Gladex Agent Logs

        FROZEN_ITEMS (111) - the historical snapshot neither grew nor shrank
    R2  the NEWEST `## [x.y.z]` entry owns a `### Queue` section - the pointer
        was not dropped by a run that forgot it
    R3  that newest `### Queue` section contains NO list item of any kind
        (`- `, `* `, `+ `, `1.`) - no second list may be written there
    R4  that newest `### Queue` section names `agent-logs/PROGRESS.md`
    R5  agent-logs/PROGRESS.md carries at least one `Next-candidate queued`
        bullet - the authoritative list still exists
    R6  NO `## [x.y.z]` entry owns more than ONE `### Queue` section - one
        pointer per entry. A duplicate is invisible to R1-R5 by construction:
        it is pointer-only, so the freeze stays 111, and `newest_section` is
        the LAST one, so R2-R4 judge only the copy a run meant to leave. This
        is the defect [0.4.102] carried at EOF for one run - found by counting
        (61 raw headings against 60 read as sections), while this tool exited 0
        and the suite stayed green.
    R7  every `### Queue` heading is accounted for: count_queue_headings() is
        an INDEPENDENT census with its own fence walk, and raw == sections +
        fenced must hold. parse() deciding not to read a heading (a `### Queue`
        written above the first `## [x.y.z]` heading, or a future edit that
        stops appending) is a dropped heading, not a file without one. The two
        readers walk the same file separately precisely so they can disagree;
        one pass reporting both numbers could not.

What R5 COUNTS (it is reported as `progress_bullets`, and until [0.4.80] that
    number was `pg_text.count("Next-candidate queued")` - raw occurrences
    anywhere in the file: measured 56 against 49 real entry lines on the live
    tree when the fix was written, a number that moves whenever an entry
    merely QUOTES the phrase):
    a queue bullet is a line whose first non-space text is the phrase, with or
    without a list marker (`- **Next-candidate queued…**` and the six entries
    written as a bare `**Next-candidate queued…**` paragraph both count, once
    each), outside any fenced block. A line that mentions the phrase in the
    middle of a sentence - the R5 prose itself, a dashboard detail line, a
    health summary - is not an entry, and neither is a second occurrence on a
    line already counted.

R1 is deliberately an EQUALITY against a constant rather than a "greater than":
    a count that may only grow cannot see a history rewrite, and a count
    compared against nothing is the shape of arithmetic this repo has been
    removing since [0.4.65]. Changing FROZEN_ITEMS is a deliberate act - edit
    the constant and say so in the changelog entry that does it.

Fenced code blocks are NOT structure: an entry that quotes the shape of a
    `### Queue` section - as [0.4.67] does - must not gain a section from the
    quotation, and list items inside the fence do not count towards
    FROZEN_ITEMS either.

exit 3 is "cannot verify", never "ok": a sandbox repo with no CHANGELOG.md is
    an environment where the rule was not evaluated, and reporting that as a
    pass is how a check stops being a check.

`detail` in the JSON output never contains a `"` or a `\\`, so a shell caller
    can embed it in its own JSON object without re-escaping it.

Usage:
    queue-source-check                 # human output against the repo it lives in
    queue-source-check --format json   # one JSON object, machine-readable
    queue-source-check --repo /tmp/sb  # check a sandbox copy instead

Environment:
    GLADEX_REPO_DIR   repo to check (default: this script's parent repo,
                      realpath'd so the documented /data/tools symlink
                      resolves to /data/repo - the same trap repo-lint
                      documents). Overridden by --repo.
"""

from __future__ import annotations

import argparse
import contextlib
import datetime
import io
import json
import os
import re
import sys

TOOL = "queue-source-check"
EXIT_OK = 0
EXIT_VIOLATION = 1
EXIT_USAGE = 2
EXIT_CANNOT = 3
> build · mimo-v2.6-flash-free
$ cd /data/repo && grep -n '^## \[0.4.14[0-9]\]' CHANGELOG.md; echo ===; sed -n "$(grep -n '^## \[0.4.144\]' CHANGELOG.md | cut -d: -f1),\$p" CHANGELOG.md | head -80; echo "=== total lines ==="; wc -l CHANGELOG.md
5924:## [0.4.140] - 2026-09-30 — the value half of a metavar (queue item (73)) was committed with no heading to match: the main loop writes it here, and `repo-lint`, the `go-compile` suite and the promote-gates `commit-lint` refusal go green again
5952:## [0.4.141] - 2026-09-30 — queue item (79): the labelled `usage:` row must name every option-list pair — 25 missing pairs repaired across 11 `--help` bodies, and the reader that now holds them there
5980:## [0.4.142] - 2026-09-30 — queue item (78): a partial run can land red on `main` — every commit to this repo is now linted before it lands, by a tracked `.githooks/pre-commit` gate that lints the commit the index is about to become
6008:## [0.4.143] - 2026-09-30 — queue item (83): the `- Live:` suite figure is checked against the worktree but can only be pinned to a revision — section F now reports BOTH counts, and the red says which side moved
6028:## [0.4.144] - 2026-09-30 — queue item (85): the detached-child census reads TEXT, not runtime — the boundary now lives in the suite's own header and is enumerated as call-site records, with both of its directions asserted
===
## [0.4.144] - 2026-09-30 — queue item (85): the detached-child census reads TEXT, not runtime — the boundary now lives in the suite's own header and is enumerated as call-site records, with both of its directions asserted

### Added — the boundary, as records instead of a bullet

- **`INDIRECT <name> <lineno> <kind>` records in `SCAN_AWK`, three kinds: `eval`, `cmd-str`, `source`.** The census answers "which tool can start a child that outlives it?" by reading the committed text of `tools/`, one file at a time — so a construct that is *not* text in those files was never searched for, and item (85) existed precisely because nobody could say how big that hole was: it was a sentence in a queue bullet. The reader now reports the **call sites of indirection** — where code reaches execution through something file-text scanning cannot follow — as a fourth record type. They are reported, **never counted as a construct** (`DETACH`/`BG`/`WAIT` and their consumers are untouched, field 3 is still the line number for every kind), so the boundary is an enumerated list in every run's output instead of an unmeasured claim.
- **Every rule is anchored at COMMAND position, because a census that reports prose is a census nobody can read.** The first draft of the `source` rule was a bare prefix match, and it reported **8 lines** of `tools/` — `authoritative source (RDAP/whois)`, `Single source for the argparse epilog`, `Copying source from {repo_src}` — as indirection call sites. The shipped rule needs the **first token** *and* a **path-ish argument** (`"`, `$`, `/`, `./` or a `*.sh` name); `eval` is anchored the same way (line start or after `;`, `|`, `&`). Measured, not estimated: the naive rule reproduced those 8 false sites on the live tree when it was installed back as a mutation.
- **The boundary is declared where the census lives — the suite's own header** — so the limitation travels with the reader instead of living in a progress log: what the census reads (`tools/`, as text), what it cannot read (a payload built at runtime, a construct in a helper outside the directory), and what it does about it (name the call site, never pretend to follow it).

### Verified — 42 assertions, seven mutation controls re-run at 42, and both halves of the boundary asserted

- **`bash tests/test_detached_children.sh` → 42 passed, 0 failed (≈ 9 s)**, **A 26 / B 13 / C 3**, up from **33/0** with no assertion removed or loosened. **A17–A20** are the call sites (an `eval` whose payload arrives at argv, a `python3 -c "$SNIPPET"`, a `. "$HELPER"` reaching outside the scan, and prose that merely mentions `eval` — the first three reported, the last not); **A21/A22** are the item itself asserted from **both directions**: the helper's `setsid` is **not** seen *through* the `source` (`A21`), and the **same reader reports that very construct once its file is inside the scan** (`A22`) — so the gap is a boundary and not a missing regex; **A23/A24** keep the live tree honest (real call sites found, every record carries a kind); **A25** holds the source rule at first-token-plus-path.
- **All seven mutation controls re-run in sandbox copies on 2026-09-30 at 42 assertions, and every historical figure holds**: second tool gaining `setsid` → **3 red (`A2 C1 C2`)**; second tool gaining `&` → **1 red (`A4`)**; the real `wait` commented out → **1 red (`A5`)**; guard tokens renamed → **2 red (`A3 C1`)**; pre-fix last-vs-last reader → **1 red (`A15`)**; over-strict bare-`wait` pairing → **1 red (`A16`)**; untouched copy → **0 red / 42**. Two **new** controls for this step: the `cmd-str` rule removed → **2 red (`A18 A23`)**, 40/2; the anchored `source` rule swapped for the naive prefix match → **2 red (`A19 A25`)**, 40/2, live tree reading **10 real + 8 prose** call sites under it.
- **Live tree at this commit: 10 call sites across 3 tools, all `cmd-str`** — `go-compile-drift-verify` 337/339/399, `ip-drift-cron` 114/131, `system-status` 932/1024/1028/1316/1943 — each a `python3 -c` payload, two of them (`system-status` 1316, 1943) a whole program held in a shell variable. The construct census is unchanged by this step: **1 tool `DETACH` (`regression-run`), 1 tool `BG` (`go-compile-drift-verify`), `A2–A5` unchanged**, and **0** `eval` / `source` call sites exist today, which is why A23 is a floor and not a count.
- **STEP 0 ran first**: 0 unread `investor_to_agent` in both databases, `tools/inbox-status` → exit 0, "OK - nothing owed", **79 entries / 79 handled / 0 open**. No investor row was owed, so none was inserted. Spend **0.00** (free `*-free` models only), zero DNS writes, no mail sent, no `app/src/php` edit → no reviewer gate and no promote.

### Queue -> agent-logs/PROGRESS.md
Authoritative next-candidate queue lives in `agent-logs/PROGRESS.md`, in the `Next-candidate queued, not actioned` bullet of the newest run entry — it is not duplicated here, and has not been since `[0.4.67]`. To read the live queue, open `PROGRESS.md`; to know whether that is still true, run `tools/queue-source-check`.
=== total lines ===
6044 CHANGELOG.md
> build · mimo-v2.6-flash-free
$ cd /tmp/opencode && D=$(mktemp -d); printf '#!/usr/bin/env bash\nsleep 61 &\np=$!\nwait;\n' > "$D/wait_semicolon"; printf '#!/usr/bin/env bash\nsleep 61 &\np=$!\nwait "$p" || true\n' > "$D/wait_pipe"; printf '#!/usr/bin/env bash\nsleep 61 &\np=$!\nwait "$p"; true\n' > "$D/wait_arg_semi"
for f in "$D"/*; do printf '%-20s new=%s old=%s\n' "$(basename $f)" "$(awk -f new_bgwait.awk "$f")" "$(awk -f old_bgwait.awk "$f")"; done
echo "--- does bash wait in the 'wait;' shape? (readers say NOT waited) ---"
cat > "$D/w.sh" <<'FIX'
sleep 1 &
p=$!
wait;
FIX
bash "$D/w.sh"; echo "rc=$? (script waited for the child: rc 0 means the wait ran)"
echo "--- WAIT-regex hits in live tree for ';' or '|' terminators ---"
grep -rnE '(^|[; 	])wait[;|]' /data/repo/tools/ /data/repo/tests/ 2>/dev/null | grep -v REGISTRY | wc -l
rm -rf "$D"
wait_arg_semi        new=0 old=0
wait_pipe            new=0 old=0
wait_semicolon       new=1 old=1
--- does bash wait in the 'wait;' shape? (readers say NOT waited) ---
rc=0 (script waited for the child: rc 0 means the wait ran)
--- WAIT-regex hits in live tree for ';' or '|' terminators ---
0
> build · mimo-v2.6-flash-free
$ cd /data/repo && tail -c 400 CHANGELOG.md | cat -A | tail -5
t M-bM-^FM-^R no reviewer gate and no promote.$
$
### Queue -> agent-logs/PROGRESS.md$
Authoritative next-candidate queue lives in `agent-logs/PROGRESS.md`, in the `Next-candidate queued, not actioned` bullet of the newest run entry M-bM-^@M-^T it is not duplicated here, and has not been since `[0.4.67]`. To read the live queue, open `PROGRESS.md`; to know whether that is still true, run `tools/queue-source-check`.$
> build · mimo-v2.6-flash-free
$ cd /data/repo && cat >> CHANGELOG.md <<'ENTRY'

## [0.4.145] - 2026-09-30 — queue item (86): the `wait` argument is PARSED, not counted — arity and identity, the two opposite directions in which the counting reader was wrong

### Changed — `bg_waited` tracks jobs as identities instead of decrementing a counter

- **Each `&` becomes a numbered job, `name=$!` records WHICH job that name refers to, and one job is covered per argument.** The (84) pairing answered *"which child does this wait cover"* with a count: one decrement per argument-bearing `wait`, whatever its argument list named, and no way to tell two names that point at the same job apart. Both halves are wrong, and they fail in **opposite** directions: **arity** — `wait "$a" "$b"` names two jobs and decremented once, so a file that waits for both its children read as *not* waited (`pending 1`), i.e. the reader could **fail a tool that is correct**; **identity** — `wait "$a"` twice covered one job twice, so a file with a sibling still sleeping read as *waited* (`pending 0`), i.e. the reader could **pass a tool that orphans a child**. The reader now resolves `name=$!` to the job number it names (the newest job at assignment time, which is exactly what `$!` holds) and covers exactly that job; where a name cannot be resolved it falls back to the oldest still-uncovered job, which is the pre-(86) behaviour kept as a *fallback* so an unresolvable name can never invent coverage nobody claimed.
- **The argument list stops at the shell operator that ends it.** Fixing arity without this buys a worse false positive than the one it removes: in `wait "$a" || return 1` the tokens `return` and `1` would otherwise be read as two more job names and cover the whole file. Measured, not guessed — that is exactly what **A29** goes red for when the cut is deleted (control row 4).
- **A bare `wait` still covers every child started before it**, unchanged from `[0.4.127]`, so `sleep & sleep & wait` stays waited instead of becoming the false positive a naive count would make it.

### Verified — 50 assertions, thirteen mutation controls re-run at 50, and the identity claim settled by a process

- **`bash tests/test_detached_children.sh` → 50 passed, 0 failed (≈ 10 s)**, **A 30 / B 17 / C 3**, up from **42/0** with no assertion removed or loosened. **A26** (two children, one `wait` naming both → *waited*; the counting reader left `pending 1`), **A27** (the same job named twice → *not* waited; the counting reader left `pending 0`), **A28** (three children, a `wait` naming two → *not* waited, so arity cannot over-correct into "an argument-bearing wait covers everything"), **A29** (`|| return 1` cut at the operator).
- **B14–B17 run the identity claim on a real process instead of asserting it**: the reader's verdict on that very fixture is *not* waited, the fixture is **done in 1 s** (not 61), the job it named **is** reaped, and the unnamed sibling's `sleep 61` is **still alive** — a process that falsifies the pre-(86) verdict, which scored the same file `pending 0`.
- **All thirteen mutation controls run in sandbox copies of `tests/` + `tools/` on 2026-09-30 at 50 assertions, every patch asserted to have applied, live tree never written, untouched copy 50/0.** Four new: the **pre-(86) counting reader** back → **3 red (`A26 A27 B14`)**, 47/3; **identity resolution removed** → **2 red (`A27 B14`)**, 48/2; **arity removed** (one cover per `wait`) → **1 red (`A26`)**, 49/1; **operator cut removed** → **1 red (`A29`)**, 49/1. The nine older rows: `setsid` plant **3 red (`A2 C1 C2`)**, `&` plant **1 (`A4`)**, real `wait` commented **1 (`A5`)**, guard tokens renamed **2 (`A3 C1`)**, over-strict bare `wait` **1 (`A16`)**, `cmd-str` rule removed **2 (`A18 A23`)**, naive `source` rule **2 (`A19 A25`)** — all unchanged — and **pre-(84) last-vs-last reader 1 red → 5 red (`A15 A27 A28 A29 B14`)**, 45/5: one historical figure *moved*, because every (86) plant and B14 puts a `wait` after the file's last `&`, which is the only question that reader asks. The 42-assertion figure in `[0.4.144]`'s table stays as recorded; 5 is the measured number now.
- **Live tree, measured before and after**: exactly one registered tool backgrounds a child (`go-compile-drift-verify`, `pid=$!` + `wait "$pid" 2>/dev/null`), and the new reader returns `pending 0` for it — the same answer the counting reader gave, so no live verdict moved. **No registered tool writes either (86) shape today** (no multi-argument `wait`, no job named twice), which is why the step is a reader fix with plants rather than a tool fix.
- **STEP 0 ran first**: 0 unread `investor_to_agent` in both databases, `tools/inbox-status` → exit 0, "OK - nothing owed", **79 entries / 79 handled / 0 open**, `grep '^## ' INBOX.md | grep -v HANDLED` empty. No investor row was owed, so none was inserted. Spend **0.00** (free `*-free` models only), zero DNS writes, no mail sent, no `app/src/php` edit → no reviewer gate and no promote.

### Queue -> agent-logs/PROGRESS.md
Authoritative next-candidate queue lives in `agent-logs/PROGRESS.md`, in the `Next-candidate queued, not actioned` bullet of the newest run entry — it is not duplicated here, and has not been since `[0.4.67]`. To read the live queue, open `PROGRESS.md`; to know whether that is still true, run `tools/queue-source-check`.
ENTRY
echo "appended"; tail -3 CHANGELOG.md
appended

### Queue -> agent-logs/PROGRESS.md
Authoritative next-candidate queue lives in `agent-logs/PROGRESS.md`, in the `Next-candidate queued, not actioned` bullet of the newest run entry — it is not duplicated here, and has not been since `[0.4.67]`. To read the live queue, open `PROGRESS.md`; to know whether that is still true, run `tools/queue-source-check`.
> build · mimo-v2.6-flash-free
$ cd /data/repo && grep -n "^## 2026-09-30T2" agent-logs/PROGRESS.md | tail -3; wc -l agent-logs/PROGRESS.md; ./tools/budget-show 2>&1 | tail -8
7211 agent-logs/PROGRESS.md
Spent:       1.50 € (gladex.de — registered + DNS live since 2026-09-20, 1.50/mo)
Remaining:   3.50 €

Log:
  - 2026-09-20: init, no domains ordered yet
  - 2026-09-20: investor APPROVED gladex.de (1.50/mo). Awaiting registrar purchase + DNS creds.
  - 2026-09-20: purchase completed same day — RDAP status active (changed 2026-09-20T11:10:57+02:00), DNS wired via pdns wrapper (A apex/dev/git/info/stats/log/team -> 77.90.15.49, MX 10 gladex.de), dig-verified; DOMAIN.md holds evidence.
  - 2026-09-24: ledger narrative corrected — spent-line still said "purchase pending at registrar" 4 days after registration (queued candidate #1); numbers unchanged (1.50/3.50, no money moved), log history above kept verbatim.
> build · mimo-v2.6-flash-free
$ cd /data/repo && grep -n '^## ' agent-logs/PROGRESS.md | tail -3; echo ===; sed -n '7100,7112p' agent-logs/PROGRESS.md
7082:## 2026-09-30T17:27Z main-loop run — STEP 0 verified first (0 unread, 0 open, nothing owed), then **(78) EXECUTED: a partial run can no longer land red on `main` — every commit is linted before it lands, by `.githooks/pre-commit`**
7150:## 2026-09-30T18:41Z main-loop run — STEP 0 verified first (0 unread, 0 open, nothing owed), then **(83) EXECUTED: section F now reports BOTH counts — what is committed and what is checked out — so a stale `- Live:` figure names the side that moved**
7179:## 2026-09-30T19:48Z main-loop run — STEP 0 verified first (0 unread, 0 open, nothing owed), then **(85) EXECUTED: the census reads text, not runtime — the boundary now lives in the suite's header and is enumerated as call-site records, asserted from both sides**
===
### Closing measurements for the entry above (`9f907f3` pushed; the figure refresh history inside `e658b7b`)

- **The step's own commit, staged by path and none of anyone else's work in it.** `9f907f3` at **2026-09-30T16:29Z**, **15 paths, 315 insertions / 25 deletions** — the 11 tools' usage lines, `tests/test_registry_coverage.sh`, `tools/REGISTRY.md`, `CHANGELOG.md` `0.4.141` and the entry above. Two other identities were mid-edit while I staged (`app/src/php/app.php`, `tests/test_trust_a11y.php`, untracked `tests/test_trust_form.php`), so `git add -A` would have shipped their half-finished `app/src/php` work under my message; `git show --stat 9f907f3` carries exactly my 15 paths. Pushed `edcbb24..9f907f3`.

- **Full regression on the pushed tree, every red attributed and none of them this step's**: `./tools/regression-run --format json` → **71 suites / 6,247 passed / 29 failed / 0 skipped**, `ok false`, exit 1. Two red suites:
  - `tests/test_registry_coverage.sh` **220 passed / 28 failed** — one root cause, read out of the child rather than guessed: `VIOLATION figure_stale 70` with `SUMMARY … fig_disc=71 fig_claim=70 fig_stale=1 violations=1`. Three suites landed between **16:36Z** and **17:10Z** while `tools/REGISTRY.md`'s `- Live:` line still claimed **70**: `tests/test_trust_form.php` in `7cd78e0` ("run 410", which as of 17:20Z still has **no `PROGRESS.md` entry**), then the concurrent run's `tests/test_hire_agent.sh`, then `tests/test_commit_gate.sh`. The **28** are not 28 findings: the file's `exactly one violation - the plant, and nothing else` family turns *one* standing violation into 26 control reds (each control still passing its own plant assertion), plus F3 `want=71 got=70`, F5 and F6. **Not one of the 28 failed on the (79) half** — the child printed `synp_n=61 synp_ok=61` through every one of those runs, so the step this run shipped was green inside a red file.
  - `tests/test_gladex_monitor.sh` **26/1** inside that run and **25/2** (A3+A15) when run directly — both assert a *clean shared working tree*, and both were reading other identities' in-flight `git add` (first theirs: `app/src/php/app.php` + `tests/test_trust_a11y.php`; later mine: the figure edit below, uncommitted at that moment). The suite's own shape is **27/0**; it reports the tree as it was at that second, and `tools/system-status` printed `git-tree [OK] clean` for the same tree one minute later.

- **The figure refresh: measured three times while it was being written, and it is not in my commit.** `70` (at my push, 16:29Z) → **71** read at 16:59Z → **72** at 17:05Z → **73** at 17:12Z, three suites in thirteen minutes with the line refreshed by none of their authors. I wrote the count read at 17:12Z (`**73 suites** discovered` plus a parenthetical recording the whole series and the rule that *the run adding the next suite owns the next refresh*), re-measured it: child → `fig_disc=73 fig_claim=73 fig_stale=0 violations=0 synp_n=61 synp_ok=61`, full suite → **248 passed / 0 failed** at 17:19Z, re-run **248/0** after the push. Then the commit came back **`nothing to commit, working tree clean`**: **`e658b7b` (Sofia, 17:16Z) had already swept my staged `tools/REGISTRY.md` into her commit** with a repo-wide add, alongside `.githooks/pre-commit`, `tests/test_commit_gate.sh`, her `## test_commit_gate.sh` registry section and her two shift logs. So the refresh **is in history, inside someone else's commit**, verified rather than assumed — `git show HEAD:tools/REGISTRY.md | grep -c '70 → 73'` → **1**, Live figure → **73 suites**. The same hazard ran in both directions within ten minutes: staging **by path** is what kept her in-flight `app/src/php` out of my commit, and her repo-wide add is what put my line into hers.

- **Queue item (78) landed under this run, not by it — and it now gates this loop's own commits.** `e658b7b` carries `.githooks/pre-commit` (installed: `core.hooksPath` = `.githooks`), `tests/test_commit_gate.sh` and its `REGISTRY.md` section: the gate writes the *index* to a tree, wraps it in a throwaway commit with `git commit-tree` and lints **that**, so the verdict before the commit and the verdict after the push are the same verdict over the same blobs — the fix for the shape that produced this run's own red file. **I wrote none of it, so (78) is left for its owner to strike**; what I can report is that the first commit I made after it landed went through the gate.

- **Gates on the pushed HEAD (`9bf7103`)**: `tools/repo-lint --format json` → **exit 0, `ok true`, `failures []`, `errors []`, `citations_missing []`**, **146 headings / 146 unique / 5,059 citations checked**, `go_compile` **45 module files compile clean**; `tools/system-status` → **Overall: ALL SYSTEMS HEALTHY**, `git-tree [OK] clean`, `go-compile [OK] … commit 9bf7103`, `queue-source [OK] [0.4.141] pointer-only, 111 item line(s) frozen across 98 section(s), 99 PROGRESS bullet(s)`, `investor-duty [OK] owed=0`, and `promote-gates` = **`commit-lint OK | dev-sync OK | ship-tree OK`** with the single remaining refusal `verdict STALE (… re-review required)`, which is the **reviewer's** next review; `tools/queue-source-check` → exit 0; `tools/source-sync-check` → **in sync, 44 files, exit 0**; `tools/healthcheck` → **dev HEALTHY / prod HEALTHY**, both `0.4.28`. **Served, not assumed**: `GET /changelog` → **200, 1,114,274 bytes, one md5 `ab5d5acd84aebbaf8d516ba4ab2f762f` across dev `:8000` and prod `:8001`**, `v0.4.141` present in the page; `GET /api/changelog` → **total 146, first `0.4.28`, last `0.4.141`** (the first heading did not move, so `GLADEX_APP_VERSION` stays **0.4.28**).
> build · mimo-v2.6-flash-free
$ date -u +%Y-%m-%dT%H:%MZ; cd /data/repo && git status --short
2026-09-30T20:56Z
 M CHANGELOG.md
 M tests/test_detached_children.sh
 M tools/REGISTRY.md
> build · mimo-v2.6-flash-free

Select another run log from the list above. Only files matching run-YYYYMMDD-HHMMSS-N.log are readable.

[Wed Sep 30 22:22:00 2026] 127.0.0.1:58472 Accepted
[Wed Sep 30 22:22:00 2026] 127.0.0.1:58472 Closing
[Wed Sep 30 22:22:24 2026] 127.0.0.1:48498 Accepted
[Wed Sep 30 22:22:24 2026] 127.0.0.1:48498 Closing
[Wed Sep 30 22:22:24 2026] 127.0.0.1:48510 Accepted
[Wed Sep 30 22:22:24 2026] 127.0.0.1:48510 Closing
[Wed Sep 30 22:22:24 2026] 127.0.0.1:48518 Accepted
[Wed Sep 30 22:22:24 2026] 127.0.0.1:48518 Closing
[Wed Sep 30 22:22:24 2026] 127.0.0.1:48532 Accepted
[Wed Sep 30 22:22:24 2026] 127.0.0.1:48532 Closing
[Wed Sep 30 22:22:24 2026] 127.0.0.1:48544 Accepted
[Wed Sep 30 22:22:24 2026] 127.0.0.1:48544 Closing
[Wed Sep 30 22:22:24 2026] 127.0.0.1:48560 Accepted
[Wed Sep 30 22:22:24 2026] 127.0.0.1:48560 Closing
[Wed Sep 30 22:22:24 2026] 127.0.0.1:48574 Accepted
[Wed Sep 30 22:22:24 2026] 127.0.0.1:48574 Closing
[Wed Sep 30 22:22:24 2026] 127.0.0.1:48590 Accepted
[Wed Sep 30 22:22:24 2026] 127.0.0.1:48590 Closing
[Wed Sep 30 22:22:24 2026] 127.0.0.1:48596 Accepted
[Wed Sep 30 22:22:25 2026] 127.0.0.1:48596 Closing
[Wed Sep 30 22:22:25 2026] 127.0.0.1:46770 Accepted
[Wed Sep 30 22:22:25 2026] 127.0.0.1:46770 Closing
[Wed Sep 30 22:35:13 2026] 127.0.0.1:48770 Accepted
[Wed Sep 30 22:35:13 2026] 127.0.0.1:48770 Closing
[Wed Sep 30 22:35:13 2026] 127.0.0.1:48784 Accepted
[Wed Sep 30 22:35:13 2026] 127.0.0.1:48784 Closing
[Wed Sep 30 22:35:13 2026] 127.0.0.1:48798 Accepted
[Wed Sep 30 22:35:13 2026] 127.0.0.1:48798 Closing
[Wed Sep 30 22:35:14 2026] 127.0.0.1:48814 Accepted
[Wed Sep 30 22:35:14 2026] 127.0.0.1:48814 Closing
[Wed Sep 30 22:35:14 2026] 127.0.0.1:48828 Accepted
[Wed Sep 30 22:35:14 2026] 127.0.0.1:48828 Closing
[Wed Sep 30 22:36:08 2026] 127.0.0.1:50134 Accepted
[Wed Sep 30 22:36:08 2026] 127.0.0.1:50134 Closing
[Wed Sep 30 22:36:08 2026] 127.0.0.1:50142 Accepted
[Wed Sep 30 22:36:08 2026] 127.0.0.1:50142 Closing
[Wed Sep 30 22:36:08 2026] 127.0.0.1:50152 Accepted
[Wed Sep 30 22:36:08 2026] 127.0.0.1:50152 Closing
[Wed Sep 30 22:46:36 2026] 127.0.0.1:36362 Accepted
[Wed Sep 30 22:46:36 2026] 127.0.0.1:36362 Closing
[Wed Sep 30 22:54:36 2026] 127.0.0.1:51110 Accepted
[Wed Sep 30 22:54:36 2026] 127.0.0.1:51110 Closing
[Wed Sep 30 22:54:36 2026] 127.0.0.1:51126 Accepted
[Wed Sep 30 22:54:36 2026] 127.0.0.1:51126 Closing
[Wed Sep 30 22:54:38 2026] 127.0.0.1:51142 Accepted
[Wed Sep 30 22:54:38 2026] 127.0.0.1:51142 Closing
[Wed Sep 30 22:54:38 2026] 127.0.0.1:51152 Accepted
[Wed Sep 30 22:54:38 2026] 127.0.0.1:51152 Closing
[Wed Sep 30 22:54:39 2026] 127.0.0.1:51154 Accepted
[Wed Sep 30 22:54:39 2026] 127.0.0.1:51154 Closing
[Wed Sep 30 22:54:39 2026] 127.0.0.1:51168 Accepted
[Wed Sep 30 22:54:39 2026] 127.0.0.1:51168 Closing
[Wed Sep 30 22:54:41 2026] 127.0.0.1:51176 Accepted
[Wed Sep 30 22:54:41 2026] 127.0.0.1:51176 Closing
[Wed Sep 30 22:54:41 2026] 127.0.0.1:51192 Accepted
[Wed Sep 30 22:54:41 2026] 127.0.0.1:51192 Closing
[Wed Sep 30 22:56:29 2026] 127.0.0.1:42212 Accepted
[Wed Sep 30 22:56:29 2026] 127.0.0.1:42212 Closing
[Wed Sep 30 22:56:29 2026] 127.0.0.1:42228 Accepted