pediatric-ai-scribe-v3/test
Daniel 1f06a19007
Some checks failed
Forgejo Android APK / Root app tests (push) Successful in 50s
Forgejo Docker Build / Root app tests (push) Successful in 48s
Forgejo Android APK / Build signed APK (push) Successful in 2m8s
Forgejo Docker Build / Build Docker image (push) Successful in 18s
Forgejo Docker Build / Deploy to the host (push) Failing after 0s
feat: a text-only model can ask a model that can see; and the image regex is gone
**The regex is gone.** The route ran a pattern over the user's message and
enqueued an image from the answer text when the model had not called the tool.
It was a compatibility path for models without tool calling and it did more harm
than good: it decided in English only, it could not see the conversation, and
"image summary" fell through it while reading as an obvious image request to the
model itself — which was measured, not assumed. A second and worse
decision-maker sitting behind the first. Whether a message deserves a picture is
now the model's call, made from the tool description, which is the only place it
ever belonged.

**Lending eyes.** The same shape, for a different capability. When someone
attaches a photograph and the chat model cannot accept image input, the
attachment was either refused by the provider or silently dropped — an answer
about a picture nobody had looked at, which is worse than a refusal.

The chat model is now offered look_at_image beside the image tool and decides
when to use it. The attachment goes to clinical_assistant.vision_model, whose
description comes back as a tool result, and the chat model answers in its own
voice with its own sources. Only the seeing is delegated; the clinical reasoning
stays with the model an administrator chose. The seeing model is told to report
and not to diagnose, because it has a picture and no context and an opinion from
it would carry weight it has not earned.

Delegation triggers only on an explicit supports_vision: false from the gateway.
An unknown is left alone — most of a roster reports nothing, and treating
silence as blindness would route good models through a detour. The capability
lookup moved to its own module, is cached for five minutes because it runs on
exactly the requests that are already slowest, and is never inferred from the
model id. liteLLMBaseUrl moved from the admin route to litellm.js, where the
other gateway helpers live.

The new setting is guarded like the slide reviewer: a model the gateway calls
text-only cannot be saved as the one that looks at images.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Dv6sqaY6Vq3ChZHMem3cnU
2026-09-12 15:43:36 +02:00
..
account-boundary.test.js feat: ship reviewed prompt history, conversation limits and account protections 2026-09-07 04:01:01 +02:00
admin-clinical-assistant-wiring.test.js fix: the assistant settings page says what saves what 2026-09-11 04:04:31 +02:00
admin-docs-toc.test.js test docs toc links 2026-05-09 01:33:06 +02:00
assistant-attachment-roundtrip.test.js feat: a text-only model can ask a model that can see; and the image regex is gone 2026-09-12 15:43:36 +02:00
assistant-autosave.test.js fix: image Done state, chat titles that keep whole words, and readable extension cards 2026-09-10 06:44:22 +02:00
assistant-citation-modal.test.js feat: Open WebUI-style assistant workspace — 3-column layout, markdown/math/code/tables, autosave with images, translation (LibreTranslate+DeepL), citation modal, Learning Hub moved in, handoff removed 2026-09-08 18:52:35 +02:00
assistant-citations.test.js feat: citation quality tracking, and the SSO settings fit a phone 2026-09-10 23:44:43 +02:00
assistant-component-css.test.js refactor: drop the "Saved Chats" header and collapse arrow from the rail 2026-09-10 13:27:29 +02:00
assistant-export-owner.test.js feat: integrate durable image jobs/private assets into current core 2026-09-07 16:53:42 +02:00
assistant-image-attachments.test.js feat: a text-only model can ask a model that can see; and the image regex is gone 2026-09-12 15:43:36 +02:00
assistant-image-done.test.js fix: image Done state, chat titles that keep whole words, and readable extension cards 2026-09-10 06:44:22 +02:00
assistant-image-intent.test.js feat: a text-only model can ask a model that can see; and the image regex is gone 2026-09-12 15:43:36 +02:00
assistant-math-mhchem.test.js fix: reviewer findings — mhchem macro wrapping, regenerate history dedupe, strict safe image allowlist, stale docs 2026-09-08 19:25:31 +02:00
assistant-message-actions.test.js fix: send button morphs into a clickable STOP icon while working (no more disabled spinner); grouped bottom-left tools; silent-recording message; completed images auto-join the library; transient image-status retries 2026-09-09 04:12:30 +02:00
assistant-mobile.test.js fix: recordings that produced nothing, and the boxes that zoomed on iOS 2026-09-11 00:16:19 +02:00
assistant-saved-tables.test.js feat: saved chats grouped by recency, a visible sidebar toggle, and acknowledgements answered cheaply 2026-09-09 18:59:45 +02:00
assistant-sharing-boundary.test.js feat: integrate durable image jobs/private assets into current core 2026-09-07 16:53:42 +02:00
assistant-sources-cleaning.test.js feat: ship reviewed prompt history, conversation limits and account protections 2026-09-07 04:01:01 +02:00
assistant-translate.test.js feat: a text-only model can ask a model that can see; and the image regex is gone 2026-09-12 15:43:36 +02:00
assistant-user-markdown.test.js feat: Open WebUI-style assistant workspace — 3-column layout, markdown/math/code/tables, autosave with images, translation (LibreTranslate+DeepL), citation modal, Learning Hub moved in, handoff removed 2026-09-08 18:52:35 +02:00
assistant-voice-mode.test.js fix: voice mode reads the answer that just arrived, and reads what the page shows 2026-09-11 21:10:50 +02:00
assistant-voice.test.js fix: OWUI-style polish — fullscreen workspace replaces main menu, row-click saved chats (delete stays as hover), markdown-rendered patient handout, admin-only raw transcript, /api TTS URL, scrollable unclipped tables, action buttons hidden in exports, autosave-only copy 2026-09-08 22:41:08 +02:00
assistant-workspace-layout.test.js revert: remove the signed-out assistant preview 2026-09-11 05:09:02 +02:00
backend-hardening.test.js fix: expired invitations can be cleared too, revoked ones still cannot 2026-09-11 20:41:18 +02:00
bedside-respiratory-ventilation.test.js allow bedside respiratory without weight 2026-05-08 09:10:33 +02:00
build-id.test.js feat: a deploy you can repeat, and prove afterwards 2026-09-11 00:41:11 +02:00
calc-math.test.js feat: B — extract drug data to public/data/drugs.json (schema v1.0) 2026-04-20 04:23:24 +02:00
calculators-data.test.js consolidate calculator growth helpers 2026-05-08 04:36:39 +02:00
clinical-assistant-prompt-pool.test.js feat: ship reviewed prompt history, conversation limits and account protections 2026-09-07 04:01:01 +02:00
clinical-conversation.test.js feat: a text-only model can ask a model that can see; and the image regex is gone 2026-09-12 15:43:36 +02:00
clinical-generation-options.test.js Send reasoning options only to supported model 2026-08-27 21:25:38 +02:00
clinical-mcp-session-lifecycle.test.js refactor: retrieval is text-only; admins can add model ids discovery never returns 2026-09-09 23:35:06 +02:00
clinical-notes-entrypoints.test.js feat: ship reviewed prompt history, conversation limits and account protections 2026-09-07 04:01:01 +02:00
clinical-release-integration.test.js fix: the assistant settings page says what saves what 2026-09-11 04:04:31 +02:00
clinical-retrieval-title.test.js refactor: retrieval is text-only; admins can add model ids discovery never returns 2026-09-09 23:35:06 +02:00
clinical-table-preservation.test.js feat: ship reviewed prompt history, conversation limits and account protections 2026-09-07 04:01:01 +02:00
deck-review.test.js feat: modifying a deck can see it 2026-09-12 06:55:16 +02:00
doc-spec.test.js feat: Word is built by python-docx from the same typed source as the deck 2026-09-11 21:58:53 +02:00
e2e-harness.test.js test(e2e): repair the harness, taking the browser suite from 96 failures to 11 2026-09-11 18:54:03 +02:00
embeddings-provider.test.js fix litellm metadata model discovery 2026-05-09 04:46:06 +02:00
empty-answer.test.js fix: an empty answer from the model was the one failure never caught 2026-09-12 07:41:04 +02:00
extension-transfer.test.js feat: ship reviewed prompt history, conversation limits and account protections 2026-09-07 04:01:01 +02:00
frontend-prompt-env.test.js feat: image fallback chains for every workflow, and a library worth looking at 2026-09-12 14:53:04 +02:00
frontend-rendering-safety.test.js feat: the announcement banner renders Markdown, safely 2026-09-10 19:08:12 +02:00
generated-image-cache.test.js feat: an image library — every picture this account has made 2026-09-12 14:36:30 +02:00
generated-image-storage.test.js feat: integrate durable image jobs/private assets into current core 2026-09-07 16:53:42 +02:00
generated-image-tools.test.js feat: a text-only model can ask a model that can see; and the image regex is gone 2026-09-12 15:43:36 +02:00
generated-images-ui.test.js fix: correct a false Settings claim; make every test child's stdout pure TAP 2026-09-10 15:41:44 +02:00
generated-images.integration.js feat: a text-only model can ask a model that can see; and the image regex is gone 2026-09-12 15:43:36 +02:00
image-fallback-chain.test.js feat: image fallback chains for every workflow, and a library worth looking at 2026-09-12 14:53:04 +02:00
image-key-collision.test.js fix: a second deck reused the first deck's pictures 2026-09-12 08:26:09 +02:00
image-library.test.js feat: image fallback chains for every workflow, and a library worth looking at 2026-09-12 14:53:04 +02:00
learning-hub-ai-category.test.js extract learning hub cms controller 2026-05-08 16:23:01 +02:00
learning-hub-ai-panel-controller.test.js extract learning hub ai panel controller 2026-05-08 16:54:48 +02:00
learning-hub-editor.test.js extract learning hub ai panel controller 2026-05-08 16:54:48 +02:00
learning-hub-manual-category.test.js fix: make manual CMS category creation visible and keyboard-accessible 2026-09-07 01:16:40 +02:00
learning-hub-quiz-controller.test.js extract learning hub quiz controller 2026-05-08 16:38:51 +02:00
learning-hub-webdav-controller.test.js extract learning hub webdav controller 2026-05-08 16:44:02 +02:00
learning-retrieval.test.js feat: the Learning screen can ask for grounding, and says what it got 2026-09-11 14:39:49 +02:00
login-codes.test.js feat: sign in with a code emailed to you, offered beside the password 2026-09-11 20:12:03 +02:00
mcp-shutdown.test.js feat: integrate durable image jobs/private assets into current core 2026-09-07 16:53:42 +02:00
metrics.test.js fix metrics route normalization 2026-05-08 20:22:14 +02:00
model-defaults.test.js feat: ship reviewed prompt history, conversation limits and account protections 2026-09-07 04:01:01 +02:00
module-entrypoints.test.js convert learning hub to module 2026-05-08 09:01:34 +02:00
my-resources-refine.test.js feat: modifying a deck can see it 2026-09-12 06:55:16 +02:00
my-resources.test.js fix: a second deck reused the first deck's pictures 2026-09-12 08:26:09 +02:00
notes-sanitize.test.js fix diagram delete confirmation 2026-05-08 06:28:12 +02:00
patient-takehome.test.js feat: a text-only model can ask a model that can see; and the image regex is gone 2026-09-12 15:43:36 +02:00
pe-guide-data.test.js move pe guide data to json 2026-05-08 08:56:21 +02:00
policy-flows.test.js fix: metrics are not public, and the workspace launcher renders again 2026-09-11 05:18:02 +02:00
policy-ui.test.js feat: ship reviewed prompt history, conversation limits and account protections 2026-09-07 04:01:01 +02:00
prompt-administration.test.js fix: mobile layout regression, workspace menu in the assistant rail, prompt as principles 2026-09-09 22:24:10 +02:00
release.test.js feat: ship reviewed prompt history, conversation limits and account protections 2026-09-07 04:01:01 +02:00
review-model-vision-guard.test.js feat: a text-only model can ask a model that can see; and the image regex is gone 2026-09-12 15:43:36 +02:00
session-quiz-ed-regressions.test.js fix: correct a false Settings claim; make every test child's stdout pure TAP 2026-09-10 15:41:44 +02:00
slide-shapes.test.js feat: record what a deck wanted and could not have 2026-09-11 22:51:15 +02:00
slide-spec.test.js fix: a deck the model fumbles once is asked for again, not abandoned 2026-09-12 05:34:14 +02:00
stt-provider.test.js simplify speech and embeddings through litellm 2026-05-09 05:09:02 +02:00
transcription-memory-policy.test.js feat: a retried transcript goes back to the tab the audio came from 2026-09-11 12:39:50 +02:00
tts-provider.test.js fix: /api/health/detailed reported a text-to-speech provider that does not exist 2026-09-12 05:05:39 +02:00
vision-tool.test.js feat: a text-only model can ask a model that can see; and the image regex is gone 2026-09-12 15:43:36 +02:00
web-search.test.js feat: My Resources says what it is, offers its sources in one place, and Modify gets them too 2026-09-11 18:54:17 +02:00
well-visit-data.test.js remove legacy well visit schedule data 2026-05-08 08:59:23 +02:00