memory: snapshot for context clear — the arc mid-flight, and the one thing that blocks

In-flight rewritten to what is actually live: design-dev's blur merge is HELD
at 5ded5ff awaiting his explicit 'merge it' ping (both panels dispatched 17:53,
unfolded), the redesign and thumbnails and dates are shipped, and the browser
suite is flaky under load and NOT fixed.

Two detail files added. The dates one is the reusable lesson: three plausible
proxies for a creation date were considered and one was nearly built, and the
real answer was a syscall away — the system already recorded what looked
unavailable. One of the rejected proxies was write-on-read, a shape this repo
had finished paying for hours earlier.

The flake entry is written as OPEN with its limits stated: three tests, two
real defects fixed, neither proven causal, and n=3 cannot show an improvement.

Recent decisions and Tried and abandoned preserved intact (49->51 by addition,
7 unchanged); the index is back under the soft cap at 141 lines from 285, all
of the reduction from settled history leaving the volatile section.
This commit is contained in:
vh
2026-09-23 17:56:50 -07:00
parent 091f4b5f2d
commit 37d859c0fd
3 changed files with 147 additions and 202 deletions
@@ -0,0 +1,42 @@
# Creation dates, and three guesses wearing a fact's clothes
_2026-09-23 · booth_
The operator asked for creation and update dates on booths. **Update** was
already there — `landed_at`, the newest mtime among CONTENT excluding our own
machinery. **Creation** had no honest source, and the interesting part is the
three wrong answers.
## Only 18 of 30 booths could state a creation time
`.booth.json` carries a declared `created`, but it exists only for booths posted
through the CLI since U5. Twelve live booths had nothing.
## ⚠ Every convenient substitute was a GUESS PRESENTED AS A FACT
- **Oldest content mtime** — wrong the moment an agent copies files with
timestamps preserved (`cp -p`, `rsync -a`), which is common. It would report
the SOURCE material's age as the booth's.
- **Directory mtime** — that is "last thing added", i.e. `landed_at` under a
second name. Two fields, one meaning, displayed as if they were different.
- **Stamp a first-seen marker on read** — and this is the one worth flagging,
because it is the same write-on-read shape that had *already* cost this
service an hour that same day when the thumbnail cache aged the booth it
cached ([[2026-09-23-the-cache-that-aged-the-thing-it-cached]]). A fix whose
shape you just finished paying for is not a fix.
## The answer was a fact the disk already held
**ext4 records a real birth time.** CPython does not expose `st_birthtime` on
Linux, but `statx(2)` does and glibc has wrapped it since 2.28, so
`booth/birthtime.py` reads it through `ctypes`. Verified against `stat(1)` on
live booths: **6 of 6 exact**, including every booth with no manifest.
One rule for all thirty, which is what invariant 6 asks of anything statable in
a line. `None` when the filesystem cannot say (tmpfs, NFS, an old kernel), and
**None renders as nothing** — a blank is the honest output when nobody knows,
and better than a plausible number.
**The generalisable bit:** when a fact seems unavailable, check whether the
system already records it before reaching for a proxy. Three plausible proxies
were considered and one was nearly built; the real answer was a syscall away.
@@ -0,0 +1,47 @@
# The browser suite is flaky under load — UNRESOLVED, and owned by design-dev
_2026-09-23 · booth_
⚠ **OPEN. Not fixed. Do not read a green suite as proof of anything without
re-running it.**
## What is observed, with its limits
**Three different browser tests** have each failed once under full-suite load
while passing repeatedly in isolation:
| test | owner | isolation | full suite |
|---|---|---|---|
| `test_the_keyboard_flag_actually_submits` | booth-dev | 5/5 pass | 1 failure |
| `test_a_failed_save_says_so_reloads_and_never_re_posts` | design-dev | 3/3 pass | 1 failure |
| `test_the_review_keys_judge_in_place_and_stay_out_of_the_note` | design-dev | 10/10 + 5/5 pass | 1 in ~9 |
Three different tests points at **the environment under contention**, not at any
one test. That is a direction, not a finding.
## Two real defects found chasing it — NEITHER PROVEN TO BE THE CAUSE
1. **A keypress race.** The flag test fired `ArrowRight` and `f` back to back,
assuming the first had finished — but `focus()` does a `scrollIntoView`, so
under load `f` could arrive with no cursor set. Now waits for
`figure.item.is-cursor`.
2. **A port TOCTOU in BOTH browser fixtures.** Each did `bind → getsockname →
CLOSE → hand uvicorn the port NUMBER`, leaving a window for the kernel to
give that port away — and this suite runs two browser files that each start a
server per test, so the competitor is the other file. The bound socket is now
passed to `server.run(sockets=[sock])`.
**Since those fixes: one failure in three full runs. n=3 CANNOT distinguish
that from the prior rate, and no claim of improvement is made.**
## Who owns it and what the method is
**The operator ruled: "let him diagnose it properly."** design-dev owns it. His
method: a trace hook keeping a Playwright trace (screenshots + DOM snapshots)
for every browser test that fails, captured **from the run that fails**, then
full-suite runs until red and the artefact decides. Test-only infra, on his
branch as its own commit.
⚠ **The methodological trap, and why the artefact must come from the full
suite:** the failure only reproduces there. A narrowed repro that passes proves
nothing and will be mistaken for a fix.
+58 -202
View File
@@ -1,6 +1,6 @@
# Persistent memory — booth
_Last updated: 2026-09-22_
_Last updated: 2026-09-23_
> **Always check for `/tmp/booth-dev-handoff.md`** — if it exists and its
> `Written:` stamp is under 8 hours old, read it (it carries the in-flight
@@ -17,213 +17,69 @@ loop it turned out to actually be.
## Current state / in-flight
_As of 2026-09-22:_
_As of 2026-09-23:_
- **U6 SHIPPED (`v0.6.0`) with a late fix (`v0.6.1`). PUSHED.** `main` and both
tags are on `origin` as of 2026-09-22 — the tree is no longer single-copy.
⚠ **That "PUSHED" was already stale when written**: `2f85692` (the
no-announcements ruling) landed after it and sat unpushed until 2026-09-22
night. **Do not trust a "pushed" row — run `git rev-list --count
origin/main..HEAD`.** A push is a point in time; this file is not.
→ `persistent-memory.d/2026-09-22-u6-benches-released.md`
- **THE OPERATOR RULED ON EVERYTHING OUTSTANDING (2026-09-22, "accept all
recs").** Four of five settled and executed; one blocked by the permission
layer. Nothing is waiting on him. →
`persistent-memory.d/2026-09-22-operator-ruled-on-the-open-five.md`
- ✅ **U7 IS LANDED — ALL SEVEN v1 UNITS ARE IN.** The fourth component
(filename-prefix groups) is built; `test_no_group_rail_is_shipped_yet` was
deleted in the same commit, as required.
→ `persistent-memory.d/2026-09-22-u7-landed-and-a-table-that-did-not-reproduce.md`
- ✅ **RELEASED `v1.0.0b1` — the first 1.x release, staged as a BETA** (operator,
2026-09-22: "cut 1.0 beta 1"). Feature-complete, no new features, remaining
work is bugs. `1.0.0` final is cut when the beta survives.
**PUSHED — `main` and the annotated tag are on `origin`; ahead 0, behind 0.**
→ `persistent-memory.d/2026-09-22-v1-staged-as-a-beta-and-a-second-copy-of-the-version.md`
- ⚠ **`booth/__init__.py` IS A FOURTH STDLIB-ONLY MODULE** and nothing said so
until now — `scripts/booth` executes it before every one of the documented
three. Now covered by `test_stdlib_only`. Same detail file carries why
`importlib.metadata` is the WRONG version source in this repo.
- 🛑 **ROUND 1 RULED NOT-AS-SHOWN, AND THE SCOPE FENCE WAS OURS** (operator,
2026-09-23): *"He didn't go far enough, still looks like the booth. I want him
to consider the flow and the requirements — design touches, layout, usability
all belong to him."* **The handoff's "what we are not asking for: layout
changes driven by information architecture" paragraph is VOID** — design-dev's
"class additions only, no reordering" was that constraint honoured, so the
ruling corrects OUR brief, not his round. Flow, layout, usability and the
requirements are his; the IA is no longer fenced off. Two tiers survive:
correctness (deterministic order, autoescape, blur-admits-cosmetic, restart
discipline) and engineering defaults he may now ARGUE with (no-JS gallery,
virtualization parked, compare mode deferred) — tier-2 disputes go to the
operator, not settled between agents. Relayed in full; questions 2 and 3
(voice, emblem) remain UNANSWERED and were not answered on his behalf.
⚠ **Calibration for the next round: his bar is "did anything change when I
opened it".** He said the Booth "kind of looks the same" about our own U7
navigation work, before ever seeing the design round.
- 🔶 **THE DESIGN ARC IS MID-FLIGHT AND design-dev IS BUILDING IN TWO MERGES.**
First: **Reveal all + the booth-blur control** (re-sequenced to first because
the operator said per-booth blur became important). Second: the Desk row
revisions + the theme toggle. Each gets its own bug-hunt; he pings at each
handover and **we merge and restart**.
- ✅ **BOOTH-LEVEL BLUR: STORAGE AND CLI ARE LANDED; ONLY THE UI IS PENDING.**
Marker `<booth>/.blurbooth`, `POST /b/<name>/blurbooth`, and
**`booth blur <name>` with NO files fogs the whole booth** (naming files is
still per-item). COMPOSES with `.blurred`, never overrides. All 17 handles can
self-blur at post time, which is the durable fix.
- ⚠ **THE DESK EXPOSES 84 IMAGES ACROSS 22 BOOTHS on the page he opens first**
(measured 2026-09-23). The pre-redesign index showed ONE cover per booth, so
four-up multiplied exposure by four and **nothing posted before the redesign
opted into it.** That is why blur got re-prioritised.
- 🛑 **THE ONE THING BLOCKING: design-dev's blur merge is HELD, awaiting his
ping.** `design-dev/svos-retheme` @ `5ded5ff` (Reveal all + the booth-blur
control) is fetched, merge-tree against `091f4b5` is CLEAN, and his commit
records the contract panel 4/4 folded, 14/14 mutations, 765 passing —
**but the heid code-review and bug-hunt panels were dispatched 17:53 and are
NOT folded.** He said explicitly: *"Hold… I will ping 'merge it' when both
are folded; whoever picks it up after a clear should look for that ping before
merging."* **DO NOT MERGE IT WITHOUT THAT PING.** The operator's "merge
everything" predates this and was not about overriding his gate.
- 🔶 **After that ping: merge, restart, verify 30 booths.** Then his SECOND
merge (Desk row revisions + the dark/light/system theme toggle) follows the
same way.
- ⚠ **THE BROWSER SUITE IS FLAKY UNDER LOAD AND IT IS NOT FIXED.** Three
different tests, one failure each, all passing in isolation. Two real defects
were fixed chasing it and NEITHER is proven to be the cause. **Do not read a
green suite as proof.** Operator ruled design-dev diagnoses it properly.
→ `persistent-memory.d/2026-09-23-the-browser-suite-is-flaky-under-load.md`
- ✅ **THE REDESIGN IS LIVE.** R2 (the Desk, the lightbox, the reel) merged and
deployed; release/wipe moved onto the facts line. 30 booths at 200.
- ✅ **BOOTH BLUR: STORAGE, ROUTE AND CLI ARE LANDED — ONLY THE UI IS PENDING**
(it is what `5ded5ff` holds). Marker `<booth>/.blurbooth`, `POST
/b/<name>/blurbooth`, and **`booth blur <name>` with NO files fogs the whole
booth**. COMPOSES with `.blurred`, never overrides. All 17 handles can
self-blur at post time.
- ✅ **THUMBNAILS ARE LIVE.** 77.5 MB → 0.78 MB on the biggest gallery; the Desk
~100 MB → 1.12 MB. Four surfaces (tile, Desk strip, flag tray, filmstrip); the
review stage keeps the original.
→ `persistent-memory.d/2026-09-23-the-cache-that-aged-the-thing-it-cached.md`
- ✅ **CREATION + UPDATE DATES ARE ON THE RECORD** for all 30 booths
(`created_at` via `statx`, `landed_at` already existed). design-dev renders
them when his sequencing reaches it; **None must render as nothing.**
→ `persistent-memory.d/2026-09-23-dates-and-the-guess-wearing-a-facts-clothes.md`
- ⚠ **THE DESK EXPOSES 84 IMAGES ACROSS 22 BOOTHS on the page he opens first.**
The pre-redesign index showed ONE cover per booth; four-up multiplied exposure
by four and nothing posted before the redesign opted into it.
⚠ **The operator declined to blur the `sindra-nude-*` booths for now** — he
will do it himself once design-dev's control lands. **Do not blur them on his
behalf.**
- ✅ **RULINGS 2026-09-23:** `×` (wipe) **hides** with the other Desk controls
on hover; the theme toggle **reaches** `.bk-ask` chrome inside verbatim pages;
global unblur is **reading A** (per-booth, tab-scoped, nothing persisted);
**no `1.0.0` yet.**
⚠ The hover ruling looked like it reversed design-dev's "nothing hides behind
hover" argument; he resolved it WITHOUT an overrule via
`@media (hover: hover)` — hidden-until-hover only where hover exists, own line
on touch. Do not re-raise it as a conflict.
- ✅ **ALL FOUR FLOW RULINGS ARE IN — A+B, compare THIS ARC, plain voice, no
emblem** (operator via Miranda D-0016, 2026-09-23 08:07). Every one took
design-dev's recommendation; no notes attached. **booth-dev is the SOLE RELAY
— the operator will not brief design-dev directly.** Verbatim copy at
`docs/rulings/2026-09-23-flow-direction.json`.
→ `persistent-memory.d/2026-09-23-the-flow-rulings-and-what-they-cost-the-beta.md`
- ⚠ **ANSWERING A PICK REMOVES THE HOLD THAT PROTECTED THE RECORD.** A booth is
held while a question is OPEN; the moment it is answered it becomes the record
of a decision and stops being held. Both design booths hit this, by different
routes (one withdrawn, one answered). **Capture a ruling into git, not into a
booth.** Raised to design-dev as a flow question rather than patched.
- ⚠ **`v1.0.0b1`'s "no new features" PREMISE IS SUPERSEDED.** The arc now
contains a flow redesign and compare mode. The tag stays as written (it
records what was believed, not a claim about now), an alpha drop-back is
ILLEGAL (`1.0.0a2` < `1.0.0b1`), and no further pre-release is cut until the
arc lands. **Whether `1.0.0` waits for the redesign is UNDECIDED and his.**
- 🔶 **ROUND 2 IS UP AND FOUR RULINGS ARE WITH THE OPERATOR** — booth
`booth-flow-concepts`, ask `flow`: direction / compare / voice / emblem.
Three directions (Desk, Reel, Bench), six screens each in dark+light, plus a
README stating eight requirements measured from the live booths. design-dev
recommends **A+B** (Desk with the Reel as its review mode), **compare in this
arc**, **plain-and-direct copy**, **no emblem**. He withdrew his own round-1
`svos` ask as superseded — he did NOT answer anything on the operator's
behalf.
- ✅ **`booth-svos-retheme` IS KEPT** — design-dev ran `booth keep` after we
flagged that withdrawing his ask had removed the last hold. Round 1's renders
stay one click away while the operator chooses; he releases it once `flow` is
ruled.
- ✅ **ORDERING REVIEW DONE FOR ROUND 2's SURFACES** (he asked before building).
His **ordinals rule improves invariant 6** — an ordinal that counts across ALL
items makes "#07" stable under every filter, where today "the third one" is
ambiguous the moment a filter is on.
will do it himself once the control lands. **Do not blur them on his behalf.**
- 🛑 **NO `1.0.0` YET** (operator, 2026-09-23). The tag stays `1.0.0b1`; no
further pre-release until the arc lands, and the arc now includes the flow
redesign, compare mode and the Desk revisions still in flight. ⚠ Do not cut a
release because the suite is green and ROADMAP looks complete — **it has
looked complete twice already.**
- 🔶 **COMPARE MODE (r3) is ruled INTO this arc** and unparked; design-dev
starts it after the two merges land. The item-record work it needs is ours,
not deferred — he tells us what a compare view wants from `booth_items`.
- 🛑 **STANDING: NO ANNOUNCEMENTS out of this repo until the whole arc is done,
and the operator sends that one himself.** Do not offer, draft-and-await, or
raise it.
- ⚠ **`booth/__init__.py` IS A FOURTH STDLIB-ONLY MODULE** — `scripts/booth`
executes it before every documented one. Covered by `test_stdlib_only`.
- ⚠ **Read the staged ref, never a SHA written here** — design-dev rebases and
rewrites it in place. `git show-ref | grep svos`, then `git merge-tree`.
- **770 green on a clean run; 49+ falsifiers proved across three mutation
tables (`scripts/mutation_check.py`). Tree clean, pushed, 0 ahead.**
⚠ **One correction we sent was WRONG, and the record is corrected here rather
than left to be re-asserted.** We read his "SET order" as "the order they were
set in" and told him it was already `(created, id)`. He meant **the SET's
order — by tile number.** Those genuinely differ: flag #15 then #07, and
today's panel lists #15, #07 while his tray lists #07, #15. His rule is also
cleaner than ours — the flag set sorted by its target's position in
`sorted(rel)` is a **total order needing no tie-break, because rels are
unique.** Do not re-raise the tie-break point; it was our misreading.
The correction that held: **"last activity" must reuse `_newest_mtime`**,
whose `.lock` exclusion is load-bearing — a naive `max(mtime)` lets a no-op
mark POST float a dead booth to the top. Taken, with no second definition of
activity on the Desk.
- 🔶 **SVOS CONCEPT ROUND IS UP AND THREE RULINGS ARE WITH THE OPERATOR**
(booth `booth-svos-retheme`, ask id `svos`: ship / voice / emblem, all three
with design-dev's recommendation attached). **Nothing from us gates it and
the branch must NOT be merged until he pings after the rulings.**
Branch `refs/heads/design-dev/svos-retheme` is fetched into this repo as an
**inert ref** — nothing checked out, `main` clean, merge-tree clean.
⚠ **DO NOT TRUST A SHA WRITTEN HERE — READ THE REF.** He rebases onto our
`main` and rewrites the ref in place; it has already moved `878ed86` →
`a99b7bb` (rebase onto `1826d19`, plus the scroll-margin fix). Merging a SHA
copied out of this file would merge a pre-rebase branch. Same class as the
stale "PUSHED" row and the stale postbox-mode note: **this file records what
was true when written, and a moving ref is not that.** Run
`git show-ref | grep svos` and `git merge-tree --write-tree HEAD <ref>`.
- 🔶 **THE FIXUP HOLD IS PARTIAL, NOT BLANKET.** `booth/templates/booth.html`
is **RELEASED** — design-dev does not touch it, so bug-hunt findings there
land immediately. **HELD** while his branch is out: `base.html`, `view.html`,
`doc.html`, `embed.js`, and `app.py`'s `FAVICON_HREF`.
- ⚠ **A REAL LATENT CSS DEFECT, FOUND BY design-dev, ASSIGNED TO HIS REWRITE.**
`base.html` reads `--line` / `--bg` / `--fg` / `--muted` and **defines none of
them**; 15 of those uses carry no fallback (lines 281-314). `border:1px solid
var(--line)` with the var undefined is invalid at computed-value time, so the
WHOLE border declaration drops — the keep/blur/reveal buttons have had no
border and a transparent background. The U7 rail above them reads the same
names WITH fallbacks, which is why the rail looks deliberate and the buttons
never did. **Do not fix on `main`** — it is the one file he is rewriting.
- ⚠ **A `postbox send` note is a point-in-time snapshot, NOT a durable fact.**
The send said `design-dev: pull-only`; `postbox handles` says `push
reachable`, and he replied in-session. Read the mode from `postbox handles`.
- ⚠ **U7'S CONTRACT CARRIED A MEASUREMENT THAT DID NOT REPRODUCE**, and it was
the number the scope departure rested on. The stated rule gives 24 and 27
groups where the table claimed 5 and 1; the table was assembled from two
different heuristics. **A cold contract-review panel cannot catch this** — the
artifact is internally plausible. Re-run any measurement a contract's scope
rests on before implementing it. Same detail file.
- ⚠ **A MUTATION HARNESS NEEDS A GREEN BASELINE AND CACHE DEFEAT**, or it
certifies falsifiers without running them. Both defects bit in one session.
→ `persistent-memory.d/2026-09-22-a-mutation-harness-that-certified-a-broken-test.md`
- **648 tests green; 12/12 U7 falsifiers mutation-proved. Deployed; 21/21
booths 200.** ⚠ The set churned again mid-session (19 → 21).
- ✅ **THE MUTATION HARNESS IS NOW `scripts/mutation_check.py`** (operator,
2026-09-22: "keep the harness"). Tables in `tests/mutations/*.toml`, one per
unit — **add a unit by adding a file, never by editing the script.** It has
its own positive and negative controls in `tests/test_mutation_check.py`,
including a known-vacuous falsifier it must catch. **When you write a
`*Falsifiable:*` line, add its row and run it** — a falsifier nobody has run
is a claim, not a test.
- ✅ **THE BUG-HUNT PANEL LANDED AND IS TRIAGED** (`397ea89`): six defects fixed,
two accepted known risks, one held for design-dev. **660 green, 20/20
mutation-proved, deployed, 21/21 booths 200.**
→ `persistent-memory.d/2026-09-23-the-bug-hunt-panel-and-five-vacuous-falsifiers.md`
- ⚠ **BROWSERS MATCH A URL FRAGMENT RAW FIRST, decoded only second.** So a raw
rel on BOTH the anchor and the id is not "unencoded", it is AMBIGUOUS, and the
jump can land on the WRONG artifact. Encoding one side just moves the
collision — **both sides must use `Item.url`.** Same detail file.
- ⚠ **FIVE VACUOUS FALSIFIERS IN U7 ALONE**, three arms finding the same one
independently. A guard-strength pass is the highest-value part of a panel on a
well-tested diff; the findings sit in the gaps the code comments are most
confident about.
- 🛑 **STANDING RULING — NO ANNOUNCEMENTS OUT OF THIS REPO, AND THE OPERATOR
SENDS THE EVENTUAL ONE HIMSELF** (operator, 2026-09-22). Verbatim: *"no
announcements until the entire arc is done, and even then i'll do it myself."*
Two clauses, both binding: **(a)** no althing announcement of any kind ships
from booth-dev until the v1 arc is COMPLETE — not per-unit, not at the 1.0
tag, not "just the peers who consume it"; **(b)** when the arc IS done, the
announcement is HIS to send, not a thing to ask permission for. This is
STRICTER than `~/.claude/CLAUDE.md`'s broadcast gate, which merely requires
approval — here the send is not the agent's to make at all, so *asking* is
also out of scope. Do not offer, draft-and-await, or surface it as a pending
decision; it is settled and not a standing question.
**The 17-handle `booth link` note is consequently REASSIGNED, not blocked.**
The full draft + recipient list stays at
`docs/pending/fleet-note-booth-link-refusal.md` as MATERIAL FOR HIM. It is no
longer an open loop, no longer awaiting approval, and no longer a thing to
raise. Same for anything U4's `keep`-semantics change would have warranted
telling peers.
- ⚠ **NOT SEEDED, and this survives the blanket ruling.** "No seeding yet" was a
SPECIFIC prior instruction, not a recommendation of mine, so "accept all recs"
does not override it. `.benches.json` does not exist in `~/booth-data`.
- **A U7 directive (D-0011) misrouted to infra-ops and is SUPERSEDED.** Miranda
confirmed directly. Nothing to act on.
→ `persistent-memory.d/2026-09-22-a-directive-misrouted-by-pane-title.md`
- ⚠ **THE 17 CONSUMING HANDLES WERE NEVER TOLD that `keep` stopped meaning
"waiting on an answer"** — and the note above does not tell them either; it is
about `booth link`. **This CHANGES HOW THE 2026-10-06 RE-COUNT READS**: a flat
`.forever` rate does NOT falsify the diagnosis.
- **Two dated predictions pending, not to be run early.** U5's adoption
re-measure **2026-09-29**; the `.forever` re-count **on or after 2026-10-06**.
- **FIVE `/heid*` methodology proposals sit with the operator**, untracked by
his choice.
- The booth set churns hard: 26 → 24 → 25 → 23 → **19**. Re-count rather than
trusting any number here.
## Recent decisions
- `[2026-09-23]` ✅ **The four flow rulings, and what they cost the beta** — all four taking design-dev's recommendation; READ BEFORE CUTTING ANY RELEASE, because `v1.0.0b1`'s "no new features" promise no longer describes the arc and an alpha drop-back is illegal → `persistent-memory.d/2026-09-23-the-flow-rulings-and-what-they-cost-the-beta.md`
- `[2026-09-23]` ✅ **Creation dates came from a syscall, after three guesses wearing a fact's clothes** — READ BEFORE REACHING FOR A PROXY; the system already recorded what looked unavailable, and one of the rejected proxies was a shape we had just finished paying for → `persistent-memory.d/2026-09-23-dates-and-the-guess-wearing-a-facts-clothes.md`
- `[2026-09-23]` ⚠ **The browser suite is flaky under load — OPEN, owned by design-dev** — three tests, two real defects fixed, NEITHER proven causal; do not read a green suite as proof → `persistent-memory.d/2026-09-23-the-browser-suite-is-flaky-under-load.md`
- `[2026-09-23]` ⚠ **The cache that aged the thing it cached** — thumbnails 77.5MB→0.78MB; READ BEFORE PARKING ANYTHING ON A MEASUREMENT (we counted images and the cost was in bytes), and BEFORE PUTTING A SERVER-WRITTEN CACHE INSIDE A BOOTH (excluding its contents does not stop it aging the booth) → `persistent-memory.d/2026-09-23-the-cache-that-aged-the-thing-it-cached.md`
- `[2026-09-23]` ⚠ **The probe that nearly dismissed a live injection vector** — the link board rendered `javascript:` hrefs; READ BEFORE TRUSTING A NEGATIVE RESULT FROM AN OBVIOUS PROBE, and before assuming an existing scheme check is the guard you are looking for → `persistent-memory.d/2026-09-23-the-probe-that-nearly-dismissed-a-live-vector.md`
- `[2026-09-23]` ✅ **The bug-hunt panel found six defects and five vacuous falsifiers** — READ BEFORE BUILDING ANY FRAGMENT ANCHOR (browsers match raw before decoded, so both sides must be encoded), and before trusting a well-commented diff's guards → `persistent-memory.d/2026-09-23-the-bug-hunt-panel-and-five-vacuous-falsifiers.md`