§04 · LAB LOG · WHAT ACTUALLY MOVED

The lab, 2026-11-09.

Daily records retain archival work-date labels. Work, preparation, review, and publication dates stay separate.

LAB/2026-11-09 // DATED WORK RECEIPTS

Lab / evidence companion2026-11-09work date

Restore the bridge without hiding the held state

Bridge and UI observed; capture still held

A transient bridge recovered and the control-room UI was installed and observed. Capture remained held, and a separate reboot-identity fix stayed a candidate.

Verification
Agents
Process control

Evidence scope

Retrospective evidence from the dated bridge and Hub acceptance receipts. A recovered service response and an installed interface are separate from boot persistence and capture recovery.

Read original narrative All lab records

At this checkpoint

Run receipts preserved: 80historical receipts
Preservation during the transient bridge recovery.
Hub source files checked: 7matching hashes
Separate actual native UI acceptance followed.

Follow the evidence

Open a stage to inspect what happened.

01Rebuild the boot-specific planreviewed
The previous process identity was not reused as current evidence.
02Restore a bounded bridgerecorded
Health responded and the existing 80 receipts were preserved.
03Install and observe the UIrecorded
The new Work/provider presentation was checked in the real desktop Hub.
04Keep capture heldopen
Restart-state verification remained incomplete after reopen.

Controls & proof2 documented controls

Preserved run historyPASS in retained receipt
The bridge recovery retained all 80 existing receipts.
Installed-versus-live distinctionNative UI observed
Seven source hashes and the visible Work/provider views were checked separately; capture remained held.

A dated evidence view. Publication and review dates remain in the original record below.

2026-11-09

Lab log — 2026-11-09

Retrospective UTC work date 2026-11-09; prepared 2026-11-12 by Codex from retained engineering receipts and separately labelled dated reports. First published 2026-11-12. Source review does not establish current runtime health or owner editorial approval of each claim.

Recovery exposed two different kinds of stale state. A worker identity check still referred to a pre-reboot process, while the Hub bridge needed a fresh launch plan for the current boot. The identity correction was tested in isolation; it was not evidence of a live worker recovery.

The native bridge was restored for that boot. Its local health checks returned 200 and the 80 existing run receipts were preserved. That proved a working response at the checkpoint. It did not prove automatic startup after another reboot or resolve the separate held journal.

The control-room update then reached the actual desktop Hub. Seven installed source hashes matched, and native UI checks showed dated provider evidence, a work preview and clearer status presentation. The eight-file retained journal stayed unchanged.

After reopen, capture was explicitly held because restart state was not verified. The interface improvement was real, but its value depended on keeping that limitation visible instead of turning a successful installation into a claim that the whole system was operational.

What did not ship

No verified capture recovery, genuine collection import, continuously fresh fleet view or accepted live deployment of the separate reboot-identity candidate was established by these receipts.

Verify the held restart state against the current journal before admitting a new observation, and check bridge startup on a separately authorized reboot rather than assuming a transient launch is permanent.

Raw telemetry

Bridge recovery preserved 80 existing run receipts.
Hub 1.16.66: seven source hashes matched and the native Work/provider view was observed.
Eight-file journal remained unchanged; post-reopen capture stayed held.
Separate identity-fix validation: offline candidate evidence, not live acceptance.

← all lab logs

Explore ABL.