Salter-Harris I–V, the parabolic profile and its two lanes, the conducting and
respiratory zones, the four types of respiratory failure, and the mole as the
bridge between one atom and a weighable mass. Each is in the image bank with a
title, a description and a source, and embedded in the article it explains — so
it behaves like every other figure: a thumbnail in the prose, the full view on
a click, findable by what it shows.
Drawn as SVG and not generated as pictures, for two reasons. A schematic has to
have every label exactly right and stay right at any size, and one assembled
from shapes is corrected by editing a line rather than by rolling the dice
again. And a generated raster of anything clinical would be a fabrication
dressed as evidence, which is the one thing a teaching figure must never be.
The sources are kept here beside the docs because they are text, and text
belongs in the repository where it can be diffed.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01TqXevQJhxFrM7jJg82cgZN
The half of "adaptive" that was never built. Difficulty was a filter a learner
could set and nothing the session did on its own, so somebody at 30% on a topic
and somebody at 90% were asked the same questions in the same order.
A question's value is now scaled by how far its level is from the one the
learner's readiness on *that topic* calls for: under 0.45 wants easy, up to
0.78 medium, above that hard. Per topic because strong on growth and weak on
arrhythmias is the normal case and one number across the two describes nobody.
A multiplier and not a filter, for the same reason the reranker is a
permutation: thinning the bank to one level makes it three times smaller, and
on a narrow topic that is the same eight questions every time. One step away
keeps 0.72 of its value, two steps 0.5, and an unlabelled question 0.92 —
unknown is not wrong.
The first band's edge is 0.45 rather than 0.55 so that a topic with no evidence
— which sits at exactly NEUTRAL_RECALL — gets medium questions. Handing
somebody the easiest questions in a subject nobody has measured them on is a
poor way to find out what they know.
Written up in docs/adaptive-sessions.md, which now says five rules.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01TqXevQJhxFrM7jJg82cgZN
A 2,000px radiograph written into an article rendered at whatever width it
happened to be — a wall of greyscale in the middle of a sentence, four
megabytes to draw it, and no way to look at it properly. Every image in prose
is now a 256px thumbnail with the author's label under it, and a click gives it
the screen.
The viewer puts what is known about the figure beside it: its title, what it
shows, and where it came from. `media_assets` gains `source` and `source_url`
for that — a citation belongs to the file, because the same figure used in
three articles is cited the same way in all three, and a licence that turns out
to be wrong is one row to fix rather than three paragraphs to find. Asked for
when the figure is opened, not when the page is drawn.
And `overlay`: the regions an educator has marked, as vector shapes in
normalised coordinates on the unit square, so one drawing is correct in a
thumbnail, in the viewer and on a projector. Off until the learner turns it on
— marks shown before they have looked answer the question for them. Vectors
rather than a second burnt-in picture, for four reasons written down in
docs/image-overlays.md. The tool that draws them is next; this is the storage,
the contract and the reader's half.
On a narrow screen the description stacks above the image rather than beside
it, where it can be read before scrolling to the picture.
Also here: `classify_question_difficulty`, which labelled all 2,924 questions
in batches of twenty-five against a written rubric — 622 easy, 1,634 medium,
668 hard, no failed batches. The column had been NULL on every row since it
existed. Migration l2c3d4e5f6a7.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01TqXevQJhxFrM7jJg82cgZN
Two halves of one TODO, settled with a measurement rather than a guess.
*Shrunk readiness* was already done — `CandidateRanking.accuracy()` pulls a
topic towards NEUTRAL_RECALL by PRIOR_ANSWERS, so one miss does not read as 0%
— and there is now a test pinning it, because the note claiming otherwise
outlived the fix by weeks.
*Difficulty as a dimension the session moves along* cannot be built, and the
reason is a number: all 2,924 questions have a NULL `difficulty`, and the
empirical route is no better at 788 answers over 706 questions from 5 learners,
about one answer each. A ladder scored against that would be scoring noise
while looking as though it worked.
What is built instead is honesty in the control that already exists: the
Difficulty facet counts each level under the other filters and disables one
that would empty the bank, so nobody picks Hard and watches the count fall to
zero with no explanation. Reopen the ordering when something writes that
column.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01TqXevQJhxFrM7jJg82cgZN
The syntax was known only to the code that parses it. This is the author's
side of it: `[[264|respiratory failure]]` and why the id form is the one that
cannot rot, `==key points==` and the three rules the parser follows,
`{{phrase|tip}}`, what each of the three views is for, how to link a question
or a whole topic, and what deleting does to a draft versus to something that
has been published.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01TqXevQJhxFrM7jJg82cgZN
Four things that share a spine, so they arrive together.
**Folders.** A hand-picked set of questions, and the fourth thing a grant can
name beside exam, discipline and category. Deliberately not `user_collections`
with a sharing flag: a library is a consequence of access — you save what you
can already see — while a folder is a source of it, and one table holding
thousands of private lists beside a handful that confer permission is one
mistake away from a leak. Built from the question manager, granted on /access.
Membership stays with the owner and moderators so a grantee cannot widen their
own reach, and deleting a folder takes its grants with it.
Two live constraints had to be rewritten to accept it: `ck_grant_has_a_dimension`
and `uq_grant_dimensions` both predate `folder_id`, so a folder-only grant
failed the check and two folder grants collided on the unique index.
**Per-question feedback.** The learner's half already existed. What was wrong
was who could read it: any grant at all let an educator list and delete reports
about the whole bank. Reports are now scoped by `question_scope_predicate`, the
same predicate that decides which questions that educator can see, and a reply
thread makes the report a conversation the learner can follow rather than a
form that swallows what they said.
**Per-section notes and article feedback.** Two tables on purpose:
`article_section_notes` is private to whoever wrote it, `article_feedback` goes
to whoever maintains the article. Both point at the section id inside
`articles.sections` rather than at `article_section_index`, whose rows are
dropped on unpublish — a cascade from there would delete a learner's writing
because an educator took an article down for an afternoon. A rename keeps a
note attached; a deleted section leaves it marked orphaned under the heading it
was written on, for its writer alone to remove.
The header's feedback badge covers both, because questions and reading are the
same job to whoever is doing it.
Migration i9f0a1b2c3d4. 556 backend and 572 frontend tests pass.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01TqXevQJhxFrM7jJg82cgZN
Retrieval fused a bi-encoder and BM25 by reciprocal rank. A bi-encoder embeds a
document long before the question exists, so the two never meet: it is good at
"same topic" and mediocre at "answers this". A cross-encoder reads the pair.
The proxy already serves three — `cohere-rerank-v4.0-pro` is the default and
measurably better than the fast variant. Query text goes exactly where the
embeddings already go, and nothing new was signed up for.
It found a defect nobody was looking for. In AI Mode each finder scored
`1/(1+rank)` *within its own corpus*, so the best article, section, question and
card all scored 1.0 and the shortlist was a meaningless round-robin. A
cross-encoder is the first thing in this system that can compare a question
with a section. Candidates per kind widened so it can select rather than merely
reorder.
Measured against labels neither ranker produced. Questions, 60 disease tags:
precision@3 0.394 → 0.483. Sections, 60 article titles: 0.772 → 0.833.
"Management of bronchiolitis" led with influenza transmission and a pregnancy
question; "when do you image a first febrile seizure" returned the definition
rather than the sentence saying imaging is unnecessary.
And the honest negative, in docs/reranking.md: board vignettes are written
*not* to name their diagnosis, so on "what causes croup" it prefers a question
that says the word in passing over the barking-cough vignette that never says
it. Some of the bi-encoder's strength is traded away.
Not on the typeahead. A page of results is a choice being made and worth a
third of a second; a typeahead is a word being finished, runs on every
keystroke, and has nothing to judge yet.
The three-state thresholds stay on cosine, argued at the constant: a reranker
only ever sees a shortlist and structurally cannot answer the corpus-wide
question those numbers ask, and whether an answer claims to come from the
library is a promise that must not depend on a network hop.
Every failure returns None and leaves the order alone — unconfigured, no proxy,
connect error, bare 502, timeout, non-JSON, a duplicate or out-of-range index,
a non-numeric score, a list the wrong length. Verified against the running site
with a bogus model name: same results, fused order, no error to the reader.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01TqXevQJhxFrM7jJg82cgZN
The taxonomy rebuild regressed a previously-closed item without anybody
noticing: merging duplicate topics moved questions up onto the surviving root,
so 254 sat at system level where 108 had a perfectly good condition beneath
them. All 108 promoted, zero ambiguous — the script already knew where each
one went; nothing had told it to look again after the merge.
The remaining 146 stay put. There is no condition inside their system to hold
them, which is a gap in what has been written rather than a filing mistake, and
the tally by system says where the writing is missing.
And `Radial Head Subluxation` is `Nursemaid's Elbow`. Two rows for one idea
split its questions and offered a browsing learner the same topic twice; the
survivor carries both names so the merge loses nothing that anyone searches for.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01TqXevQJhxFrM7jJg82cgZN
**Missing views.** The note said 10 articles without a clinical view and 13
without a short one; the real numbers were 8 and 11, and the eight originals
turned out to carry no `variant` key at all, so everything in them was being
read as the long view. 33 sections written across 11 articles, every sentence
traceable to that article's own body, and the legacy sections stamped with the
variant they were always being treated as. No doses appear anywhere, because
not one of those eight states one. Every article now has a high-yield view and
`mdm_pass check` reads 330 of 331 conforming with no placeholder sections.
**Pediatric Dentistry** was neither an educator's job nor a delete. Its body is
real — scope, disparities, records and consent, public health, biomaterials —
and only its clinical view was the disclaimer. The three placeholder sections
are gone and the article stays: the reader hides a view an article does not
have, and a discipline overview about consent and public health has no bedside
path to write. Inventing one was the thing not to do.
**The 15 tied conditions were stale**, ticked with the evidence rather than
worked: the taxonomy rebuild gave every one of them a single home, and the
duplicate-name query now returns nothing.
**The 31 tied questions were 3**, and reading all three showed none was a real
tie. The best of them: a question about acute flaccid paralysis after minor
trauma was filed between "Nursemaid's Elbow" and "Radial Head Subluxation" —
two names for one condition, neither anything to do with the question. It is
under Spinal Cord Injury now, cross-filed to Acute Flaccid Paralysis, and the
two elbow links are removed rather than cross-filed: a wrong link surfaces a
question under a topic it does not belong to, which is the opposite of what
cross-filing is for.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01TqXevQJhxFrM7jJg82cgZN
A tour is a confession that the product does not explain itself, so this one
names the three things a learner would otherwise go looking for and stops:
the session chosen for them, search against AI Mode, and what the ring is
measuring. Anything needing six steps wants rebuilding rather than narrating.
Steps are declared against a `data-tour` attribute rather than a class or a
ref, because a tour that reaches into another component's internals breaks the
first time somebody renames something for a visual reason. A step whose target
is not on the page is dropped from the count rather than shown pointing at
nothing — an educator and a learner do not see the same dashboard, and "Step 2
of 6" must not promise a step that will never come. The count is fixed when the
tour opens, so a block arriving late cannot renumber it under somebody midway.
Only on the overview, and after the objective rather than beside it: two things
asking for attention at once is one of them being dismissed unread. Seen-ness
is per browser, because a tour is about the interface rather than the account.
The dimming is four rectangles rather than one enormous box-shadow, which would
repaint the whole screen on every scroll event — and this follows scrolling.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01TqXevQJhxFrM7jJg82cgZN
The classifier review is moot: it reviewed a per-question organ-system pick,
and systems hang off the topic now — `question_tag_links` holds no system rows
at all, every one of the 569 topics carries a system, and all 2,922 questions
reach one through their topic. There is nothing left to review.
Admin settings was revamped by the Settings rebuild a day earlier and the list
never caught up. The short view is called high yield in the interface, so the
outstanding item says so.
Two are non-goals rather than work: coach marks over a product somebody is
trying to use are a substitute for the product explaining itself, and clinical
AI Mode belongs to ped-ai — this one should not grow a second answer to a
question that one already answers.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01TqXevQJhxFrM7jJg82cgZN
The voice picker was a dropdown in the quiz player, beside the question — the
one control on that screen with nothing to do with answering it, and one a
learner sets once and never touches. It is a setting now, on the user rather
than in a Redis blob, with a play button beside each voice because a voice is
worth hearing before it is chosen. Choosing nothing stays a real choice: it
means whatever an administrator marked default, so a site that changes its
default reaches everybody without a row being edited.
The tutor reads figures from `question_media` rather than the two legacy path
columns. Those agree exactly today, so nothing was being lost — the first
question given a second figure in the editor would have been the one that
broke it, silently and only for the tutor. The legacy columns remain as a
fallback for anything not projected into that table yet.
And the retrieval thresholds are written down in docs/retrieval-thresholds.md:
the three answers, the sixteen queries they were measured against, why they are
deliberately not the retrieval floor, and how to re-measure when the corpus
grows. Worth keeping the headline in mind — "discuss love" scores 0.491,
alongside "tell me a joke". A number in the 0.4s is noise, not a weak signal.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01TqXevQJhxFrM7jJg82cgZN
**Prepared sessions.** Most of this existed: unanswered first, weakest topic
next, wrong-before-right after that, all scaled by what share of the real paper
each topic carries. What it could not do was change with time, say anything
about itself, or be reached without filling in a form.
Evidence now decays on a thirty-day half-life. Exponential rather than a fixed
window because memory has a slope, not a cliff — under a window, 29 days counts
fully and 31 counts for nothing — and because it is memoryless, so an answer's
weight does not shift when unrelated questions are answered, which is what lets
the preview stay a valid forecast. Spring is worth an eighth of last week. Two
things decay: a question's recall probability, drifting towards even rather
than past it, so an old right answer becomes eligible rather than wrong; and a
topic's accuracy, against a prior of two "no idea" answers, which fixes "right
once, known forever".
Strict unanswered-first meant that on a bank of 2,900 nothing was ever
recycled — spaced repetition existed and was unreachable. Review now takes up
to two fifths of a session. And the damping that spread the picks across topics
was applied only to seen material, so a learner with no history was handed the
heaviest domain entire instead of a spread; that was live.
The plan is the product. It is computed, shown, and then the session is built
from that plan's own ids and the plan returned with it, so the two cannot
differ; every figure in it is a tally over the chosen questions rather than a
forecast. No model touches the ranking — a learner asking "why these twenty"
has to get the same answer twice.
**Vision.** The proxy's own `/model/info` says which models can see, so nothing
is hard-coded: 77 report yes, 11 no, and 328 say nothing at all, which means
absent rather than incapable — so those are asked once with an 8px PNG and the
refusal cached. The deployment's main model turns out not to see, and questions
carry figures the learner is looking at, so the tutor was answering about an
image it had never been shown. It routes to a configured tool model now, folds
the description back in as text saying plainly where it came from, and caches
on the bytes because the same figure is re-sent every turn.
Also fixed on the way: `article` was missing from the admin's task list, so
article drafting always ran on the fallback model whatever an administrator
chose; and `.jpx` stem images were sent as JPEG because `mimetypes` guesses
that from the name, so the provider rejected them two hops later.
An administrator must pick a tool model in Settings → AI models. Until then the
tutor says a figure exists that nothing could read, rather than describing one
it cannot see.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01TqXevQJhxFrM7jJg82cgZN
A question's stem image is two to four megabytes of scanned radiograph, and a
media grid is forty of those pulled at full size to draw forty postage stamps.
`?w=256` and `?w=640` now serve a WebP copy instead, made on the first ask and
kept beside the original under `thumbs/{width}/{key}` — same bucket, so nothing
new has to be configured for them to be backed up or thrown away.
Three rules, all about not making this a way to spend the server's afternoon.
Those two widths and no others: any other `?w=` is refused with a 400, because
an endpoint that resizes to whatever the query string asks for is a CPU sink
anybody can point at. Never enlarged: a 180px image asked for at 640 is served
as it is, since scaling up invents detail and charges bytes for it. And best
effort throughout — a PDF, an SVG, a truncated upload or a file that is not the
image its name claims all serve their original rather than failing, because a
preview must never take down the page that wanted it.
Authorisation is unchanged and still runs first: a thumbnail of a file you may
not read is a file you may not read. They stay `private, no-store` like
everything else here — they are behind authentication, so there is nothing for
a shared cache to do with them, and the win is the byte count.
EXIF rotation is read before anything measures the image. Every phone stores a
portrait photograph sideways with a flag; a thumbnail made without reading it
is a sideways thumbnail.
Pillow rather than sharp, which is Node. It is not pinned in requirements: the
pin invalidates the pip layer, and that layer no longer builds because
litellm==1.28.13 has been withdrawn from PyPI. Re-pinning litellm is a
deliberate upgrade of the AI layer, not something to slip into this. Noted in
the TODO.
Also: the article hover-card excerpt was printing `[[288|eczema]]` at readers.
The generic markdown-link rule does not know our own cross-reference syntax, so
it left the brackets and the id behind.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01TqXevQJhxFrM7jJg82cgZN
Nothing had imported it since editing moved to the full page; the only thing
that still referred to it was its own test.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01TqXevQJhxFrM7jJg82cgZN
Ctrl+K anywhere. Searching and asking are the same act from where the
learner sits — they have a question and want it answered — so which one
happens is a toggle rather than two destinations to choose between before
typing. It opens over whatever is on screen, because a question arrives
while you are reading something and having to leave that page to ask is
how the question gets dropped.
Search keeps its furniture: suggestions, the history of what was asked,
and the keys spelled out. Space fills the box with a suggestion rather
than following it, so a near-miss is edited instead of retyped. AI Mode
drops all of that — a list of half-remembered searches is not help while
writing a question — and gains a clear button and a submit arrow.
Submitting in AI Mode does not show results. It hands the question to the
conversation at /ai?ask=, which is where the follow-up will be asked from
anyway.
The header's search box becomes the door to it rather than a second box
beside it: two fields that do nearly the same thing is how people learn
to trust neither.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01TqXevQJhxFrM7jJg82cgZN
landing page describes this product
An objective is now required to build a session, not only asked for in
the interface — the interface asks, and this is the same rule where it
cannot be walked past. Only where there is something to choose: a
deployment with no exams, and the first administrator of a fresh one,
must still be able to build a session. A rule that locks an empty site is
not a rule, it is a fault.
Saving a question into a folder is one box that searches what you have
and offers to make what you do not. It used to say "make one in the
question bank" and leave you to go and do it, which means leaving the
question you were reading and coming back to find your place. A name
that already exists exactly is not offered twice; a partial match offers
both, because wanting a narrower folder called "cardio" is not the same
as wanting the one called "Cardiology misses".
The landing page is rebuilt. Its copy described a product from months ago
— "upload a PDF, AI extracts questions", which is one feature of many now
— and it was 568 lines of inline style objects, which cannot express a
hover, a media query or a keyframe. The figures come from
/api/public/stats and count up; a failed fetch renders the section
without them rather than showing noughts, which would be a lie about an
empty bank. Motion is CSS and SVG, and prefers-reduced-motion turns all
of it off — including forcing the scroll-revealed elements visible,
since a hidden element with its animation removed is how respecting that
setting turns into a blank page.
Two smaller ones from the screenshots: the collections shelf is boxed
rather than scrolling past everything else on the page, and its rows no
longer carry the entire stem — lab tables and all — in a native tooltip
that covered half the screen and could not be dismissed.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01TqXevQJhxFrM7jJg82cgZN
Proof-of-work rather than a puzzle, and — the reason for it — nothing
about the person signing up is described to a third party in order to let
them in. Turnstile and then hCaptcha were both here; both told Cloudflare
who was at the door.
The `cap` service runs on the compose network with its own Redis
database, kept apart from the app's so a flush of one cannot clear the
other's challenges. The widget talks to /cap/ on this origin, proxied by
the frontend's nginx, so the browser reaches nobody else either. Caddy
passes the whole host through to that container, so it needed no change.
Two things that had to be found rather than read:
Cap's key API is undocumented. The routes are `/auth/login` and
`/server/keys`, and the Bearer value is base64 JSON of `{token, hash}` —
not the session token itself, which is why the obvious call returns
"Malformed session token". The site key and secret were created that way
rather than by hand in a dashboard.
And an nginx proxy_pass whose target is a variable passes the URI through
untouched: the trailing slash that strips a location prefix on a literal
target does nothing. Cap was being asked for /cap/<key>/challenge and
answering NOT_FOUND until the prefix was stripped by an explicit rewrite.
Verified end to end against the running service: a challenge is issued
through the public path, and a token that was never issued is refused
rather than waved through.
Also here: the register modal's Name and Email were bare labels that
neither wrapped their input nor named it, so a screen reader met two
boxes with no names and clicking the word did nothing.
And the knowledge profile paginates ten to a page and expands each row to
its two bars beside the next step. "Correct using hints" is missing from
that bar because /study-tools/recommendations does not carry it per
topic — inferring it from the lifetime figure would be a different set of
answers, so the bar is honestly two-tone until the backend offers it.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01TqXevQJhxFrM7jJg82cgZN
One verifier, backend/app/services/captcha.py, and one widget,
components/Captcha.jsx. There were two copies of each and they had
drifted: the register widget loaded the script itself while the landing
one relied on a page-level effect elsewhere in its file, and on the
backend auth failed *open* on an unreachable Turnstile while contact
failed *shut*.
Both failure modes were kept rather than one quietly chosen, as an
explicit `fail_open` argument with the reason written down: an outage that
stops people creating accounts costs the site its users, while an outage
that bounces a contact message costs the sender one retry.
An unconfigured secret still skips verification entirely, as before, so a
site with no keys keeps working.
The keys in .env are empty. The Cloudflare ones there were live and are
now dead, so **there is no captcha on register or contact until hCaptcha
keys are issued** — this is not a state to leave a public site in.
Also corrected on the way: docs/frontend.md still documented
`login(email, password, turnstileToken)`, whose third argument had
already gone from AuthContext.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01TqXevQJhxFrM7jJg82cgZN
analysis is a real tab
Adaptive selection knew what you were weak at and nothing about what the
exam is made of, so being weak at something worth 5% of the paper ranked
the same as being weak at something worth 1%. Every score is now
multiplied by the weight the board publishes for that topic's domain —
the same `exam_blueprints.weight` behind the Relevance column.
A topic the blueprint does not cover takes the median published weight. A
zero would make unmapped material unreachable and the highest would make
it the priority; neither is a claim the blueprint supports. With no study
objective the multiplier is absent and selection is about weakness alone,
exactly as before.
Weight scales weakness, it does not replace it: a topic you are certain of
does not surface because it is worth 5% of the paper, because (1 −
accuracy) is near zero and no multiplier rescues that. docs/adaptive-
sessions.md says all of this, including what is still open.
Session analysis is the third tab rather than a link out of the page —
two of the three used to change what you were reading and the third took
you somewhere else. The tab bar is one component both routes wear,
AnalysisSessionPage's body is a component the tab renders in place, and
the tab lives in the address so a link opens where it says.
Two things that were wrong turned up in that work: a session nobody had
sat showed 0% in the figures and "0% correct" in the donut — two separate
statements of a score on a session that had none — and the old third tab
disappeared entirely for anyone with no attempts, so the strip silently
changed shape.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01TqXevQJhxFrM7jJg82cgZN
written down
The yellow was a gradient stripe under the x-height, which reads as an
underline rather than a highlight and all but vanishes on a wrapped line.
It is a solid band now, the way a highlighter leaves one.
A block that ends without warning ends on whatever question you happened
to be reading. Five minutes out it says so — once, because a warning that
returns every second is a warning nobody reads.
And docs/adaptive-sessions.md, because "prioritised by impact" was a
phrase with no written meaning. It says where the code is, what the three
rules are — unanswered first, weakest topic among those, then wrong ones
oldest first with the category damped so twenty questions do not all come
from your worst subject — and, as plainly, what it does not do: it is not
weighted by the exam blueprint. Weakest and most-of-the-paper are
different questions and we answer only the first. The refinement that
would combine them is written down too, with the column that already
holds the weights.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01TqXevQJhxFrM7jJg82cgZN
Favorites and the question libraries in one place. Card and Table views
with the choice remembered, sort by last used / created / name / size
with a direction control, a count line, and a search over name and date.
Favorites leads as a fixed row: it is the one shelf nobody made and
everybody has, so it cannot be renamed or deleted.
Sorted by when each was last used, not when it was made — the order
things were created in is nobody's mental model of their own shelf. A
library nobody has opened falls back to its age, because it is newer to
the learner than it is to the database. That needed
`user_collections.last_used_at`: null on every existing row, since
backfilling from created_at would invent a use that never happened.
A shelf opens in place rather than linking away. The obvious link would
have been /questions?collection=N, and there is no page there that reads
it — the old bank browser was dismantled — so the card would have led
nowhere. Questions can be taken back out from the open shelf, and any
shelf can be sat as a session through the existing explicit_ids builder.
The ⋯ menu moved out of QuizPage into components/MoreMenu; the player
keeps its own look and its own children through className props. It no
longer closes on any click inside, which the player's feedback form and
share dialog were relying on by accident.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01TqXevQJhxFrM7jJg82cgZN
The session analysis ranked its weakest topics by primary category only,
while the Analysis page asked the same question three ways and rolled
answers up the category tree. Two sets of rules for "where does this
question belong" is two pages that can disagree about a learner and
neither able to explain why.
So the rules moved to services/knowledge_groups.py: ancestor roll-up,
article reached through its category, organ system reached through the
symptom keyword. study_tools now asks that service instead of building
the lookups inline, and GET /attempts/{id}/recommendations gives one
session the same Articles / Disciplines / Systems switch. Grouping is its
own call, so changing it does not re-read the question table and the peer
statistics beside it. A running exam ranks nothing — marking it there
would answer the question the exam is asking.
The ungrouped `recommendations` key is gone from the analysis payload
along with the code that built it.
And the document page had no way back. It is reached from the Tools
workbench, which by design has no menu of its own, so leaving it meant
the browser button. It opens onto Tools now, as Tools opens onto
Settings.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01TqXevQJhxFrM7jJg82cgZN
Your score is the share of questions right at your most recent answer to
each. It is deliberately not called an equated score: AMBOSS's EPC rests
on psychometrics we do not have, and a number dressed up as one would be
a claim we cannot support. The card says so.
Against everyone else compares you with other learners on the questions
you have in common — not with their scores on whatever they happened to
sit. A percentile over different question sets reads someone who worked
through the hardest fifty in the bank as weaker than someone who did
fifty easy ones, which is the opposite of true.
Neither appears before it means anything, and each says which half is
missing: more questions of your own, more questions shared with others,
or more learners. The cohort reported is the most any one shared question
saw — distinct learners cannot be summed across questions without
counting the same person once per question.
The "readiness is still locked" note sat above the tab switch and so
appeared on Performance, where it described a table that is on the other
tab. Moved down to the table it is about.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01TqXevQJhxFrM7jJg82cgZN
Uploading, extracting and reviewing were three places and the middle one
wrote straight into the bank. They are one job, so they are one page:
the documents, the batches a run produced, and the drafts themselves,
with the counts of what is still to be read said before anything else.
The review is an editor, because that is what the staging table is for.
A draft that is not ready says why on its own row rather than on a
failed attempt to accept it, and cannot be selected until it is fixed.
Fixing happens in place, with nothing in the bank. Only the chosen,
ready ones cross over, and that is the single moment a question id is
taken — an accepted draft then says which question it became.
Nextcloud is the administrator's. It was in every learner's settings as
though each of them had a cloud to connect; it is an import path for
whoever loads the corpus.
The image work is restated in the TODO now that the ped-ai design has
actually been read rather than remembered: allow-listed widths only,
EXIF rotate, resize without enlarging, WebP q82, derivatives in the same
bucket under a thumbs/ prefix, generated best-effort. Pillow rather than
sharp — sharp is Node and this backend is Python — and no Caddy caching,
because these are behind auth and ped-ai serves its own private,
no-store for that reason.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01TqXevQJhxFrM7jJg82cgZN
Exit asked "Suspend quiz?" and offered Stay. Nobody presses Exit by
accident, nothing is lost — the answers are saved and the clock pauses,
which the dialog was explaining rather than deciding — and a
confirmation for a reversible act is a step, not a safeguard. One press
now, and if the save fails it stays put and says so, because leaving
then would lose the answers.
Resume and Start go straight in for the same reason. The overview that
asks whether you meant it belongs to opening a link to a session you
have not seen, not to a button you came to this page and chose.
The actions moved out of the page heading and into the card that holds
the result — Repeat session, Review answers, Resume session — because a
row of buttons beside the title puts the decision as far from the
figures it follows from as the layout allows.
Delete session stays, but not as a red button next to Resume. It throws
away answers the analysis is built from, and it is the one thing here
AMBOSS has no need of — sessions are made freely, so a mis-made one is
clutter worth removing. It is a quiet control at the end of the row.
The session rail was sticky at top:0 with height:100vh, which started it
underneath the 98px header and ran it past the bottom of the window, so
the end of the list and the scroll that would reach it were never on
screen. It measures from --app-header now, and a wheel over the list
moves the list rather than carrying on into the page.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01TqXevQJhxFrM7jJg82cgZN
Settings linked to a second dashboard with its own tab bar and its own
visual language. The admin sections are rendered in Settings now, under
headings that say who they are for — You, Content, The site — and each
has its own address, so People, AI models, Safety and Search are links.
/admin redirects into Settings for anyone who bookmarked it. AdminPage
takes a `section` prop and drops its tab row when embedded; it is loaded
lazily, so it is not in a learner's download.
Comments are gone: router, model, table and the half of the test file
that covered them. They were a discussion thread nobody was obliged to
answer, and feedback replaced them with a message addressed to whoever
maintains the question. The table was empty, so nothing was lost —
verified before dropping it.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01TqXevQJhxFrM7jJg82cgZN
The quiz player is a box the height of the window. The question used to
scroll the whole page, which took the session rail and the navigation off
screen exactly when you wanted them; now each column scrolls on its own
and the bar — Exit session, Previous, Next, Review — stays put.
Two site-wide switches, together under Settings → Site policy because
both are the administrator's and both apply to everyone:
* Sharing can be turned off. That stops new links being made; one
already handed to somebody keeps working, since revoking it would
break something a learner has already given away.
* Sign-up can be made invite-only, with single-use codes carrying a
note of who each is for and, afterwards, who it let in. A spent code
is kept rather than deleted — that record is the point of invite-only.
The alphabet has no O/0 or I/1/l, because these get read aloud.
The registration form asks for a code only when the site needs one, via
an unauthenticated policy endpoint — it has to know before there is an
account to ask with. It never says whether a given code is valid before
the account exists, which would make it somewhere to guess them. The
first account is always allowed, or a new install would lock itself out
before an administrator existed to issue a code.
Flags fall back to their defaults when Redis is down, in the safe
direction each way: sharing keeps working, sign-up does not silently
open.
Found on the way: the registration form's three labels named nothing —
no `for`, no wrapping — so a screen reader announced unlabelled boxes.
Backend 261/261, frontend 328/328.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01TqXevQJhxFrM7jJg82cgZN
The study objective was a <select> capped at 220px, so it clipped
"Pediatrics Boards (2948)" mid-word, and it offered "USMLE Step 2 CK (0)"
as though it were ready — choosing it would have scoped the bank to
nothing. It is a dialog now: grouped by family, searchable, each option
saying how many questions stand behind it, an objective with none shown
but not selectable, and the change confirmed rather than applied the
instant the pointer crosses an option.
The section bar outran its width because four of its entries were not
about the material. Home is the signed-out landing page and is gone from
it; Dashboard, Account, Settings and signing out are about the person
and now sit behind one account control. Nine entries instead of
thirteen, which is the actual fix — the arrows added earlier make a long
bar usable, they do not make it right.
Frontend 320/320.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01TqXevQJhxFrM7jJg82cgZN
Three things asked for.
New question opened a cramped modal in both places it was offered — the
question bank and the question manager — while Edit opened the full
page. Both now go to /questions/new, carrying where they came from so
the back link returns there. CreateQuestionModal had no callers left and
is gone.
The bank's "Study" action is "Preview", and shows the question whole:
the correct option, the per-option reasoning, the explanation and the
key points, all at once. Making someone answer first is the right shape
for practice and the wrong one in the bank, where the question is being
inspected rather than sat. Option reasoning and the explanation render
as Markdown there now too, instead of raw text.
Option explanations get the same formatting toolbar as the stem — an
option's reasoning is prose as well, and often carries a list or table.
Frontend 309/309.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01TqXevQJhxFrM7jJg82cgZN
323 of 323. It was 413 free-form sections under 242 different titles,
most of them one block called "Management" holding everything from the
presenting complaint to discharge advice. They now read Clinical paths →
Diagnosis → Management, with Prognosis and outcome where the source
actually said something about course — 95 articles earned one, the rest
were left without rather than padded.
Six batches, and the gate earned its keep: it caught a dropped
`[[206|epilepsy]]`, a duplicated "24", a repeated sentence that doubled
two numbers, and several numbered lists flattened into bullets. Two
articles were refused on a first import and passed on a second — I had
imported a file its agent was still correcting.
Then the linker re-run over the moved text: 2,150 links, 0 broken, 0
nested, 0 escaped. Instructions to the reader across the whole corpus
are down to 21 from 819 + 71.
Backend 242/242.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01TqXevQJhxFrM7jJg82cgZN
Correcting the note this was tracked under: "114 tied conditions need a
human call" counted duplicated *names*, not ties. 99 of them had a clear
plurality and were merged; 15 are genuine even splits.
Same script and same safety property as the earlier pass — every moved
question gains a question_category_links row to the system folded away,
so filtering by that system still finds it. 786 categories to 662, 124
rows removed. Checked after: 0 questions with no category at all, 0
links pointing at a deleted row, and Pneumonia's questions still reach
Pulmonology and Allergy/Immunology.
The remaining 15 are written up in docs/tied-conditions.md as a decision
rather than left as a task: what each is tied between, and the note that
the choice only sets where a condition lives, since the other systems
survive as cross-links either way.
Backend 242/242.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01TqXevQJhxFrM7jJg82cgZN
'1/1 article · 50/50 questions' beside each block, the way a course
module reports itself. Counts rather than a percentage: '0/6 articles'
is something you can act on, '0%' only tells you how to feel about it.
The state stays alongside for the one-word answer.
Frontend 308/308.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01TqXevQJhxFrM7jJg82cgZN
"How do I bold? add a list etc?" — answered, and the interesting part is
what it is not.
Milkdown was the obvious choice; it is already in the project. Round-
tripping a question stem through it first showed why not: bullets come
back as `*` with blank lines inserted between them, tables are repadded,
and anything it does not recognise is escaped. The first two reflow text
nobody edited — and learners' highlights are stored as character offsets
into that exact string, so a reflow on any save moves every one of them.
So the text stays byte-for-byte as typed and the buttons only insert
syntax at the cursor: bold, italic, code, bullet and numbered lists
(every selected line, and already-bulleted lines left alone), heading,
a table skeleton, inline maths. The Markdown preview that was already
there shows the result.
Frontend 308/308.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01TqXevQJhxFrM7jJg82cgZN
Per-question notes were already built and saved on blur. The global
notes tab was still floating over the same screen beside them, so it was
never clear which notepad a note was going into. The global note stays
where it belongs, on the dashboard.
The quiz test mock had no `put`, which is why nothing had ever covered
the per-question note path; it does now.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01TqXevQJhxFrM7jJg82cgZN
Four time bands rather than three — someone revising at 2am is not
having an evening — and the wording is keyed on the date, so it differs
when you come back and holds steady while you are here. A line that
changes under the reader on re-render looks like a glitch, not warmth.
Also closed: "References with in-text superscript markers". That was on
the list against an explicit instruction — "with refernec, but you dont
need in text reference". Leaving it open would have meant building it
eventually.
Frontend 293/293.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01TqXevQJhxFrM7jJg82cgZN
Answers the two questions owed: the readiness shrinkage, the three
groupings, the priority ranking, and the four steps of adaptive
selection — including where it is weaker than it looks.
Writing it up surfaced two defects, both fixed here:
* adaptive_select took the first 2,000 candidate rows. The bank is
2,948, so about a third of it could never be selected, and which
third depended on database order. The cap is gone; two integer
columns per question is not a size worth protecting against.
* category lookup was a linear scan through every candidate for every
recorded answer — O(answers x candidates), the slowest part of
building a session. It is a dict now.
Left alone and documented instead, because changing them changes which
questions a learner is given and that is not a silent decision: adaptive
ordering uses raw category accuracy rather than the shrunk readiness the
recommendations page uses, and difficulty is a filter rather than
something the session moves along.
Backend 231/231.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01TqXevQJhxFrM7jJg82cgZN
An unsuspended exam keeps running. When its clock runs out it is
submitted with what was answered and the score counts — a learner who
ran out of time sat an exam, which is a result and not an accident to
hide. Previously it was graded, flagged expired=1, excluded from every
statistic, and the client was told the opposite ("submit manually").
- attempt_expiry.settle_if_expired: one path, used by resume and by the
sessions list, so an exam left open elsewhere shows its score rather
than "in progress" forever. Suspended attempts hold their clock and
never expire.
- resume returns {expired_submitted, attempt_id}; the client opens the
analysis. The suspend dialog and the leave warning now say what
actually happens.
- delete: saved progress and device lock cleared; a study-plan block
whose only completed attempt is deleted goes back to unfinished.
- POST /attempts/reset-all: typed RESET, removes attempts, answers,
in-progress state, plan progress, reading marks, saved questions and
question notes; leaves the account, authored content and AI chats.
Settings → Your data, with the counts reported afterwards.
Also fixed on the way: the first version of the sessions-list change
mutated the dict it was iterating; the test only passed because it had
one attempt. Now two.
Backend 223/223, frontend 258/258.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01TqXevQJhxFrM7jJg82cgZN
The linking was the gap
The marker system was built weeks ago — resolves by id, survives a rename, shows
a preview on hover — and not one of 333 articles used it. Every article was
written in isolation, so a piece on croup named stridor and epiglottitis and
offered no way to reach either. `scripts/link_articles.py` reads what is written
and links it: 3,718 cross-references across 307 articles, by id, so a later
rename cannot break them.
Conservative on purpose, because a wrong link is worse than a missing one: only
the first mention in a section, whole words, longest title first so "Otitis media
with effusion" beats "Otitis media", never inside an existing link, marker,
heading, code span or table, and never an article to itself.
That exposed a second thing: the reading view had its own Markdown pipeline with
its own cross-reference regex, and it only understood the old slug form. It would
have printed every one of those 3,718 links as literal brackets. Article prose
now goes through the same renderer as the rest of the site.
Short and Clinical looked empty
Both are usually a single section, and everything starts collapsed, so the tab
showed one heading over blank space. A view of one section is not a contents
page; it opens.
Removed
Quiz reminders — emailed nudges to retake anything under 75%, with a scheduler
that existed solely to send them: the model, the service, the scheduler, the
email, the table. Article comments. The dashboard's in-progress list and its
stat cards, both of which the analysis page now answers better.
One mistake worth recording: the first pass at removing the reminder cleanup used
a regex that took 109 lines with it, including an unrelated endpoint. The test
suite caught it (`/attempts/quiz/{id}/in-progress` returning 404 instead of 403),
and the file was restored and edited by exact match instead.
208 backend, 249 frontend green.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01TqXevQJhxFrM7jJg82cgZN
Every link from the history went to /quizzes/:id, which auto-started. Clicking
"Board Review IX" out of curiosity began a 243-question exam, clock running.
That is why the links looked broken: they worked exactly as built, and what they
did was wrong.
Landing on a session now shows it — mode, length, and what the clock does — and
starts when asked. The session rows and the analysis rail open the session's own
analysis rather than the raw answer list, which is what a rail of past sessions
is for.
Sessions carry no category, so the row no longer prints one; that field was
still showing the retired "PREP" label.
Twenty tests asserted the old behaviour by mounting the player and waiting for
the stem. They now start the session the way a person does, which is the point:
the assertion was encoding a bug.
Settings is on the list. Before restyling it, the question is what belongs there
at all — several things have drifted in because there was nowhere else to put
them.
249 frontend tests green.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01TqXevQJhxFrM7jJg82cgZN
Figures
A question could carry exactly one stem image and one explanation image, each a
bare path with no title, no legend, and no way for the prose to refer to it.
`question_media` makes a figure a row: it points at an image already in the bank,
carries a role, a label the text can name ("Figure 1"), a caption and an order,
and there can be as many as the question needs. The same radiograph can serve two
questions without being stored twice.
The 346 existing paths were backfilled into figure records and retitled —
`page_339_img_0.png` says where a file came from and nothing about what it shows,
so the filename moved into the caption where it is still searchable, and the
title became something a person can read.
On the editor question: no new platform needed. Milkdown is already installed —
ProseMirror-based, MIT, GFM tables, code blocks, LaTeX — and already used for
articles, courses and the quick question modal. Only the question *page* still
has plain textareas, and that swap is written down rather than rushed, because
the stem carries manual-highlight offsets and a WYSIWYG rewrite would move them.
Fewer hints during a quiz
The category trail and the difficulty pill were shown beside every stem. Being
told a question is filed under Neonatology, or that it is "hard", narrows the
answer before the stem has been read. Both now wait until the answer is in,
where the trail becomes a way to more of the same topic.
The dashboard is about questions
Quizzes and attempts describe how the material happens to be packaged. What a
learner is working through is questions: how many of the bank they have seen,
how many they have answered correctly, and their average. The old per-quiz
performance card — which needed two attempts before it showed anything — is
gone, superseded by the session analysis. The greeting sits above "continue your
study" rather than below it, where it read as a heading for the wrong section.
208 backend, 249 frontend green.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01TqXevQJhxFrM7jJg82cgZN
The results page showed a score and a wall of explanations. What a learner needs
afterwards is where the time went and what to go back to, so
/analysis/session/:attemptId gives them: a rail of recent sessions, the four
figures they act on — correct, completed, time per question, total time — a
donut, the weakest topics, and a paginated table of every question with its
status, difficulty, time and how peers did on it.
Time per question was not recorded at all, so it could not be reported. It is
now (`attempt_answers.seconds_spent`), banked when you leave a question and
including the one still open at submission — without that the last question of
every session would show nothing. Answers from before this read "—" rather than
claiming zero, and a question nobody else has answered has no peer rate rather
than 0%, which would read as everyone having failed it.
Also in this pass, from the review:
* quiz categories are gone from the library — a second taxonomy beside the
real one, putting a heading above every test;
* the board review sets are numbered rather than dated, in both the quizzes
and the study plans built from the same material, so a learner does not meet
2019 in one place and VII in another;
* the footer's standing note is one clause, and the gap above it no longer
looks like the page ended early.
Everything else asked for today is written down in docs/TODO.md rather than
half-built: resume instead of restart, an unsuspended exam that keeps running,
deleting a session's data, reset-all-data with a warning, recommendations split
by article/discipline/system, and the adaptive session. Two questions I owe
answers to are in there too.
208 backend, 249 frontend green.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01TqXevQJhxFrM7jJg82cgZN
The design settled earlier, built as described: retrieval decides what may be
cited, and the server enforces it.
The model is handed a shortlist of at most fourteen sources from the learner's
own library and told to cite them by marker. Afterwards every citation it wrote
is checked against that shortlist and anything else is deleted before it is
stored or shown. A hallucinated citation is not unlikely here, it is impossible
— surviving is not a decision the model gets to make. A URL it invents is not a
citation either: only the marker form counts, so a plausible-looking link stays
in the prose citing nothing.
Retrieval reuses the hybrid search already in place, and each corpus keeps its
own visibility rules — the bank predicate and exam scope for questions, the
draft rule for articles, deck ownership for cards. A question source carries the
stem only: a chat that printed the answer would hand away the practice it exists
to prepare you for.
Curated links do the job they were built for. A retrieved row an educator tied
to another retrieved row is boosted, because two things somebody already linked
surfacing for one query is evidence rather than coincidence. Nothing is stored
for this; the boost lives only in that ordering, and the answer marks those
sources so the reader knows which claim rests on an educator's judgement rather
than on a ranking.
Citations are stored with the answer as filtered, so reopening a thread shows
the links it showed at the time rather than a fresh retrieval that may now rank
differently. In the page the markers become numbers and each number opens its
source; a section citation deep-links into that section.
Two smaller decisions worth naming: a question appears in the thread the moment
you send it and is handed back to the input if the answer fails, because typed
words are not something to lose on a 502; and someone else's thread returns 404
rather than 403, since whether it exists is not your business either.
182 backend, 206 frontend green — 16 of the backend tests are the citation
contract and the retrieval boundary.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01TqXevQJhxFrM7jJg82cgZN
