The pathway existed but was reachable only by API. It now has a tab of its own
next to the Learning Hub — related, not the same thing, and sitting together is
how someone discovers the difference — visible to every signed-in user with no
role gate in the markup.
Generate a deck or an article, see everything you have made, download each as
PowerPoint, Word or PDF, delete what you no longer want. The screen says
"Private to you" and "Nobody else sees these", because the distinction from
published Learning content is the thing a person needs to understand before
typing a patient's condition into it.
Downloads are fetched rather than linked: an <a href> cannot carry the
Authorization header. The blob is saved under the filename the server chose and
the object URL is revoked afterwards. Resource titles come from a model, so rows
are built as elements and a title is only ever assigned to textContent.
The e2e stack now joins danvics_convert too. It could previously reach only
Postgres and Redis, so a PDF download failed there in a way production would
not — which did at least prove the degradation path works: with Gotenberg
unreachable the response is "PDF conversion is unavailable right now. PowerPoint
and Word still work", and the other two formats download unaffected.
Verified in a browser as an ordinary user: the tab appears and opens, the form
swaps slide count for word count when the format changes, the library lists
their own work, and pptx, docx and pdf all download with sensible filenames
(36360, 13285 and 68310 bytes).
Also documents retrieval sizing in docs/retrieval-tuning.md — the per-feature
budgets, and RERANKER_TOP_K, which caps all of them and had until now appeared
in no configuration file at all.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Dv6sqaY6Vq3ChZHMem3cnU
Learning is moderator-owned: content published into categories that everyone
sees, behind router.use(moderatorMiddleware). That is right for institutional
material and wrong as the only way in — an ordinary user could not generate
anything at all.
So this is a separate pathway rather than a loosening of that one. Learning is
untouched; the moderator gate stays exactly where it was. A signed-in user can
generate a deck or an article for their own use, keep it, refine it and export
it, and nobody else ever sees it.
Private by construction. Every statement filters on the owner and there is no
route that returns another person's work, which a test asserts statement by
statement rather than trusting. The foreign key cascades, so deleting an account
takes its drafts with it. There is no category, no publish state and no sharing:
adding sharing later should be a deliberate feature, not something that leaks
out of a forgotten WHERE clause.
Markdown is the artifact. Every format is rendered from it on demand — pptx and
docx by pandoc, both carrying the house reference deck, and PDF by Gotenberg,
whose LibreOffice preserves a deck's layout in a way rendering from markdown
would not. That is what makes "add a slide on when to admit" a text edit rather
than a binary patch.
Gotenberg was published on the host but on a network of its own, so reaching it
from a container went out and back through the host gateway. It now joins
danvics_convert, owned by danvics-net like the others. PDF is the one export
allowed to fail: if that service is down, the deck and the document still
download and the error says which.
Verified end to end as a plain user: the moderator route still refuses with 403,
generation returned a deck grounded on 12 corpus excerpts, the library lists
only their own, pptx/docx/pdf all downloaded valid, "add a slide on when to
admit" put the slide in the right place and left References last, and an
unauthenticated request gets 401 while someone else's id gets 404.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Dv6sqaY6Vq3ChZHMem3cnU
The route accepted useCorpus and reported grounding, but nothing in the admin
screen sent the flag or showed the result — so the feature existed and was
unreachable.
Opt-out in the UI rather than opt-in. For clinical teaching the library is
nearly always the right source, so someone who never notices the checkbox
should get the grounded version. The help text explains when to turn it OFF,
which is the non-obvious case: a topic the library does not cover is better
written without grounding than padded with the nearest unrelated excerpts.
Afterwards it says what happened — "Written from 12 library excerpts", or "Not
grounded — nothing indexed matched. Written from the model alone." Ungrounded
material presented as grounded is the failure worth preventing here, so the
wording never implies the library was used when it was not.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Dv6sqaY6Vq3ChZHMem3cnU
A slide carrying [1] markers is unreadable from the back of a room, and an
article that cites inline reads as a paper rather than as teaching material. The
model is now told explicitly not to cite in the body — no bracketed numbers, no
parenthetical "(Nelson, p. 2604)" inside sentences — and to put everything it
drew on in a References section at the end, which in a presentation is the final
slide.
Checked rather than assumed: a six-slide deck generated through the grounded
path contains zero in-text citation markers, and ends with a References slide.
The prose keeps the specificity that grounding is for — bilirubin produced at
two to three times the adult rate, conjugation immature until about two weeks,
thresholds in mg/dL — without a single marker interrupting it.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Dv6sqaY6Vq3ChZHMem3cnU
The grounding metadata went to the API response and the logs, which is no use to
someone holding the deck. A teaching resource shown to trainees should carry its
own provenance, so a grounded one now ends with a References section — the final
slide in a presentation — listing the library excerpts it actually used, by
title and page.
Restricted deliberately: only excerpts actually drawn on, nothing invented. That
was worth checking rather than trusting. Generated a deck and compared every
citation against the source metadata: "Kliegman R. Nelson Textbook of
Pediatrics, 22nd ed., 2024, p. 2604" against a stored title of "Kliegman R.
Nelson Textbook of Pediatrics 2-Volume Set 22ed 2024" at page 2604, and the same
for Fleisher & Ludwig, Rosen's, Understanding Pathophysiology and the AAP
compendium. The model reformatted filename-derived titles into readable
citations using only what it was given.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Dv6sqaY6Vq3ChZHMem3cnU
Learning generated everything from the model alone. A deck on bronchiolitis was
whatever the model remembered about bronchiolitis, with no connection to the
documents this institution actually indexed — while the assistant had been
searching that corpus all along.
Same collection, deliberately. mcp_bge_m3_1024 is already embedded with
openrouter-bge-m3 at 1024 dimensions; a second index over the same documents
with the same embedder would be a copy that drifts. What differs is the budget:
a chat answer wants a few tight excerpts because the reader is waiting, a
teaching resource synthesises a whole topic. So learning.search_limit and
learning.context_chars default to 30 and 2500 against the assistant's 8 and
1400, and are separate keys so tuning one cannot move the other.
Not unbounded, though. "No limit" only moves the ceiling from a setting to the
model's context window, where overflow truncates the middle of the prompt
silently — the worst place to lose source material. 60 results and 8000
characters per excerpt are the caps.
Opt in per generation: a resource on something the library does not cover is
better written without it than padded with the nearest unrelated excerpts.
Retrieval never fails a generation — the resource is then written from the model
alone, which is what happened before this existed — and every response reports
what it was grounded on, so a caller can say "24 excerpts" or "the library had
nothing on this" rather than quietly serving ungrounded material.
Verified against the live corpus: bronchiolitis, neonatal jaundice and febrile
seizure each returned 12 excerpts and ~23k characters from Nelson, Rudolph and
the Pediatric Clinical Practice Guidelines. A deck generated through the full
chain came back with textbook specificity that is not general recall —
bronchiolar diameter, birth-weight thresholds, the full pathogen list.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Dv6sqaY6Vq3ChZHMem3cnU
npm audit reports 0 vulnerabilities. It reported 2 high and 2 moderate this
morning.
@google-cloud/vertexai was the last source of findings — gaxios and a uuid with
a missing buffer bounds check, neither reachable in this deployment because
GOOGLE_VERTEX_PROJECT is unset and the require sits inside that check. Dormant
is not the same as gone, and the provider is available through the gateway
anyway, so the direct path has been removed rather than left to rot:
- the SDK client and callVertex, which without the package could never run
- the dispatch and discovery branches that reached them
- VERTEX_MODELS, a list of ids nothing could route any more, and the two
places in adminConfig that concatenated it into the built-in set
- the health endpoint's vertex line, and the env vars documented for it
AI_PROVIDER=vertex now says where to configure the model instead of quietly
becoming something else. The Google STT and TTS paths keyed off the same
variable are untouched; neither ever used this SDK.
Verified after deploy: provider litellm, the assistant answers with 8 sources,
/api/models returns 10, and @aws-sdk/s3-request-presigner — which documents.js
needs for presigned MinIO URLs — is still declared and resolvable.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Dv6sqaY6Vq3ChZHMem3cnU
Generated a deck with ds-deepseek-v4-flash and rendered it to look at. The model
produced exactly the six headings it was asked for; the deck came out with eight
slides. The extra ones were pandoc's, not the model's.
Two rules, both found by rendering rather than reading:
- pandoc splits a slide after a table. Anything following one becomes a new
slide with no title — that was the stray "Key differentials to consider:"
slide floating with no heading.
- A table with no blank line before it is not parsed as a table at all. It
renders as literal pipe characters in the preceding paragraph.
And one that was visible on the slide itself: a nested ordered list inside a
bullet ran off the bottom.
None of these are the model failing. A cheap model writes perfectly good slide
markdown — bold, italics, nested lists and a table with a subscript all came
through correctly. It just needs to be told the shape the renderer wants.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Dv6sqaY6Vq3ChZHMem3cnU
pptxgenjs is gone, and with it 269 lines of hand-rolled markdown parsing.
It stretched every image. Reading the slide XML it emitted shows why: it writes
the target box verbatim with <a:stretch/> and a no-op srcRect, so a 200x800
image handed an 11.8x3.9 box came out 1:4 squashed to 3:1. It could not do
better — it never measures an image, and its own getSizeFromImage is commented
out and marked "currently unused", reaching for a package called sizeof that
does not exist. pandoc measures them: a 300x175 source renders at aspect 1.714
and a 160x360 at 0.445, verified by rendering the deck to PDF and looking at it.
Tables, ordered and unordered lists, bold, italic and subscripts all come out
natively, and the fonts, palette and slide layouts come from
assets/learning/slides-reference.pptx. Design now lives in that file: restyling
the decks means editing it in PowerPoint, not editing this route.
Only images the requester owns can reach a deck. pandoc resolves an image link
against the filesystem, so a markdown link naming any local path would read that
file into the presentation. Images are fetched by id through the ownership
check, written into a per-request temporary directory under names we choose, and
every image link that did not resolve is removed rather than passed through. The
directory is removed in a finally block, and the conversion has a 60s timeout so
it cannot hang a request.
pandoc is in the image rather than a sidecar, because an export must not fail
for reasons outside this container. It costs 197MB (307 -> 504).
Removing pptxgenjs also removed image-size, and with it both high-severity
advisories — GHSA-w3rx-r6r6-pgpr and GHSA-5p2g-fcmc-qvqq, ICNS/JXL/HEIF parser
denial of service, ranged <=2.0.2 with no fixed release to upgrade to. npm audit
goes from 2 high and 2 moderate to 2 moderate.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Dv6sqaY6Vq3ChZHMem3cnU
Every generated image in an exported deck was distorted. PowerPoint scales an
image to whatever extent it is given, and the export handed it the content box
verbatim. pptxgenjs has a `sizing: contain` option that looks like it solves
this; reading the emitted slide XML shows it does not — a 200x800 image in an
11.8x3.9 box came out as cx=10789920 cy=3566160 with <a:stretch/>, stretched
from 1:4 to 3:1. It cannot do better: it never measures the image, and its own
getSizeFromImage is commented out and marked "currently unused".
So the export measures the image itself and hands PowerPoint an extent that
already has the right shape, centred in the space available. Verified: a 200x800
image now places 0.97x3.90 and a 4x3 places 5.20x3.90, both matching their
source aspect exactly, neither overflowing. An image that cannot be measured
keeps the old behaviour rather than failing the export.
image-size becomes a real dependency rather than one borrowed transitively, and
an override collapses it to a single copy — pptxgenjs declares it but the string
appears in none of its four shipped bundles, so npm was placing a second copy in
the production image that nothing could load.
Its ICNS, JXL and HEIF parsers have open denial-of-service advisories against
every published version (GHSA-w3rx-r6r6-pgpr, GHSA-5p2g-fcmc-qvqq, both ranged
<=2.0.2), so there is no release to upgrade to. They are disabled instead:
this application measures PNG, JPEG, WebP and GIF and nothing else. An ICNS
buffer is now refused and falls back to the box rather than entering the parser.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Dv6sqaY6Vq3ChZHMem3cnU
Retrying a kept recording transcribed it and put the text on the clipboard,
leaving you to find the right tab and paste. The app already had the answer: the
module was recorded with the audio. Retry now opens that tab and puts the text
in its transcript box.
Two things had to be true first.
The module was not actually being recorded. transcribeAudio never sent one, so
the server stored its default for every upload — all 28 rows in audio_backups
said "recording", and a retry had nowhere to send anything back to. Each
module's call now tags its own upload.
And the names disagreed. The recorders tagged 'encounter', 'soap', 'dictation'
while the recording-started events said 'enc', 'sick', 'dict'. One table now
holds the mapping and resolves the aliases, so the recorder that tags the
upload, the backup row that labels it and the retry that delivers it cannot
drift apart again.
Existing text is appended to, never replaced: a retry usually recovers
something on top of a live transcript, and overwriting would lose the words the
browser did hear. The box only exists once its tab's markup has been fetched, so
delivery polls briefly rather than guessing a delay, and falls back to the
clipboard if the tab never opens. An empty result says so rather than claiming
success. Backup rows now name their source and the button reads "Retry into
SOAP Note" instead of "Retry".
Verified in a browser: enc resolves to encounter, delivery switched tabs and
produced 'existing live transcript\n\nRECOVERED TEXT'.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Dv6sqaY6Vq3ChZHMem3cnU
/metrics answered on every public hostname — app.pedshub.com, peds.danvics.com
and scribe.pedshub.com — with 201 lines naming routes, traffic volumes,
event-loop timings and process internals. No credential in it, but a free map of
the application for anyone probing. Prometheus scrapes the container directly
over the Docker network and never goes through the proxy, so a request carrying
X-Forwarded-For is by definition not Prometheus and now gets a 404.
METRICS_TOKEN allows an explicit override; unset, it can never match.
Verified: 404 on all three hostnames, and up{job="ped-ai"} still 1.
The workspace launcher was blank. [hidden]{display:none!important}, added to
stop "Retry loading settings" showing permanently, outranked
`body.assistant-mode-workspace .assistant-workspace-view { display:block }` —
and that view is the one element toggled by CSS rather than by JavaScript
removing the attribute, so it could never be revealed again. The attribute is
gone from the markup; the class already defaults it to display:none, so it still
starts hidden. Everything else using the attribute is JS-toggled and unaffected.
Verified with a real login: 17 cards at 208x40 in an 860x240 grid.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Dv6sqaY6Vq3ChZHMem3cnU
Removed at the owner's request, entirely rather than switched off: the route's
allow-list, anonymous identity and flag lookup; the client's entry path, the
authFetch exception that let four endpoints out without an account, and the
workspace guard; the CSS, the in-page note, the admin flag and its save/load;
the test file and the assertions elsewhere that pinned it. Both settings rows
are deleted from app_settings.
Two things were checked rather than assumed on the way out. Removing the
anonymous identity collapsed every `if (!req.user.preview)` branch to its
authenticated side, so image tools, audit logging and citation storage now run
unconditionally — which is what they did before preview existed. And the route's
gate went back to a bare router.use(authMiddleware), which on a /api mount gates
every path below it in server.js; it is scoped to /clinical-assistant again, the
guard test catches it either way.
Verified after deploy: signed out, status, examples and chat all refuse with
401; signed in, chat still answers with 8 sources; extensions, encounters,
documents and admin remain shut.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Dv6sqaY6Vq3ChZHMem3cnU
The server side has worked since this morning, but no browser could reach it.
authFetch rejects every /api request that has no account before it is sent,
and the four preview endpoints were not on its short list — so the status call
that decides whether to show the login screen never left the browser, and the
screen was always shown. The list now mirrors the server's own allow-list
exactly: status, examples, chat, chat/stream, and nothing else.
Preview now begins from any path. A visitor landing on the root met the login
wall while /assistant did not, which read as "preview doesn't work"; both now
enter the assistant, and the URL follows.
Reaching for anything that needs an account raises the sign-in screen through
one hook in authFetch rather than a check on every control — but only for
something the visitor did. The page also fetches saved chats and config in the
background on load, and the first version raised the screen for those too,
burying the assistant before a word was typed. The hook is gated on
navigator.userActivation.
The HIPAA notice is hidden on that screen in preview: it is an invitation to
sign in, not the compliance notice a clinician sees on first login.
Verified in a browser: landing on / and on /assistant both show the assistant
with no login wall and no HIPAA text; clicking Workspace raises sign-in.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Dv6sqaY6Vq3ChZHMem3cnU
The card holds more than one Save button and nothing said so. "Save image
settings" is injected directly above "Save model & retrieval settings", with no
indication of where one stops and the other starts, and the page saves nothing
automatically. It now states that up front, and the bottom button says which
settings it applies.
"Retry loading settings" sat beside Save looking like an ordinary control,
because it did: a bare button with the hidden attribute, which the browser's own
[hidden] rule could not hide once .btn-sm set a display. It is now inside an
error message that exists only on failure, says what failed, and says that
nothing typed has been lost.
The status line used to read "Settings ready." forever, which answers a question
nobody asks. It now reports the thing an admin actually wants to know when they
come back: whether the last save went through, and at what time. A toast is gone
in three seconds; this stays on the page.
The signed-out preview moves to Feature Flags, where it belongs. It was a second
checkbox under a row labelled "Sources", followed by two paragraphs, the first
about preview and the second about citations — so neither paragraph clearly
belonged to either checkbox. It is stored as feature.assistant_preview now, with
the old clinical_assistant.preview_enabled still honoured when the new key has
never been written. That also means an ordinary admin can toggle it under
ADMIN_LOCKDOWN: clinical_assistant.* is locked, and putting a day-to-day switch
behind host access was never the intent.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Dv6sqaY6Vq3ChZHMem3cnU
Eleven routers are mounted on '/api' and called router.use(authMiddleware) with
no path. Mounted that way, the gate applies to every /api request that reaches
the router — including routes belonging to routers mounted further down
server.js. extensions.js did it from line 295; the assistant is mounted at 305.
So a signed-out request to /api/clinical-assistant/status was refused ten lines
before the preview middleware could look at it, whatever the admin setting said.
server.js line 250 already warned about this shape.
Each gate now names its own prefix, so a router protects its own routes and
nothing else. Verified afterwards that every namespace which must stay shut
still answers 401 signed out: extensions, encounters, memories, notes, diagrams,
generated images, image jobs, documents, audio backups, ED encounters,
don't-miss, patient education, billing, well visit, admin, transcribe and the
rest. Two of these routers were gating routes nobody realised they were gating.
Second defect in the same path: authMiddleware only ever looks for a token, so
calling it unconditionally after the preview identity had been assigned rejected
exactly the requests preview exists to serve. Only that identity may skip it;
authMiddleware stays strict everywhere else.
Preview now answers with a real cited answer, and stays as narrow as it was
designed to be — four allow-listed paths, no identity, nothing ownable. A test
now walks every /api router and fails on a blanket gate, which is how the last
six were found.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Dv6sqaY6Vq3ChZHMem3cnU
Reproducibility means two things here: the same commit builds the same image,
and the running container can be asked which commit it is.
- Base images are pinned by digest, not by tag. A tag moves; two builds of one
commit could otherwise differ. These are manifest-list digests, so buildx
still picks the right architecture.
- scripts/build-image.sh also writes ped-ai-local:<revision>, an immutable
name a deploy can refer to instead of chasing :latest. Its summary goes to
stderr so stdout stays the Compose invocation.
- Compose takes the image from PED_AI_IMAGE, so a deploy runs a specific
revision-tagged image while a local build still uses the local tag.
- scripts/deploy.sh pins that image in the file Compose interpolates from,
waits for health, then asks /api/build which revision is actually serving
and rolls back to the previous image if it does not match. Healthy is not
the same as running what you asked for. The rollback path was exercised.
- The entrypoint applies migrations before the app starts, so code and schema
arrive together. node-pg-migrate takes an advisory lock; losing it is not an
error, it waits and looks again, so a rolling restart does not fail. A real
migration failure stops the container rather than serving on a schema that
does not match the build. RUN_MIGRATIONS=false opts out.
- The Forgejo workflow builds through that same script, tags by full revision,
and has an opt-in deploy job. It refuses to run if the deploy directory has
uncommitted work rather than resetting over it.
The running image was labelled revision=unknown, and /api/build said "unknown",
because `docker compose up --build` never passes GIT_REVISION. That is exactly
the hole this closes.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Dv6sqaY6Vq3ChZHMem3cnU
Measured in a real browser against the app rather than reasoned about.
The transcript boxes are contenteditable divs, and an editable div zooms on
focus exactly like an <input>. The earlier 16px sweep covered input, textarea
and select, so every workspace tab still zoomed while the calculators did not —
which is exactly what was reported. Every focusable text control in every tab
now measures 16px at phone width; the count of ones below it is zero.
Three ways a recording could end with nothing to show for it:
- Safari supports none of the audio/webm types and throws NotSupportedError
when handed one. Six modules built their own recorder on resume with
"opus, else audio/webm", so resuming threw there and the recording stopped.
There is now one codec chain in the app, and no module constructs a
MediaRecorder of its own.
- audio-recorder-failed is dispatched on document, and the encounter tab
stopped its recording on any of them. The assistant's microphone failing
ended a consultation being recorded in another tab. The recorder now
travels with the event and the listener checks it is its own.
- The server answers {success:true, text:''} for silence, and five modules
assigned that straight into the transcript — emptying the box the browser
had been filling live. It reads as a recording that vanished. Text is now
required before overwriting, and a recording that captured nothing says so
instead of resetting the button over an empty box.
Also: the citation counters were registered on prom-client's default registry
while the app serves its own, so they were never scraped. They read zero at
/metrics now instead of being absent, which is what the Grafana panels need.
And the reference linter passes for the first time, so scripts/e2e.sh gets past
its preflight: KaTeX is vendored (it was referenced by the assistant's LaTeX
rendering but never shipped — three 404s a page load and no math), and the
JavaScript left behind by the removed image picker, saved-chats toggle, image
gallery and visual-output panel is gone.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Dv6sqaY6Vq3ChZHMem3cnU
Citation quality
- A citation naming a source that never came back is never rendered as a
link, so it appears as plain text and nobody learns it happened. It is now
measured on the server, where the answer and the sources both exist, so it
is seen whether or not a browser rendered it.
- Four Prometheus counters feed a Grafana dashboard (Ped-AI Citation
Quality): answers, citations written, answers affected, and individual
unresolved markers. Only answers with at least one unresolved citation are
stored, with the question and the titles retrieval returned, so an operator
can judge whether retrieval came back thin or the model over-cited. Rows
expire after 30 days: this is a quality signal, not a transcript log.
- Both answer paths are covered. /chat/stream is normal; /chat is the
fallback the client uses when streaming fails, so auditing only the first
would have hidden exactly the answers produced under failure.
- The tracker is resolved on demand and allowed to be absent. Seven test
files load this route with a hand-built list of permitted imports, and
adding a hard dependency would mean editing all seven — and the eighth
written later would break. Observation must never be able to fail an
answer, so a missing module simply means no tracking.
- Metric registration reuses an already-registered counter, because this
module can legitimately load twice in one process.
SSO settings on mobile
- Six rows were laid out inline: flex with a 160px label and an input that
would not shrink, so on a phone the row was wider than the screen with
nothing to scroll and no way to reach the rest. They use .admin-row now,
which already stacks below 640px. Verified at 390px and 360px: nothing
off-screen, no sideways overflow.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Dv6sqaY6Vq3ChZHMem3cnU
With more than one admin, every setting in the panel was editable by all of
them — prompts, model policy, retrieval budgets, SMTP, email templates.
ADMIN_LOCKDOWN separates running the service from changing how it behaves.
It is an environment variable on purpose: a setting could be switched off by
the very admin it restrains, so lifting this needs host access and a restart.
The server is the control. One gate refuses configuration writes rather than
a check in each of the fifteen write routes, because that list grows and a
route added later would quietly miss it. Reads always pass — lockdown hides
nothing. Day-to-day operation stays available: invitations, announcements,
registration, feature flags, and the test endpoints, which persist nothing.
A setting invented later is locked until someone deliberately makes it
editable, rather than defaulting to open.
The panel disables what it cannot save and says why, but that is courtesy;
the refusal is what enforces it.
Two things this taught me, both fixed: my first version painted the panel
from an IIFE, which the module conventions forbid, and fetched the whole
config a second time just to read one flag — breaking the test that pins
admin loaders firing exactly once. The state now rides on the invites
response the panel already requests.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Dv6sqaY6Vq3ChZHMem3cnU
registration_enabled was a single switch: open to anyone, or closed to
everyone. This adds the setting an operator actually wants in between — open
to people you invited.
A code is single-use, expires (7 days by default, 90 maximum), and can be
revoked or deleted. It is stored hashed with only its last four characters
kept, because an invite grants account creation and a database dump should
not hand someone a working one. The code is readable exactly once, in the
response that creates it.
The claim is a single conditional UPDATE carrying every condition, so two
registrations racing the same code cannot both succeed. It happens after the
account exists, so a code is never spent on a failed registration — and if
the race is lost, the just-created account is removed rather than left behind
as a free registration. The rejection never says which of the four reasons
applied; distinguishing them would tell someone probing codes which guesses
were closer.
Codes avoid I, L, O and U so they survive being read aloud or copied off a
screen, and matching ignores case and separators.
The sign-up field appears only when the server says a code is required. The
admin card creates, lists, revokes and deletes, and carries the toggle.
Verified against the live database: create, claim, second claim refused,
unknown code refused, revoking a used code refused, delete. 684 tests pass.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Dv6sqaY6Vq3ChZHMem3cnU
It documented 18 of 67, so anyone setting up a deployment had to find the
rest by reading source — including CORS_ORIGINS, REDIS_URL, NODE_ENV and the
whole Clinical Assistant retrieval group. Each entry now shows the default
the code falls back to, so an unset line reads as "this default" rather than
"broken".
Also documents the OpenBao switch: the entrypoint only fetches secrets when
OPENBAO_ADDR is set, a compose or env_file value always beats the OpenBao
copy, and turning it off means unsetting that one variable and putting the
keys here — with the command to list the key names without printing values.
A test now derives the set of variables from source, so this cannot drift
again.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Dv6sqaY6Vq3ChZHMem3cnU
It was textContent, so formatting appeared literally. It now renders inline
Markdown only — bold, italic, code, links, breaks — because this text is
injected into every page and must not be able to introduce headings, images
or block layout that shifts the app around.
parseInline rather than parse, a tag allow-list, event-handler and style
attributes forbidden, data attributes off, and one innerHTML assignment that
can only be reached through DOMPurify. Missing either library, or any failure
while rendering, falls back to the literal text rather than to unsanitised
markup.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Dv6sqaY6Vq3ChZHMem3cnU
Three S3 configurations had grown separately — S3_* for documents,
GENERATED_IMAGES_S3_* for images, and AUDIO_BACKUPS_S3_* after them — with
different key names and their own client construction. That is why moving
storage meant hunting through several files.
src/utils/objectStorage.js now resolves settings for any purpose: its own
variables first, then the shared S3_* ones, with a per-purpose bucket name
(S3_BUCKET_AUDIO_BACKUPS). One endpoint plus three bucket names is enough
for the whole app, and a purpose that needs its own account still overrides
everything. Audio backups and documents use it; generated images keeps its
own tested storage module, whose variable names the resolver already
understands.
Nothing existing has to change: S3_ACCESS_KEY_ID, S3_SECRET_ACCESS_KEY and
the AWS_* fallbacks still resolve, and path-style addressing keeps each
purpose's previous default — off for documents, so a Backblaze endpoint
behaves as before, on where a custom endpoint implies MinIO. A _FILE
credential now always beats an inline one, so a mounted secret cannot be
shadowed by an inherited environment variable.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Dv6sqaY6Vq3ChZHMem3cnU
Switching to the Assistant set window.location, which reloads the document
and silently ended any running recording. The assistant is a tab in the same
page and activateTab already rewrites the URL to /assistant, so the switch
now happens in place; the reload stays as a fallback. Verified in Chromium:
recorder still running, no reload, URL /assistant, assistant visible.
Signing out mid-recording used to end it with nothing kept. It now says so
first — "the audio will be saved for 24 hours so you can transcribe it
later" — and stores the audio either way, tagged with the module that
produced it ('encounter', 'soap', 'dictation'), which is what makes it
findable in Settings afterwards. Sign-out completes whether or not the save
worked, and the rescue never rejects, because the caller is on its way out.
Verified: the warning appears and the upload carries module=encounter.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Dv6sqaY6Vq3ChZHMem3cnU
Recording
- A screen wake lock is held for as long as a recording runs. Browsers drop
the lock whenever the page is hidden, so it is taken again on return —
without that, one glance away ended it for the session. The lock is
reference counted (two recorders cannot release each other's), never
requested while hidden (the request would just be rejected), and a denial
or an unsupported browser leaves the recording running.
- Signing out releases it and stops the recording; nothing is sent, because
the session that owned the audio is gone.
- start() on an already-running recorder is now a no-op instead of replacing
the MediaRecorder and silently dropping everything captured so far.
- A recording that ends by itself — recorder error, or the microphone taken
by another app, unplugged or revoked — takes the same path as pressing
Stop, so it is transcribed and stored rather than left in a tab that still
says "recording". Moving around the workspace already kept recording.
Retention
- Every recording is kept for 24 hours now, not only the ones whose
transcription failed. /api/transcribe already has the audio, so this costs
no second upload, and a storage failure is logged rather than thrown: it
must never lose the transcription someone is waiting for.
- One store (src/utils/audioBackupStore.js) is shared by /api/transcribe and
/api/audio-backups so the two cannot drift. Payload goes to object storage
when AUDIO_BACKUPS_S3_* is set and to the encrypted Postgres column
otherwise; metadata always stays in Postgres, so listing, ownership and
expiry behave the same either way. Object keys are scoped by owner, and
the expiry sweep deletes the object with the row.
Verified against the live database: round trip byte-identical, another user
reads null, 950 -> 48 bytes compressed, expired rows take their objects.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Dv6sqaY6Vq3ChZHMem3cnU
- adminMilestones was reached only because adminConfig is mounted on
/api/admin ahead of it and guards the whole path. adminMiddleware checks
req.user.role and nothing else, so it failed closed (403) rather than
open — but on mount order, not intent. It now states the requirement,
with a test covering all four admin routers.
- TODO.md records the whole audit: what was verified working (live
transcription round trip, voice mode wiring), what was fixed, the two
advisories that are unreachable and why, and the CI/CD and Kubernetes
work worth doing before scaling out.
Security review found nothing else exploitable: parameterised SQL
throughout (the one interpolated table name is allowlisted), CORS refuses
to start open in production, JWT_SECRET refuses to start unset in
production, rate limits on /api and each auth route, a real CSP, and no
secrets in the repo.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Dv6sqaY6Vq3ChZHMem3cnU
Export
- Both stores already held the audio (server /audio-backups/:id/audio and
the local IndexedDB record) but nothing exposed it, so a recording could
not be taken out of the app. Each backup row now has a download, naming
the file by its timestamp and using the extension actually recorded
(webm, or m4a on iOS). A local record is only handed over to the account
that owns it.
Robustness
- MediaRecorder had no onerror and nothing watched the audio track, so a
recorder that failed, or a microphone claimed by another app, unplugged,
or revoked, left the tab saying "recording" while capturing nothing.
Both are now reported once, with the chunks captured so far kept, so
stopping still returns the audio up to the failure.
Deliberately not added: a wake lock. Stopping when the screen sleeps or the
session ends is the intended behaviour — recording is meant to be
deliberate, and nothing is left behind.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Dv6sqaY6Vq3ChZHMem3cnU
Security
- nodemailer 9.0.1 -> 9.1.1, clearing four high advisories, two of which
are delivery bugs that matter for an app that sends mail: recipient-domain
validation bypass via RFC 5322 comments, and an IDN/punycode allow-list
bypass, both of which can route mail to an attacker-controlled domain.
Live transcription
- The Settings picker was a hardcoded list of six ids
(local-whisper-*, local-parakeet-v3, gemini-*). None of them resolve on
this gateway, and /api/transcribe prefers the user's choice over the admin
default, so picking one broke every recording with "Invalid model name".
Verified against the live gateway: local-whisper-large-v3-turbo -> 400.
- The picker now lists what /model/info advertises as audio_transcription,
cached for five minutes, with the built-in list kept only as a fallback
and the admin default marked.
- The pipeline itself is healthy: local-kokoro-tts produced 92KB of speech
and mistral-voxtral-mini-transcribe returned the sentence back verbatim.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Dv6sqaY6Vq3ChZHMem3cnU
Feature audit
- Settings claimed you could reference a template by saying "use my normal
physical exam" in dictation. No phrase handling exists anywhere, and the
prompt says the opposite: "Never copy clinical content from a template —
only formatting and structure." So a template can never supply findings.
The text now says what happens, and keeps the true privacy statement that
only template categories go to the AI (Custom is filtered out in
/memories/context by AI_CONTEXT_CATEGORIES).
- Templates themselves are real: CRUD plus /memories/context, injected as
style hints by hpi, soap, sickVisit, wellVisit, edEncounters and
hospitalCourse, behind the `memories` feature flag.
Docs
- docs/CLINICAL_ASSISTANT.md listed six settings and offered `deepl`, which
no longer exists in the code. The table now covers all seventeen keys the
server reads, with their fallbacks, plus how a model reaches a user.
Testing
- Every test file's stdout is now pure TAP, which is the stream node:test
parses results from. Three sources: a leftover debug console.log dumping
600 characters of HTML, page modules logging into a JSDOM without a
virtual console, and the server startup banners. The banners are guarded
by NODE_TEST_CONTEXT, set only inside node:test children, so production
and `node server.js` output is unchanged (verified both ways).
- Three consecutive full-suite runs at 671/671.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Dv6sqaY6Vq3ChZHMem3cnU
iOS zoom (and the menu button scrolling away with it)
- Safari zooms the whole page when a text field under 16px takes focus and
never zooms back; while zoomed, fixed chrome sits outside the visual
viewport, which is why the menu button had to be scrolled to. All 144
visible text controls were 13-15px. They are 16px on phones now.
!important is deliberate: per-component rules like `.demo-field select`
outrank any element selector, and a new component would slip through.
- The composer set 15px AFTER its own mobile 16px rule, so at equal
specificity the base rule won; the size is per breakpoint now.
- Pinch-zoom is untouched: no user-scalable=no. Desktop unchanged (13/15px).
Composer
- "Ready" / "Looking up sources..." / "saved" repeated what the send button
already shows by turning into Stop, and on a phone they wrapped under the
model control. Kept in the DOM as live regions; nothing is drawn.
- The model control shows the word "Model" (76px) instead of ids as long as
"openrouter-gemini-3.1-flash-image-preview", and opens the full list on
click. The select stays as the state holder, so choosing still saves
through the same change event.
Search
- Nothing typed lists the 19 destinations only; sub-items appear once there
is something to match.
- Listing, matching and opening now read one attribute list instead of three
that could drift, and bedside's emergency sections are searchable.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Dv6sqaY6Vq3ChZHMem3cnU
Admin links
- The [data-admin-jump] click handler had been pasted inside
discoverModels(), so it only registered once someone pressed Search in AI
Model Management, and re-registered on every later search. Normally
nothing intercepted the click, so href="#" did what it says: jump to the
top and leave "#" in the URL.
- This is an app, so the pointers are plain text naming the sections rather
than links, and the handler is gone.
Test flake
- admin-clinical-assistant-wiring failed about 1 run in 4 with "Unable to
deserialize cloned data due to invalid or unsupported version": node:test
reads a test file's results back over the child's stdout, and app.js's own
console.log landed inside a serialized frame.
- The page's console is now forwarded to stderr (jsdom 29: forwardTo).
Verified: child stdout clean, 0 failures in 32 stress runs (was 6 in 24),
and three full-suite runs at 668/668.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Dv6sqaY6Vq3ChZHMem3cnU
Image models
- The Clinical Assistant "Image models" list waited on an old
#assistant-image-model dropdown that no longer exists, so discovery never
reached it and only four hard-coded fallbacks appeared — with no way to
add any of the gateway's 50 image models.
- Image Generation search rows now have + Add / Added. Added models are
saved as clinical_assistant.image_model_roster (validated server-side:
up to 100 ids) and appear in the Clinical Assistant list at once; ticking
one there offers it to users. Anything already allowed or configured
stays listed. Unsaved ticks survive an add.
- The roster notification is guarded, so it can never fail the settings load.
Phone top bar
- The page is drawn under the status bar (viewport-fit=cover) and its
theme colour was the removed header's blue, so on an iPhone content showed
scrolling at the top of the screen. The row is now a real fixed element
that extends behind the status bar (env(safe-area-inset-top), 0 in a
normal tab), the menu button, sources pill and drawers clear it, and the
theme colour is white.
Verified in Chromium: + Add -> saved roster -> listed unticked; tick kept;
remove works. Top bar is the only thing in the top 48px on all 22 pages;
phone menu positions unchanged.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01W4rptBNvn6RYieQw54GXNS
.assistant-sources was a grid with an auto column, which grows to its
widest unbreakable content — so a long file name such as
"2021_Fleisher_&_Ludwig's_Textbook_..." widened every card past the screen.
The column is now minmax(0,1fr), cards shrink (min-width:0) and names break
onto new lines (overflow-wrap:anywhere, with word-break for iOS < 15.4).
On phones the list is one vertical scroll: no sideways pan and no excerpt
scroll box inside it.
Measured on iPhone 14 / SE, Android 360 and Pixel: before, cards reached
636px on every phone (list overflowed by 236-288px); after, the widest card
ends 12px inside the screen and nothing scrolls sideways.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01W4rptBNvn6RYieQw54GXNS
The list is simply there now. Its collapse preference is removed too, so a
list someone had collapsed cannot come back hidden with no control to open
it. The header's CSS (desktop, phone and preview) goes with it.
Sizes, measured in Chromium before and after: switch, New chat and the
name card are unchanged (desktop card 848, phone 790). The list takes the
header's 16px, and the first date heading sits 10px below New chat on both
desktop (167 -> 151) and phone (173 -> 165; phone list top padding 2px ->
10px), matching the menu's 10px rhythm. The rail itself still never scrolls.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01W4rptBNvn6RYieQw54GXNS
- Collapsed menu: its own plain white 52px column against the tiled ground,
no border line, in both views. Flat, it had no edge.
- Collapsed + workspace launcher: assistant.css set the rail track to 0 and
beat the 52px rule, so the collapsed menu (name card included) vanished.
- Search follows the view, not the URL: the workspace launcher keeps
/assistant, so Workspace searched chats.
- Create image, "Base it on": a native <select> list is as wide as its
longest option, so full titles pushed it past the dialog. Labels now keep
the whole words that fit the select's width; the full title is on hover.
Verified in Chromium: account menu items on top in all 8 states; collapsed
strip 52px white with the name card at the same place in all 3 views;
search placeholder and results per mode; dropdown widest label 500px in a
506px box (279 in 280 on a phone).
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01W4rptBNvn6RYieQw54GXNS
- The tiled ground moves from .main-content to .app-body, so the menu sits
on it too and the grid keeps one origin whichever view is showing. A
second tile on .main-content would have started 210px in and seamed.
- Desktop: the workspace sidebar is flat (transparent, no border), the same
as the assistant rail, which drops the white box it briefly had. The phone
sheets stay white because they slide over the page.
- Workspace page cards get the assistant chat panel's tiled-white fill.
Measured in Chromium: name card 848-892 at x=8, brand, switch and 210px
width identical in both views, expanded and collapsed; no page scroll.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01W4rptBNvn6RYieQw54GXNS
Measured in real Chromium (1440x900 and 390x844) rather than inferred.
Desktop
- The assistant rail sat inside the content area's 14px padding, on the
tiled ground. So its background differed from the workspace sidebar's,
its name card sat 6px lower and further in, and the page was 28px taller
than the window. It is now the same flush white sidebar with the same
insets; the chat and sources panels carry the 14px. Measured: name card
848-892 at x=8, brand, switch and 210px width identical in both views,
expanded and collapsed, with no page scroll.
Phone
- The menu button opens the chat history in the assistant (its drawer) and
the workspace list elsewhere. In workspace mode the drawer lists the
workspace links. Choosing a page, switching mode or pressing « closes it.
- The drawer no longer scrolls as a whole: only the chat list does, so the
name card is pinned at the foot, the same place as in the workspace menu
(790-836 at x=12 in both).
- Both menus are one plain shape: same width, text rows, grey current row,
no "Menu" header and one close control.
- The assistant page no longer overflows the phone by 16px.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01W4rptBNvn6RYieQw54GXNS
Phone
- The desktop collapse rules (52px icon rail) were unscoped, and menu-hidden
is remembered across viewports, so after collapsing on a desktop the phone
sheet opened with its links, the Assistant/Workspace switch and the account
name all display:none. Scoped them to min-width:769px.
- Workspace pill: the sheet now closes on tap, and the launcher is no longer
display:none below 640px, so the pill opens the workspace instead of
toggling a view that could never appear.
- The menu button has its own 48px row instead of floating over the page;
layouts subtract --mobile-bar (0px on desktop, so desktop is unchanged).
- Sources open as their own scrollable band under that row, and the toggle
sits in the row, instead of a sheet/FAB covering the composer.
- The sheet uses dvh, so the account card is not pushed below the visible
screen by the browser toolbar.
Admin
- Image Generation card beside TTS and STT: search the gateway's image
models and test one. No Set button: an image model is chosen per workflow.
It searches only on demand, so opening Admin still makes one discovery call.
- Account card sits at the same height in every rail state; the model
availability list points to where new models are added.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01W4rptBNvn6RYieQw54GXNS
.sidebar sets overflow:hidden and the collapsed rail is 52px wide, so the account
menu — laid out inside it — was cut to a sliver. Settings, FAQ and Log out became
unreachable behind the avatar exactly when the rail is narrow, which is the state
the screenshot shows. The collapsed menu is now positioned against the viewport
instead of the rail, so it cannot be clipped by the thing that opens it.
The card also sat lower in the assistant than in the app: the app column pads its
bottom and the assistant rail only padded the top. Both now use the same vertical
padding and the same row height, so the card lands on one baseline in either
view.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_0161xNW1z4vPusrXKGWcQdQu
Three faults, all visible on a phone at once.
Two hamburgers, neither useful. The assistant added a drawer button of its own
AND the app's button was still there, stacked on top of it. Worse, the assistant
hid the app sidebar unconditionally, so the app's button opened something
invisible. A phone now has exactly one menu — the app sidebar — and the assistant
only replaces it with its own rail on desktop.
The collapse control had nothing to collapse on a phone, where the menu is a
sheet rather than a rail. It closes the sheet instead.
Sources were `display:none` below 640px with no alternative, so every SRC chip in
an answer pointed at nothing. They now slide up as a dismissible sheet, opened by
a button that appears only when an answer actually has citations, or by tapping a
citation itself. Tapping away or pressing Escape closes it — a sheet with no way
out is a trap.
The menu itself is sized for a thumb, borrowing the reference layout: a wide
sheet, close and search leading, tappable rows, and padding clear of the home
indicator.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_0161xNW1z4vPusrXKGWcQdQu
The Create image popup showed "Generating image…" forever even after the job
finished. The status poll called fetchAssistantImageJob, which was never
imported, so every tick threw ReferenceError — and the catch treated that like a
transient network failure and rescheduled, permanently. The import is added, a
test now asserts that every api.js function the assistant calls is actually
imported, and the poll distinguishes a programming error (surface it) from a
transient one (retry, but not forever).
Chat titles were hard-cut at 60 characters mid-word, so "Rickets Radiographic
Fea" was all the Create image picker could ever show. The server already allows
160, so titles now keep whole words up to that, and each view decides its own
visible length from the width it actually has rather than inheriting one cut made
at save time. A single very long token still falls back to a hard cut.
Extension cards led with the number at 20px with word-break:break-all, so
"5616/3764/5619" wrapped as "5616/3764/56 19" — unreadable, and unsafe to dial
from. The name leads now, since that is what the eye hunts for in a list of
fifty; the number follows in tabular figures and may only break between groups,
never inside a run of digits. Cards share a minimum height so a grid reads as
rows rather than a ragged mosaic.
The collapsed rail's brand kept its expanded margin-right:auto, which pushed the
stethoscope off the axis the two buttons sat on. Every child of the collapsed
head is now the same centred fixed-size box.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_018e1PLqrKgAM9jQhFKRnbLd
Gallery tiles are 56px but were downloading the full ~280kB original. Previews
are now rendered with sharp and stored beside the originals in the same MinIO
bucket under a thumbs/ prefix, so nothing about credentials, lifecycle or backup
changes. Measured on live assets: 216-294kB originals become 13-19kB at 256px,
about 16x smaller; 640px is about 4x.
Both paths, as asked:
- Rendered when a job completes, so the first viewer never waits for a resize.
A preview failure never unmakes a finished job.
- Rendered on demand for anything that has none — the existing 26 images work
immediately with no backfill required, and the result is stored for next time.
Boundaries that matter more than the speed:
- Only 256 and 640 are honoured. An open width parameter would let a caller
drive arbitrary resizes.
- Permission is checked against the ORIGINAL before a preview is served, so a
preview can never widen who can see an image.
- Previews carry their own SHA-256 and owner headers, because the client
verifies both on every asset; sending the original's checksum would be
rejected as tampering, which is that check working correctly.
- Still private, no-store. The client asset pattern was widened to exactly
?w=256 and ?w=640 and nothing else.
Client-side downscaling stays as the fallback when a preview cannot be produced.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_018e1PLqrKgAM9jQhFKRnbLd
Generated assets are served `private, no-store`, which is right for a clinical
app on a shared workstation — but it meant every gallery render re-downloaded
every image. Measured: 26 images averaging 279kB, so 7.2MB fetched to draw 56px
tiles, on every open of the Create image popup.
Two changes, both reusable anywhere in the app:
- A session cache holding decoded blobs in MEMORY ONLY, so nothing is written to
disk and the no-store posture is unchanged. Entries are keyed by owner ticket
as well as asset, and cleared when the account boundary moves, so one account
can never read another's bytes out of memory. Concurrent tiles asking for the
same asset share one request rather than racing.
- Any img carrying data-image-thumb gets a downscaled copy instead of the
original, so a 56px tile no longer decodes a 300kB image. The gallery asks for
256px and the in-chat preview for 640px; opening the full view still gets the
original. Browsers without OffscreenCanvas, and any decode failure, fall back
to the full image rather than showing nothing.
This does not reduce the first fetch. Serving genuinely smaller bytes needs
server-side resizing, which needs an image library this project does not carry —
worth a deliberate decision rather than adding a native dependency in passing.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GmpYHPSLGmXGZMyLpn2Lbe
Search now reaches sub-navigation, because people look for "bili" rather than
"Calculators". It reads whatever a loaded component exposes — data-calc,
data-subtab, data-section — so it covers every tab with sub-navigation instead of
one hard-coded list, and the relevant components are warmed when the palette
opens. Opening a result survives the component still loading, and reaching one
from the assistant navigates first and opens it after.
Menu head order is now show/hide menu, then the mark, then search. In the
collapsed rail every item is the same centred 52px box, so the icons finally
share one axis — the brand was a flex row with a gap and sat off-centre from the
buttons beneath it.
Settings, FAQ and Admin are no longer listed in the menu: they already have a
place in the account card, and listing them twice only made the tab list longer.
Their sections remain, and activateTab already tolerates a tab whose section
exists without a sidebar button.
The retry loops are named functions rather than IIFEs, which the module
entrypoint rules forbid.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GmpYHPSLGmXGZMyLpn2Lbe
Collapsing hid the sidebar entirely, and the toggle had moved into .menu-head
while the hide rule still targeted direct children of .sidebar-nav — so the head
went, taking the toggle with it, and there was no way to reopen the menu.
Collapsed is now a 52px icon rail rather than a disappearance: the mark, the
toggle, search and the account avatar stay reachable in place, so nothing has to
be floated over the content and there is always a way back. The assistant rail
collapses to the same strip.
Scrolling. The workspace list in the assistant rail had no overflow rule, so its
list of every app tab ran off the bottom and scrolling only appeared once you
left for the app. In the app sidebar the opposite: the whole sidebar scrolled, so
the account card scrolled away with the list and Settings and Log out became
something to hunt for. Both now scroll only the list, with the head and the
account card pinned.
The toggle was also being styled twice — as a head-row icon button and by its own
older rule with a different size and margin. It is now only the behavioural hook
that flips the icon.
Brand is PedAI, which is short enough to sit in the head row without squeezing
the controls; the collapsed rail shows the mark alone.
The reachability test now walks the toggle's real ancestor chain and fails if any
collapse rule hides a branch it sits on, rather than asserting one rule's text —
this class of bug has broken twice by moving the toggle deeper.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GmpYHPSLGmXGZMyLpn2Lbe
The header bar is gone. Settings and Log out moved into the account card, so a
whole band of chrome had nothing left to hold. The brand now heads the menu
alongside the search and collapse controls, on one row rather than three stacked
full-width ones. btn-settings and btn-logout are retained hidden as the canonical
handlers other code already calls, rather than duplicating that logic.
One shell for every view. Each tab was a plain white page while the assistant and
the workspace launcher sat on a tiled card, so opening any menu item changed the
whole face of the app. The tiled ground and the card edge now belong to
.main-content, and only the contents differ. The assistant replaces the shell
card rather than nesting a second one inside it.
Search, in both views, from the same palette: the workspace searches the app
menu, the assistant searches saved chats. Both sources are already in memory, so
this needs no new endpoint. Cmd/Ctrl+K opens it, arrows and Enter work, and the
assistant publishes its chats rather than the palette reaching into its
internals.
The collapse icon is now directional angles that rotate when collapsed, so it
always states what the click will do; fa-table-columns said nothing.
Mobile: the sidebar slides in whole, so the in-sidebar collapse control is hidden
there, the account card sticks to the foot of a scrolling menu, and the palette
is full-bleed instead of a centred dialog on a small screen.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GmpYHPSLGmXGZMyLpn2Lbe
The view now decides only what the menu LISTS. Everything structural is shared,
so switching redraws the content and nothing else:
- One width. The app sidebar is 210px; the assistant rail was 260px, so every
switch shifted the content sideways.
- One surface. The tiled ground belongs to the shared panel rather than to one
view's empty state, so the transcript and the workspace launcher sit on the
same background.
- One menu ending. An account card closes both menus with Settings, FAQ, Log out
and — for admins only — Admin. These were unlabelled icon buttons in the old
header; log out reuses the existing flow rather than a second implementation.
The blue gradient header is now a slim white bar. Being the loudest element on
screen, it made any view without it look like a different product. Everything
measured against it follows from 66px to 52px.
The composer has room around it and centres on an empty transcript instead of
sitting at the very top, and Create image joins take home, export and attach in
the + menu — it acts on the conversation, so it belongs with them rather than as
its own button on the rail.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GmpYHPSLGmXGZMyLpn2Lbe
The assistant carried its own bar above the layout, on top of the blue app
header that both pages now share. On desktop it held a visually-hidden title, a
mobile-only drawer button and a status pill — so it was mostly dead space, and it
made the assistant taller than the app page it is supposed to match.
The bar is gone. The ready/busy indicator moved into the composer footer, next to
the work it describes, and the drawer button floats on phones rather than
occupying a strip the desktop layout does not have. The rail now sticks to the
top of the layout instead of 52px below a bar that no longer exists.
The menu toggle used fa-table-columns, a split-pane glyph that does not read as
"menu". Both copies now use fa-bars and rotate when collapsed to show which way
they will act.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GmpYHPSLGmXGZMyLpn2Lbe
Sources (correcting what I built earlier)
The previous toggle branched the SYSTEM PROMPT, so the same question could get a
different answer depending on a display setting — the bias this was meant to
avoid. The prompt is now unconditional: buildSystemPrompt takes no display
argument and is byte-identical either way. Hiding sources happens on the way out
— the server omits them and strips the now-orphaned [n] markers from the copy it
sends. The answer is generated, stored and exported with citations intact, so
turning the setting back on restores them without re-asking anything. Renamed to
clinical_assistant.show_sources; the old key is still honoured.
Signed-out preview (admin opt-in, default off)
A visitor may try the assistant; reaching for the workspace asks them to sign in.
Deliberately narrow:
- Reachable paths are an exact allow-list, not a pattern, so a new endpoint is
private unless someone adds it on purpose.
- A preview visitor gets no identity at all (id: null), so nothing can be owned,
saved, billed or addressed to them.
- The image tool is withheld rather than left to fail on a null owner, and no
audit rows are written.
- A caller presenting a token is authenticated normally, so preview can never
downgrade a real session; if the setting cannot be read, authentication is
required.
- Actions needing an account are hidden rather than offered and refused.
Composer
The bar above the transcript is gone. Patient take home, Export PDF, Download
transcript and Attach images moved into a + menu in the composer, and the model
selector moved beside send — shown only when there is more than one model, as
before. Both views now start at the same top edge, so switching modes cannot
nudge the page up or down. On an empty transcript the tiled ground runs behind
and below the composer, which floats on it above centre.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GmpYHPSLGmXGZMyLpn2Lbe
