target-release.py`, `validate-pin-registrations.py`, `validate-document-catalog.py` | rc 0 each (the catalog: 0 errors, 0 warnings) | identical |
| `pytest -m "not postgres"` over `tests/doc-health`, `sequenced_after`, `proposal-support`, `former_id_arrival`, `citation_remainder`, `signed_execution_chain`, `openspec_cli_pin`, `target_release`, `code_surface`, `packet_reference`, `scope_globs` | the 8 below fail identically | running (the last full run, at `10b6ba05`: `8 failed, 3427 passed, 1 skipped`) |

The 8 failures are environmental, the same 8 at every head. Each needs the `openDox` and `openXdox` legs initialized, as CI's `pytest-suite` does, or a reachable pinned validator: `test_ideation_readiness.py` (3), `test_readiness_dispatch.py` (1), `test_sentinel_vocabulary.py` (2), `test_status_reader_real_lines.py` (1) and `test_chain_reader.py` (1). Re-run alone at `main` `8a37a087`, they read `8 failed` with the same messages. The venv was built by CI's own `pip install --require-hashes -r requirements/hermes-runtime-contracts.lock`, with `LANG=C.UTF-8` and every `GIT_*`/`XF_*` variable unset. The earlier rounds read the same: `main` `c8d80020` against `76d8ff37`, `b7f49673` and `e2efb3a1`, `main` `ba6bb870` against `bd32ea9c`, `cf2c6bdc`, `5870cd00` and `23665543`, `main` `20ce593e` against `935164af`, `main` `a43845e6` against `e873b25f`, and `main` `f2acc88c` against `10b6ba05`. At `f94be2fd` the local run was stopped at 31% on a session swap, and CI's `pytest-suite` passed there. No test, script or workflow reads plan 034's files (`git grep` over `tests`, `scripts` and `.github` finds none).

Also checked, on the inserted text at `9713f991`:

- every `repo#n → sha8` in #1144's notes (71 distinct pairs, 122 mentions) and in plan 034's bullets is that PR's merge commit, per the forge;
- every other sha (`17167481`, `1e4a57fb`, `36908480`, `5c137a90`, `626f2c8d`, `aecac805`, `dede32b4`, `f8a1ece`; in the plan, `5c663a7e`, `9153c254` and `f632d70c`) resolves in its repository, and the file digests match PyPI's JSON and the `SHA256SUMS` files;
- each of the 63 `Landed` dates equals the last landing of the box's realizing tasks (F5.1's and F7.1's notes also name T063's later re-run, and 5.4a's names the host line, #1215);
- no closing keyword, no `Arc:` line, no host path, and no unqualified reference to another repository, in the diff, the commit messages or this body;
- no inserted line is wider than 80 characters, and in #1144 none starts a list, a heading or a table.

## CI and review

- **CI at this head `9713f991`:** running.
- **CI at `f94be2fd`:** every check-run completed `success`, the 8 required among them (`pytest-suite` 17:13:56Z to 17:31:21Z).
- **CI at `10b6ba05`:** every check-run completed `success` (`pytest-suite` until 17:04:09Z).
- **CI at `e873b25f`:** every check-run completed `success`, the 8 required among them (`pytest-suite` 15:57:01Z to 16:22:01Z). `lane-line` re-ran on each edit of this description and passed.
- **CI at `935164af`, `23665543`, `cf2c6bdc`, `bd32ea9c` and `e2efb3a1`:** every check-run completed `success`, the 8 required among them (`signed-execution-chain-gate`, `lane-line`, `former-id-arrival-gate`, `openspec-cli-pin`, `wallet-validation`, `pytest-suite`, `release-tag-gate`, `openxdox-consumer-gate`). At `5870cd00`, `23665543`'s run superseded its `pytest-suite`.
- **Copilot at `76d8ff37`:** one finding, r4179674026: F7.2's note named no landing commit for T058 or T063. `b7f49673` names them (openDox-code#68 → `047bb4fa`, openxFactory#1218 → `a883bbf6`), and 7.1's changelog landing too (openDox#15 → `66758438`). The thread is answered and resolved.
- **Copilot at `b7f49673`:** no new finding, and one "previously missed" note: that 5.4a's note wrongly said T059's run deselected the three pre-arc cases. The note was right, and comment `5985467371` gives the records. The cited lines are T059's Falsifier line, written before it landed; T059's `Landed` record and openXdox-code#35's body (rulings 4 and 6, and its F5.2 run) record the deselections. `e2efb3a1` makes the note exact and cites them.
- **Copilot at `5870cd00`:** *"Findings: None"*, and one "previously missed" low note: 5.5's note read as though `consumer_reach` *"names the task lists"*. `23665543` says what it meant, that the seven `consumer_reach` names T055 lists are retired.
- **Copilot at `e2efb3a1`, `bd32ea9c`, `cf2c6bdc` and `23665543`:** *"Findings: None"*. Where its status line read *"Changes recommended"*, it named only what this DRAFT waits for.
- **Copilot at `935164af`:** one finding, r4185667778: 5.4a's note read *"the seven red at both pins"*. `e93ab9fb` reads "the seven reds", and the thread is answered and resolved.
- **Copilot at `e873b25f`:** one finding, r4186043242: this description was stale after T089's ticks. It is refreshed here, and the thread is answered and resolved. A second request at that head registered no new review.
- **Copilot at `10b6ba05`:** *"Findings: None"*.
- **Copilot at `f94be2fd`:** two findings, review `5418196219`. r4186761333: plan 034's opening status still said batch P was to land, against this PR's T007 tick. `9713f991` says that all sixteen batches have landed, A (#1171) to P (#1230), and that T007 is closed. r4186761413: this description was stale at `e873b25f`. It was refreshed at 17:16:28Z, while that review ran, and is refreshed again here for `9713f991`. Both threads are answered and resolved.
- **Copilot at `9713f991`:** review `5418607294`, on the push. Its one open item was r4186761413, this description, refreshed here for `9713f991`. Its two "previously missed" low notes, at plan 034's lines 129 ("whose message lacks") and 4496 ("the drive"), need no change, because each phrase is the wording of the record it cites, T091's and `oracle-verdict.md`'s. A comment on this PR gives the records.
- Sourcery posted a reviewer's guide, and no comment on the diff. At 17:16:25Z it reported its review budget used, and its check-run reads `skipped` at `f94be2fd` and at `9713f991`.

🤖 Generated with [Claude Code](https://claude.com/claude-code)


Lane: openxfactory-4
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants