Show the v0.2.0-beta.4 screens in the README

README.ko.md

@@ -36,7 +36,7 @@
 판단을 뒷받침하는 근거와 아직 풀리지 않은 문제를 살펴보고, 저장된 두 결과를 비교하며 다음 행동을 정합니다.
 
 > [!NOTE]
-> **개인·로컬 사용을 위한 실험적 소스 공개본입니다.** 화면은 한국어·영어를 지원합니다.
+> **개인·로컬 사용을 위한 실험적 소스 공개본입니다.** 화면은 한국어·영어를 지원합니다. 일본어·중국어(간체·번체) 화면도 고를 수 있지만 검수하지 않은 기계 번역입니다.
 > GitHub Actions에서 Python 시험, 웹 시험·빌드, 검토된 기준선을 사용하는 Python 타입 검사를 실행합니다.
 > 합성 데모에서 실제 모델 조사를 끝까지 실행한 기록이 있습니다. 답의 품질이나 실제 자료에서의 유용성을 입증한 것은 아닙니다.
 > 영어 질문·답변·근거 원문을 담은 화면이 있습니다. 전반적인 영어 답 품질, 실계정 연결, 실제 PC 재시작 검증은 일부가 남아 있습니다.
@@ -45,23 +45,23 @@
 ## 실제 화면
 
 함께 들어 있는 [합성 레이더 데모](examples/synthetic-radar-demo-v1)에 한국어 합성 완료 실험 자료를
-추가해 2026-10-25 조사한 화면입니다. 자료는 지어낸 기록이며 실제 측정값이나 장비 성능의 증거가 아닙니다.
+추가해 2026-11-12에 v0.2.0-beta.4로 조사한 화면입니다. 자료는 지어낸 기록이며 실제 측정값이나 장비 성능의 증거가 아닙니다.
 
 <table>
   <tr>
     <td width="50%"><img src="docs/screenshots/ko/01-trace-verdict.png" alt="강우 조건 탐지율 0.80이 기준 0.90 이상에 못 미친 것으로 표시된 추적표"></td>
-    <td width="50%"><img src="docs/screenshots/ko/02-question.png" alt="미달 줄에서 한국어 조사 질문을 채우고 답변 언어를 한국어로 선택한 입력창"></td>
+    <td width="50%"><img src="docs/screenshots/ko/02-question.png" alt="미달 줄에서 채운 한국어 조사 질문이 들어 있는 입력창. 그 아래에 답변이 한국어로 고정돼 있다고 표시됩니다"></td>
   </tr>
   <tr>
     <td><b>1 · 결과가 기준에 못 미칩니다.</b> 판정은 규칙으로 계산합니다. 0.80은 0.90 이상이라는 기준에 못 미칩니다.</td>
-    <td><b>2 · 그 줄에서 원인을 묻습니다.</b> “원인 조사”를 누르면 결과와 기준이 담긴 질문이 채워집니다. 보내기를 눌러야 조사가 시작됩니다.</td>
+    <td><b>2 · 그 줄에서 원인을 묻습니다.</b> “원인 조사”를 누르면 결과와 기준이 담긴 질문이 채워집니다. 보내야 조사가 시작됩니다.</td>
   </tr>
   <tr>
-    <td><img src="docs/screenshots/ko/03-progress.png" alt="가설 세 개를 원문과 대조하는 진행 화면"></td>
-    <td><img src="docs/screenshots/ko/04-answer.png" alt="원래 기준 미달 판정을 유지하며 원인 후보와 그것을 가를 시험을 제안하는 한국어 부분 답변"></td>
+    <td><img src="docs/screenshots/ko/03-progress.png" alt="지금 원문과 대조하는 원인 후보 세 개, 그 이유, 확인한 과정, 근거 범위, 판단 상태를 보여 주는 진행 화면"></td>
+    <td><img src="docs/screenshots/ko/04-answer.png" alt="'판단 보류'로 표시된 한국어 부분 답변. 원래 기준 미달 판정을 유지하며 원인 후보와 그것을 가를 시험을 제안합니다"></td>
   </tr>
   <tr>
-    <td><b>3 · 원문 검토 과정을 봅니다.</b> 진행 화면에서 자료와 대조 중인 원인 후보 세 개를 확인합니다.</td>
+    <td><b>3 · 원문 검토 과정을 봅니다.</b> 진행 화면은 지금 자료와 대조하는 원인 후보, 그 이유, 확인한 과정과 근거 범위를 보여 줍니다.</td>
     <td><b>4 · 미해결 판단을 드러냅니다.</b> 부분 답변은 원래 판정을 유지하고, 가능한 원인과 제안한 시험을 나눠 보여 줍니다.</td>
   </tr>
   <tr><td colspan="2"><img src="docs/screenshots/ko/05-hypotheses.png" alt="자료에서 추론한 예상과 가설에 따른 예상을 각각 AI 제안으로 표시한 가설 비교"></td></tr>

README.md

@@ -37,7 +37,7 @@ competing explanations and recorded results connected.** Inspect what supports a
 see what remains unresolved, and compare two saved results before deciding what comes next.
 
 > [!NOTE]
-> **Experimental source preview for personal/local use.** The interface supports Korean and English.
+> **Experimental source preview for personal/local use.** The interface supports Korean and English. Japanese and Chinese (Simplified, Traditional) screens can be chosen too; they are machine-translated and unreviewed.
 > GitHub Actions checks the Python suite, Web tests/build and a reviewed Python type-error baseline.
 > Recorded synthetic-demo investigations have completed with a real model; this does not establish research quality or usefulness on real data.
 > English question, answer and source-text captures are available; broader English quality, real-account coverage and physical PC restart coverage remain incomplete.
@@ -47,26 +47,26 @@ see what remains unresolved, and compare two saved results before deciding what
 
 ## See it work
 
-English captures from a 2026-10-25 investigation on the bundled
+English captures from a 2026-11-13 investigation with v0.2.0-beta.4 on the bundled
 [synthetic radar demo](examples/synthetic-radar-demo-v1), with an English synthetic completed-experiment
 source added. The question, answer and cited source text shown here are in English. These are invented
 records, not real measurements or evidence of engineering performance.
 
 <table>
   <tr>
     <td width="50%"><img src="docs/screenshots/en/01-trace-verdict.png" alt="Trace table showing rain detection of 0.80 below the criterion of at least 0.90."></td>
-    <td width="50%"><img src="docs/screenshots/en/02-question.png" alt="An English investigation question filled from the failed rain-detection row, with English selected for the response."></td>
+    <td width="50%"><img src="docs/screenshots/en/02-question.png" alt="The question box holding the English investigation question filled from the failed rain-detection row; under it, the reply is shown as fixed to English."></td>
   </tr>
   <tr>
     <td><b>1 · A result misses its criterion.</b> Rules compute the assessment: 0.80 does not meet a requirement of 0.90 or higher.</td>
     <td><b>2 · Ask why, from that row.</b> “Investigate causes” fills in a question with the result and criterion. Sending it starts the investigation.</td>
   </tr>
   <tr>
-    <td><img src="docs/screenshots/en/03-progress.png" alt="Progress while three proposed hypotheses are compared with the original sources."></td>
-    <td><img src="docs/screenshots/en/04-answer.png" alt="A partial English answer retains the failed criterion and proposes causes and tests that could distinguish them."></td>
+    <td><img src="docs/screenshots/en/03-progress.png" alt="Progress view naming the three candidate explanations being compared with the original sources, with the reason, the checks performed, the evidence scope and the assessment state."></td>
+    <td><img src="docs/screenshots/en/04-answer.png" alt="A partial English answer marked 'Assessment on hold' retains the failed criterion and proposes causes and tests that could distinguish them."></td>
   </tr>
   <tr>
-    <td><b>3 · Watch the source review.</b> The progress view names the three candidate explanations being compared with the source documents.</td>
+    <td><b>3 · Watch the source review.</b> The progress view names the candidate explanations being compared with the source documents, and says why, what was checked and how much evidence is in scope.</td>
     <td><b>4 · Keep the unresolved judgment visible.</b> The partial answer retains the original assessment and separates possible causes from proposed tests.</td>
   </tr>
   <tr><td colspan="2"><img src="docs/screenshots/en/05-hypotheses.png" alt="Hypothesis comparison showing an inference from data and hypothesis-only predictions, each marked as an AI proposal."></td></tr>

SOURCE_MANIFEST.json

@@ -1,6 +1,6 @@
 {
   "schema_version": "thoth-source-preview-manifest-v1",
-  "created_at": "2026-11-12T11:00:04.245678+00:00",
+  "created_at": "2026-11-12T16:09:54.951120+00:00",
   "license": "MIT",
   "git_history_included": false,
   "full_suite_verified": false,
@@ -58,13 +58,13 @@
     },
     {
       "path": "README.ko.md",
-      "sha256": "f096324affdd7936709d38a027f71369679e6e3d78df53a9ec275ffcb58637d5",
-      "bytes": 18519
+      "sha256": "fe710924f12741bf8ff363a826aa6569b6aee17e991d3b3150f5b5f765719ff0",
+      "bytes": 18847
     },
     {
       "path": "README.md",
-      "sha256": "2032c1d90b0456774f61b8ba14a726887b5b566b287460ce4b7118ca1c4a6795",
-      "bytes": 16557
+      "sha256": "6f330b9066917ac15954526558d51ea6e31937ac3ff83775d799c311ea61b686",
+      "bytes": 16917
     },
     {
       "path": "SECURITY.md",
@@ -3213,108 +3213,78 @@
     },
     {
       "path": "docs/releases/v0.2.0-beta.4.md",
-      "sha256": "7950b01eb040aa12ed66535ca802780a253f789065c8a12bf4ac6665664bd714",
-      "bytes": 4520
-    },
-    {
-      "path": "docs/screenshots/01-trace-verdict.png",
-      "sha256": "800800b0e2bcfdd290d2f911618aab8a8e60e8e803d2400df9829cd61a351989",
-      "bytes": 169161
-    },
-    {
-      "path": "docs/screenshots/02-question.png",
-      "sha256": "a41bf8421b96d72c7a97c8a663d248f118dde6bad78aee75945e9e7797d74b22",
-      "bytes": 80357
-    },
-    {
-      "path": "docs/screenshots/03-progress.png",
-      "sha256": "7214715cdcc606c1c0f12368f7c6395c188204314446962be551ec875ed2235f",
-      "bytes": 164664
-    },
-    {
-      "path": "docs/screenshots/04-answer.png",
-      "sha256": "04f3cb6002ccc7ed1b58702cbfa0ebc43081021f421a9c85be6a97b05554d8ad",
-      "bytes": 242801
-    },
-    {
-      "path": "docs/screenshots/05-hypotheses.png",
-      "sha256": "a78447e9e5f8a63c8dd2bfeca76e943588cfb9e4f613d603c22c543bf987a14c",
-      "bytes": 406289
-    },
-    {
-      "path": "docs/screenshots/06-evidence.png",
-      "sha256": "2cff40afec991c105b4813fd4b1bf1d9092f836b8cd55b1239511ef480a488c5",
-      "bytes": 287062
+      "sha256": "2f6b52704e74274b374854633ac9dda22ee0bd440db82ccd98e5f64b6e270298",
+      "bytes": 4604
     },
     {
       "path": "docs/screenshots/en/01-trace-verdict.png",
-      "sha256": "63dec42925d7ea512c4caed1bc4c320caa5965e28b1c4a057342359d003c6003",
-      "bytes": 316500
+      "sha256": "c47ab8053dd899a47e8afb31a760ab2326e331b0bd8ec647e45d9fcf05297119",
+      "bytes": 339768
     },
     {
       "path": "docs/screenshots/en/02-question.png",
-      "sha256": "e411f84d3e5ea61d2594bd489bcc91dec03e09656d7d1805703435f463c0a534",
-      "bytes": 210351
+      "sha256": "5d949922c801b30a0ec04e284fe3a34f24729a7ac3a8cffbca51accb0bf898aa",
+      "bytes": 273795
     },
     {
       "path": "docs/screenshots/en/03-progress.png",
-      "sha256": "b789eae6df3d7e27c16f0281ab577d4090b352cc86bcb2ed1f565335c8f362c9",
-      "bytes": 372412
+      "sha256": "76bd8ab2feb97fdda18e516e7b0b145ccbc3806a26b104c9a0448e4de6c44449",
+      "bytes": 392994
     },
     {
       "path": "docs/screenshots/en/04-answer.png",
-      "sha256": "bcbc8f0f623be67b0d0dc398feb0c9d72bc4de801dd717ac4563b04d0bf44ec4",
-      "bytes": 403713
+      "sha256": "8be65dfd7a845f2e15770e75783058959d8a4445fd3a001daf45f431cc8ea12a",
+      "bytes": 479348
     },
     {
       "path": "docs/screenshots/en/05-hypotheses.png",
-      "sha256": "07a8d2e712cfb172168b02a439855ae012f3b51db7ffe8b1bbe65f4764130e8b",
-      "bytes": 638520
+      "sha256": "89db20e9b88abe07fd0be5099c30e339d8cb15a9548159dc613342d3fe75f6a4",
+      "bytes": 766450
     },
     {
       "path": "docs/screenshots/en/06-evidence.png",
-      "sha256": "2c75a384f76f9ac7decfcd0c6783c93bdb20845ea7cb6aa9326de47bb203da42",
-      "bytes": 587085
+      "sha256": "60c1776585c43f5bc937ca6b7741e4bfa8bc59a3eb0a20871f80d76cff147e0a",
+      "bytes": 696081
     },
     {
       "path": "docs/screenshots/en/07-inferred-evidence.png",
-      "sha256": "e518b65a1e9b912ac4f7fdee821b9ffb8f07202f2eedf2140021f1f4923e1451",
-      "bytes": 327895
+      "sha256": "a65dc2e795d773c71f603a8a35a63242f23dac37e76805aec6fa21f292e75ea5",
+      "bytes": 330593
     },
     {
       "path": "docs/screenshots/ko/01-trace-verdict.png",
-      "sha256": "127b0da5c371e3d2a2fdfb88951fd676b5831d7027effd92624ef81cc196e40f",
-      "bytes": 284179
+      "sha256": "77bbf7f2867d9f6ec76249e7fbe6f31f91d0f22ef366a4a388ece142d9140ba6",
+      "bytes": 301083
     },
     {
       "path": "docs/screenshots/ko/02-question.png",
-      "sha256": "d36f726fa03ccf2d9cb4d345ab0f1d92369f0bff786d2881df7a9dc61d3758d4",
-      "bytes": 181763
+      "sha256": "ccf6cc294035f0878cef949ec9426169df09d26c60f5926faa568e12b33cb683",
+      "bytes": 229547
     },
     {
       "path": "docs/screenshots/ko/03-progress.png",
-      "sha256": "c3e7414c59a3a08a47adad32dd6ff7edd5556b04dc5d94c51d439b972585f8e8",
-      "bytes": 374766
+      "sha256": "dc649d8bb10d7f9b8e3a6b43d9019d70c62dca78d57d34bc3a616bddd7cdce33",
+      "bytes": 393370
     },
     {
       "path": "docs/screenshots/ko/04-answer.png",
-      "sha256": "e46affb3cd5409b146648bdd1ad6aa2ef7af8efed43f2431922440550b11ab27",
-      "bytes": 414714
+      "sha256": "c24378bd7a8341693d47c6789aa6893829e7b971844c4307ef664b28cd7447b4",
+      "bytes": 415154
     },
     {
       "path": "docs/screenshots/ko/05-hypotheses.png",
-      "sha256": "93e2e1a87ec8bb6e25db4edb056afec1d24a3807ea63b15ae18f11667a3f97f1",
-      "bytes": 637091
+      "sha256": "6d4f790596cee85925dff582758aff92106977c4ae9bbf9d78981e79b3b20950",
+      "bytes": 684763
     },
     {
       "path": "docs/screenshots/ko/06-evidence.png",
-      "sha256": "c8e1463371de832e8e9806f88ac0d0d7e0ac667a98923025f9b4ccfaba01cb65",
-      "bytes": 558880
+      "sha256": "707ea5c92f00cac3ea9a53b1e94c58348ac1d661ff3fcb9052505208e344ef02",
+      "bytes": 625814
     },
     {
       "path": "docs/screenshots/ko/07-inferred-evidence.png",
-      "sha256": "c9fec121d182637080b5f532fcf190c209d464abd511fe46f1fa1420bfa6bfa9",
-      "bytes": 247568
+      "sha256": "e7acbb6173a77fb79d89fbcf5b0f69dce5accf5b7d52685a4700ae5767ea2d43",
+      "bytes": 256550
     },
     {
       "path": "examples/projectpacks/6g-sandbox-hero/criteria.json",

docs/releases/v0.2.0-beta.4.md

@@ -15,7 +15,7 @@ This release gathers the fixes from a full audit of [v0.2.0-beta.3](https://gith
 
 **Sources and settings**
 
-- The Sources tab lists connected sources first. Choosing a file connects it right away; files that cannot be uploaded are explained before sending (#75).
+- The Sources tab lists connected sources first. The usage scope is already set to "Shared within this project", so choosing files and pressing **Connect** is all it takes; files that cannot be uploaded are explained before sending (#75).
 - Project settings and computer settings are two separate screens. With no model connected, settings show one sentence and a link (#75).
 - The help for source scope now says plainly that there is no screen yet for granting a single task permission: whichever scope you choose, research in the project can read the source (#78).