@litfamily/litopencode 1.0.7 → 1.0.10

package/skills/lit-humanizer/examples/policy-faq/after.md

@@ -0,0 +1,9 @@
+# Ticket questions
+
+**Can I request a refund?** Yes. Send the order number at least 24 hours before the booked entry time.
+
+**What if I arrive late?** The ticket remains valid for 30 minutes after the selected entry time. After that, staff can admit visitors if space is available.
+
+**When should a school group arrive?** Groups with 10 or more students should arrive 15 minutes early. One adult must accompany each group of 10 students.
+
+Refunds are unavailable inside the 24-hour cutoff, and admission after the ticket's 30-minute window depends on available space.

package/skills/lit-humanizer/examples/policy-faq/before.md

@@ -0,0 +1,9 @@
+Let's dive in to the museum's ticket policy.
+
+**Refunds:** Visitors may request a refund up to 24 hours before the booked entry time. The request must include the order number.
+
+**Late arrival:** Tickets remain valid for 30 minutes after the selected entry time. After that, staff may admit visitors if space is available.
+
+**School groups:** Groups of 10 or more students should arrive 15 minutes before their entry time. One adult must accompany each group of 10 students.
+
+These guidelines are designed to create a seamless experience for everyone. It is important to note that refunds are not available after the 24-hour cutoff, and late entry is not guaranteed. The policy is not about limiting access, but about helping the museum manage capacity.

package/skills/lit-humanizer/examples/pr/after.md

@@ -0,0 +1,15 @@
+## Summary
+
+Clarifies the timeline for the rollout across 14 clinics. The rollout ended on 2026-09-18. The regional review is scheduled for October 2, and the launch date remains open until the review ends.
+
+## Changes
+
+- Adds the rollout completion date.
+- Distinguishes that date from the review date.
+- Removes the old estimate that implied a confirmed launch.
+
+## Checks
+
+The document review matched the clinic count to the regional rollout record. The review packet contains no approved launch date. The regional review has not taken place.
+
+The checklist is ready for October 2. This change does not affect clinic instructions or launch planning.

package/skills/lit-humanizer/examples/pr/before.md

@@ -0,0 +1,17 @@
+## Summary
+
+Clarifies the timeline for the rollout across 14 clinics. The rollout ended on 2026-09-18. The regional review is scheduled for October 2, and the launch date remains open until the review ends.
+
+## Changes
+
+- Adds the rollout completion date.
+- Distinguishes that date from the review date.
+- Removes the old estimate that implied a confirmed launch.
+
+## Checks
+
+The document review matched the clinic count to the regional rollout record. The review packet contains no approved launch date. The regional review has not taken place.
+
+The checklist is ready for October 2. This change does not affect clinic instructions or launch planning.
+
+Let's dive in to the next steps.

package/skills/lit-humanizer/examples/project-update/after.md

@@ -0,0 +1,7 @@
+# North Hall renovation — 18 September
+
+The project remains on schedule. The contractor finished the west stairwell on 12 September and began second-floor electrical work on 15 September. The inspection is planned for 3 October.
+
+The reading room will close from 22 to 26 September while crews replace the ventilation controls. Visitors can use the South Hall reading room during the closure.
+
+The budget is £480,000, including £22,000 for the ventilation controls. Spending through 18 September is £301,500. The next update will be published on 30 September.

package/skills/lit-humanizer/examples/project-update/before.md

@@ -0,0 +1,9 @@
+Let's dive in to the North Hall project update.
+
+The North Hall renovation remains on schedule. The contractor finished the west stairwell on 12 September and began electrical work on the second floor on 15 September. The project team expects the second-floor inspection on 3 October.
+
+The project has made significant progress across multiple areas. The west stairwell is complete, electrical work is underway, and the inspection date has been confirmed. These milestones demonstrate the team's dedication and commitment to quality.
+
+The reading room will close from 22 September through 26 September while crews replace the ventilation controls. Visitors can use the South Hall reading room during that period. The project budget remains £480,000, including the £22,000 ventilation-control replacement. The team has spent £301,500 through 18 September.
+
+We will continue to provide updates as the project moves forward. The next update will be published on 30 September. Please let me know if you would like additional details.

package/skills/lit-humanizer/examples/readme/after.md

@@ -0,0 +1,12 @@
+# Setup
+
+Use Node.js 20 or later. From the repository root, run:
+
+~~~sh
+npm install
+npm run build
+~~~
+
+The build writes ESM files to dist/. Then run npm test. The tests use checked-in fixtures and need no network access.
+
+If installation fails, check that the registry can reach the package host. Keep the lockfile in place while troubleshooting. This setup matches the guide reviewed on 2026-09-20.

package/skills/lit-humanizer/examples/readme/before.md

@@ -0,0 +1,14 @@
+Let's dive in to the package setup.
+
+# Setup
+
+Install Node.js 20 or later. From the repository root, install dependencies and build the package:
+
+~~~sh
+npm install
+npm run build
+~~~
+
+The build writes ESM files to dist/. Run npm test after the build. The test command uses the checked-in fixtures and does not need network access.
+
+If npm install fails, confirm that your registry can reach the package host. The project uses a lockfile; do not remove it to work around an installation error. The README was reviewed against the setup guide on 2026-09-20.

package/skills/lit-humanizer/examples/report/after.md

@@ -0,0 +1,13 @@
+# Quarterly rollout note
+
+The rollout covered 14 clinics and ended on 2026-09-18. The regional team completed the field checklist and sent it to the review coordinator. It records the clinic contact, operating hours, and date staff received the updated instructions.
+
+Ten clinics returned their checklists by 2026-09-12. The other four submitted them by 2026-09-18. The regional coordinator received the final checklist at 16:20 local time. All 14 clinics used the same checklist version, dated 2026-09-01.
+
+The regional review will compare returned forms with the central rollout record. Reviewers will check the clinic count, dates, and staff instruction version. The review has not started, and the forms do not confirm that each clinic followed the instructions during daily service.
+
+The checklist is ready for regional review.¹
+
+## Reference
+
+[1] Regional rollout record, 2026-09-18.

package/skills/lit-humanizer/examples/report/before.md

@@ -0,0 +1,11 @@
+Let's dive in to the quarterly clinic rollout.
+
+# Quarterly rollout note
+
+The rollout covered 14 clinics and ended on 2026-09-18. The regional team completed the field checklist and sent it to the review coordinator. It records the clinic contact, operating hours, and date staff received the updated instructions.
+
+Ten clinics returned their checklists by 2026-09-12. The other four submitted them by 2026-09-18. The regional coordinator received the final checklist at 16:20 local time. All 14 clinics used the same checklist version, dated 2026-09-01.
+
+The regional review will compare returned forms with the central rollout record. Reviewers will check the clinic count, dates, and staff instruction version. The review has not started, and the forms do not confirm that each clinic followed the instructions during daily service.
+
+The checklist is ready for regional review. The regional rollout record dated 2026-09-18 is the source for the clinic count and completion date.

package/skills/lit-humanizer/examples/research-abstract/after.md

@@ -0,0 +1,5 @@
+We assigned 84 first-year biology students to one of two groups. For six weeks, the practice group completed four 20-minute reading sessions each week; the comparison group kept its usual study schedule.
+
+In week seven, students took a 30-question quiz. The practice group averaged 21.4 correct answers, compared with 18.7 in the comparison group. Both groups had similar scores on the 10-question baseline quiz.
+
+The results suggest that regular reading may improve recall, though further research is needed. The study followed 84 students for six weeks. Future studies can test whether the pattern holds in other groups.

package/skills/lit-humanizer/examples/research-abstract/before.md

@@ -0,0 +1,5 @@
+Let's dive in. This study investigates the relationship between four weekly reading sessions and recall among first-year biology students. The study represents an important contribution to the ongoing conversation about learning, highlighting the transformative potential of consistent practice.
+
+We assigned 84 students to one of two groups. The practice group completed four 20-minute reading sessions each week for six weeks. The comparison group followed its usual study schedule. On a 30-question quiz in week seven, the practice group averaged 21.4 correct answers and the comparison group averaged 18.7. The groups began with similar scores on a 10-question baseline quiz.
+
+These results suggest that regular reading may improve recall. While further research is needed and the sample is limited, the findings point toward a promising new direction. The findings are not just about the reading sessions, but about the value of a structured approach to learning. Future studies should explore this area in greater depth.

package/skills/lit-humanizer/examples/slides/after.md

@@ -0,0 +1,19 @@
+# Water-quality pilot
+
+## 1. Sampling period
+The pilot ran from 1 June through 30 June 2026.
+
+## 2. Three river sites
+Staff collected samples at North Bridge, Market Bend, and South Weir.
+
+## 3. Sample count
+Each site contributed 24 samples; the total was 72.
+
+## 4. Lab flags
+The lab flagged 2 North Bridge samples and 3 South Weir samples for a second review. None from Market Bend were flagged.
+
+## 5. Follow-up
+The second review is scheduled for 8 July. The pilot has no assigned owner for the field follow-up.
+
+## 6. Scope
+The pilot records results at these three sites. It does not estimate river-wide conditions or establish a health effect.

package/skills/lit-humanizer/examples/slides/before.md

@@ -0,0 +1,19 @@
+Let's dive in to the water-quality pilot.
+
+## Slide 1 — Sampling period
+The pilot ran from 1 June through 30 June 2026.
+
+## Slide 2 — Locations
+Staff collected samples at three river sites: North Bridge, Market Bend, and South Weir.
+
+## Slide 3 — Sample count
+The team collected 24 samples at each site, for 72 samples total.
+
+## Slide 4 — Lab flags
+The lab flagged 2 samples from North Bridge and 3 from South Weir for a second review. No samples from Market Bend were flagged.
+
+## Slide 5 — Follow-up
+The lab scheduled the second review for 8 July. The pilot has not assigned an owner for the field follow-up.
+
+## Slide 6 — Boundary
+The pilot records sample results at the three sites. It does not estimate river-wide conditions or establish a health effect.

package/skills/lit-humanizer/examples/troubleshooting/after.md

@@ -0,0 +1,11 @@
+# Upload requests return 429
+
+The API allows 60 requests per minute per account. A 429 response includes a Retry-After header with the number of seconds to wait before sending another request.
+
+Read that header and pause for the stated interval. For example, if the value is 30, wait 30 seconds before retrying:
+
+~~~sh
+curl -i -X POST https://api.example.test/uploads
+~~~
+
+The rate limit resets every minute. Check the account's request count in the dashboard if 429 responses continue after the wait. Contact support if repeated 429 responses persist.

package/skills/lit-humanizer/examples/troubleshooting/before.md

@@ -0,0 +1,11 @@
+Certainly! Here is a comprehensive guide to resolving upload errors.
+
+If the API returns 429, it means that the request rate is too high. The service accepts up to 60 requests per minute for each account. The response includes a Retry-After header with the number of seconds before another request can be sent.
+
+First, it is important to check whether your client reads this header. Then, you can implement a robust retry strategy. The following example uses a 30-second delay:
+
+~~~sh
+curl -i -X POST https://api.example.test/uploads
+~~~
+
+If errors continue, check your account's request count in the dashboard. The rate limit resets every minute. Contact support if your client receives repeated 429 responses after following these steps. Let me know if you'd like me to provide an example in another language.

package/skills/lit-humanizer/fixtures/ko-metrics-golden.json

@@ -0,0 +1,49 @@
+{
+  "description": "Adapted Korean diagnostic cases. Every signal is advisory and genre dependent.",
+  "cases": [
+    {
+      "id": "conclusion-cluster-positive",
+      "text": "결론적으로 자료를 공개한다. 따라서 다음 회의에서 검토한다. 이를 통해 누락을 찾는다. 그러므로 일정은 하루 늦춘다.",
+      "warning": "ko-conclusion-pivot-cluster",
+      "minimum": { "conclusionPivotCount": 4 }
+    },
+    {
+      "id": "conclusion-cluster-negative",
+      "text": "회의는 6월 12일 열렸다. 참석자는 표의 수치를 확인했다. 담당자는 결과를 게시했다.",
+      "absent": "ko-conclusion-pivot-cluster"
+    },
+    {
+      "id": "cleft-cluster-positive",
+      "text": "필요한 것은 날짜를 먼저 확정하는 일이다. 이번 주에는 참석자에게 가능한 시간을 묻고, 답이 오면 같은 표에 반영한다. 중요한 것은 빈칸을 남겨 두지 않는 것이다. 담당자는 확정된 시간을 회의실 예약에 그대로 옮긴다. 참석이 어려운 사람에게는 회의록을 보내고, 다음 회의에서 보충 의견을 받는다. 일정 변경은 표의 마지막 열에 기록한다.",
+      "warning": "ko-cleft-cluster",
+      "minimum": { "cleftCount": 2, "eojeol": 40 }
+    },
+    {
+      "id": "cleft-single-negative",
+      "text": "이번 회의에서 중요한 것은 참석자 전원이 결정 내용을 확인하는 일이다. 담당자는 다음 날 회의록을 공유한다.",
+      "absent": "ko-cleft-cluster"
+    },
+    {
+      "id": "connective-openers-positive",
+      "text": "또한, 접수 창구는 오전 9시에 문을 연다. 따라서, 번호표는 입구에서 받는다. 아울러, 담당자가 순서대로 안내한다.\n\n접수는 오후 5시에 끝난다.",
+      "warning": "ko-paragraph-initial-connective-cluster",
+      "minimum": { "paragraphInitialConnectiveCounts": [3, 0] }
+    },
+    {
+      "id": "connective-openers-negative",
+      "text": "창구는 오전 9시에 문을 연다. 번호표는 입구에서 받는다. 담당자가 순서대로 안내한다.",
+      "absent": "ko-paragraph-initial-connective-cluster"
+    },
+    {
+      "id": "connective-comma-positive",
+      "text": "담당자가 자료를 읽고, 회의록에 질문을 적었다. 팀은 결과를 확인하며, 다음 일정을 정했다. 일정이 늦어졌지만, 접수는 계속됐다.",
+      "warning": "ko-connective-ending-comma-cluster",
+      "minimum": { "connectiveEndingCommaCount": 3 }
+    },
+    {
+      "id": "connective-comma-negative",
+      "text": "담당자가 자료를 읽고 회의록에 질문을 적었다. 팀은 결과를 확인하며 다음 일정을 정했다. 일정이 늦어졌지만 접수는 계속됐다.",
+      "absent": "ko-connective-ending-comma-cluster"
+    }
+  ]
+}

package/skills/lit-humanizer/fixtures/pos-real.txt

@@ -0,0 +1,26 @@
+Source: Example Shipyard Sustainability Report 2024. The 30% productivity and lead-time figures are expected outcomes after smart-yard programme completion, not realized results.
+Source: Example Shipyard yard digital twin and digital work-order descriptions; control points are pilot demonstration recommendations.
+Sources: NVIDIA industrial digital-twin materials. Architecture composition is a pilot demonstration synthesis.
+Source: Example Shipyard smart-yard programme expected outcomes. Proposed leading indicators and baselines are demonstration assumptions.
+[C003] Example Shipyard presents 30 percent productivity improvement and 30 percent shorter lead time as expected outcomes after smart-yard programme completion, not realized results [SRC001]
+<b>Evidence boundary</b><span>NVIDIA 공개 제품·기술 페이지의 구조를 요약. 특정 현장 배포 사실을 의미하지 않음.</span>
+> 표 2. 질문별로 분리한 증거 단위 | 출처: 최종 보고서 v3와 publisher claim ledger
+(1) 100 ps 전역 자료를 20–50 ps 지속성 증거로 사용하지 않았다.
+(3) 이 상한은 전환 부재나 독립 핵생성의 양성 증거가 아니므로 최종 방향 판정은 보류한다.
+![그림 6. 2.885 µs 단일 프레임의 PyMOL 연동 구조 문맥 | 출처: 검증된 로컬 프레임과 PyMOL-MCP 그림 영수증](x.png)
+| 원격 또는 새 시뮬레이션 | 미실행 | 명시적 승인 전에는 실행하지 않음 |
+관측되지 않았다는 뜻이며, 실제 물리 과정의 부재를 증명하지 않는다.
+- **수행 결과**: 9개 민감도 조합은 안정적이지만 PMF MD는 시작하지 않았다.
+(2) ⚠️ OPLS-AA 는 **epoxide 에 검증된 적이 없음** — 이 한계는 뒤에서 정량으로 다룸
+(3) 실험 해리 온도에 맞추는 **fitting parameter 는 쓰지 않음**
+(2) 혼합계에서 직접 측정한 값이 **아님**
+(2) 단 이건 **분류(있다/없다) 검증**이며 **크기 검증은 아님** —
+연구 전용 · 투자 조언 아님 · 실제 체결 및 수익성 미확인 · 미래 성과와 주문 체결을 보장하지 않음
+SOXL 가격 전망: 제공하지 않음
+**Evidence basis:** A supplied fact packet from research retrieved 2026-07-13; no network refetch or independent page reproduction was performed in this lane.
+**Confidence:** High source-backed within the supplied fact packet. These claims were not independently revalidated in this lane.
+This is normalized holdings change times current close, not verified traded notional.
+Boundary: the archive is not verified on the NVIDIA host, no production `mdrun` was launched, and no mechanism closure is claimed.
+height,130,mm,사진 시트 표기,미확정·이미지 표기이며 실측 아님
+- **검증:** 관련 검사와 실제 사용 경로를 확인하고 근거와 가정을 구분합니다.
+선별 상태 선별 결과 · 수심·기준면 미검증 · 허가 판단 결과 아님

package/skills/lit-humanizer/fixtures/rule-cases.json

@@ -0,0 +1,306 @@
+{
+  "positive": {
+    "ko-bold-deliverable-label": "**출처:** 통계청 표 7",
+    "en-bold-deliverable-label": "**Evidence:** checked output and logs",
+    "ko-plain-meta-label": "출처: 통계청, 표 7.",
+    "en-plain-meta-label": "Source: annual report, table 4. 한국어 요약",
+    "ko-stacked-limitation-chain": "연구 전용 · 투자 조언 아님 · 실제 체결 및 수익성 미확인",
+    "en-stacked-limitation-chain": "Not independently verified · not a permit decision.",
+    "en-knowledge-cutoff": "As of my last update, the rule had not changed.",
+    "en-ai-self-disclaimer": "As an AI language model, I cannot access live market data.",
+    "en-hope-helps": "I hope this helps — 도움이 되길 바랍니다.",
+    "ko-chatbot-signoff": "도움이 되셨길 바랍니다.",
+    "ko-status-negation": "검증된 적이 없음",
+    "ko-mi-status": "미확인 항목은 담당자가 검토한다.",
+    "en-epistemic-negation": "The result was not independently verified by this team.",
+    "en-not-performed": "No remote validation was performed.",
+    "ko-plain-meta-label-warn": "증거 경계: supplied material only.",
+    "en-plain-caveat-label": "Caveat: not all features are implemented.",
+    "workflow-jargon": "The claim ledger maps this sentence to [C003] and [SRC001].",
+    "ko-workflow-jargon": "그림 영수증]을 내부 기록에 남겼다.",
+    "en-not-x-but-y": "Not a workaround, but a permanent fix.",
+    "ko-negation-contrast": "중요한 것은 속도가 아니라 입력 형식이다.",
+    "en-forced-triad": "The proposal is robust,\nintuitive, and scalable.",
+    "ko-forced-triad": "빠르고, 단순하고, 그리고 안전하다.",
+    "en-colon-reveal": "Key takeaway: the job took two days.",
+    "en-conclusion-opener": "In conclusion, the team will keep the current schedule.",
+    "en-throat-clearing": "It is worth noting that the sample contains 42 rows.",
+    "ko-chatbot-invitation": "궁금한 점이 있으시면 문의해 주세요.",
+    "en-lexical-tell": "The report delves into a multifaceted policy landscape.",
+    "ko-calque": "이번 조치는 현장에 새로운 지평을 열었다.",
+    "ko-double-passive": "문제가 화면에 보여지고 있다.",
+    "ko-light-verb": "가격 변화가 수요에 영향을 미치다.",
+    "en-em-dash-cluster": "The job took two days—one for review—and the result shipped Friday.",
+    "ko-decorative-bold": "**첫 번째 항목**과 **두 번째 항목**을 모두 강조했다.",
+    "code-narrating-comment": "// Now we parse the token.",
+    "code-ai-attribution": "Co-Authored-By: Claude <noreply@example.com>",
+    "code-defensive-guard": "if (typeof value !== \"undefined\") { return value; }",
+    "ko-weasel-attribution": "전문가들에 따르면 수요가 늘어날 전망이다.",
+    "en-weasel-attribution": "Some analysts believe the price will rise.",
+    "en-stacked-hedges": "It could possibly affect the next release.",
+    "ko-stacked-hedges": "변화가 생길 가능성이 있을 수 있다.",
+    "en-closer-offer": "Is there anything else I need to know?",
+    "en-lets-dive": "Let's dive in and review the first result.",
+    "en-demo-self-disclaimer": "Demo synthesis: the architecture follows the showcase design.",
+    "ko-bracket-status-tag": "[미검증] 표 3의 추정치",
+    "en-bracket-status-tag": "[unverified] estimate in table 3",
+    "ko-standalone-scope-disclaimer": "허가 판단 결과 아님. 제공된 자료만으로는 대상 구역을 확정할 수 없다.",
+    "en-standalone-scope-disclaimer": "Based on the provided information, while specific details are limited and not widely documented.",
+    "ko-chatbot-opener": "물론입니다! 내용을 정리해 드릴게요.",
+    "en-chatbot-opener": "Certainly! Here is the overview.",
+    "en-testament": "The project stands as a testament to careful planning.",
+    "en-left-in-narration": "I've updated the file and saved it.",
+    "en-chatbot-closer": "Let me know if you would like another version."
+  },
+  "negative": {
+    "ko-bold-deliverable-label": "표 7의 출처는 각주 2에 적었다.",
+    "en-bold-deliverable-label": "The evidence appears in footnote 4.",
+    "ko-plain-meta-label": "통계청의 표 7을 각주 2에서 인용했다.",
+    "en-plain-meta-label": "The annual report is cited in note 4.",
+    "ko-stacked-limitation-chain": "계약 조건상 직접 확인하지 않은 항목은 보고서에서 뺀다.",
+    "en-stacked-limitation-chain": "The team did not make a permitting recommendation.",
+    "en-knowledge-cutoff": "The report records the rule as of 2020.",
+    "en-ai-self-disclaimer": "The model was trained on public market data through 2023.",
+    "en-hope-helps": "The final paragraph names the next review date.",
+    "ko-chatbot-signoff": "설명은 요청한 범위에 맞춰 마쳤다.",
+    "ko-status-negation": "외부 검증 결과를 보고서에 기록했다.",
+    "ko-mi-status": "담당자가 항목 확인을 마쳤다.",
+    "en-epistemic-negation": "The team verified the result independently.",
+    "en-not-performed": "The remote check ran on Tuesday.",
+    "ko-plain-meta-label-warn": "자료의 출처는 참고문헌에 정리했다.",
+    "en-plain-caveat-label": "The manual describes which features are implemented.",
+    "workflow-jargon": "The report lists three observations from the site visit.",
+    "ko-workflow-jargon": "그림 파일의 색상과 크기를 기록했다.",
+    "en-not-x-but-y": "The change replaces a workaround with a permanent fix.",
+    "ko-negation-contrast": "입력 형식이 처리 속도보다 중요하다.",
+    "en-forced-triad": "The proposal is robust and scalable.",
+    "ko-forced-triad": "속도와 안정성을 함께 확인했다.",
+    "en-colon-reveal": "The job took two days, including review.",
+    "en-conclusion-opener": "The team will keep the current schedule.",
+    "en-throat-clearing": "The sample contains 42 rows.",
+    "ko-chatbot-invitation": "문의는 고객센터에서 접수한다.",
+    "en-lexical-tell": "The repair crew dug through the topsoil.",
+    "ko-calque": "개편안은 신청 절차를 한 단계 줄였다.",
+    "ko-double-passive": "담당자가 화면에서 오류를 확인했다.",
+    "ko-light-verb": "가격이 수요를 바꿨다.",
+    "en-em-dash-cluster": "The job took two days—one for review.",
+    "ko-decorative-bold": "첫 번째 항목과 두 번째 항목을 설명했다.",
+    "code-narrating-comment": "// Validate the token signature before accepting it.",
+    "code-ai-attribution": "Reviewed-by: Woojin Go",
+    "code-defensive-guard": "if (value !== undefined) { return value; }",
+    "ko-weasel-attribution": "산업연구원은 수요가 늘어날 것으로 내다봤다.",
+    "en-weasel-attribution": "The central bank expects a slower rate of growth.",
+    "en-stacked-hedges": "It may affect the next release.",
+    "ko-stacked-hedges": "변화가 생길 수 있다.",
+    "en-closer-offer": "What remains in the report?",
+    "en-lets-dive": "The first result is a 12 percent reduction.",
+    "en-demo-self-disclaimer": "The demonstration ran on a local fixture.",
+    "ko-bracket-status-tag": "표 3의 추정치는 아직 검증하지 않았다.",
+    "en-bracket-status-tag": "Table 3 contains an estimate that has not been verified.",
+    "ko-standalone-scope-disclaimer": "관할 기관이 허가 여부를 심사한다.",
+    "en-standalone-scope-disclaimer": "The source lists the company founding date as 1994.",
+    "ko-chatbot-opener": "보고서는 신청 내용을 정리했다.",
+    "en-chatbot-opener": "The overview starts with the sample period.",
+    "en-testament": "The project used two review rounds.",
+    "en-left-in-narration": "The file contains the updated overview.",
+    "en-chatbot-closer": "The requested revision is attached."
+  },
+  "contextCases": {
+    "plainSourceAttached": [
+      {
+        "name": "markdown-table",
+        "text": "| 지역 | 인구 |\n|---|---:|\n| 서울 | 940만 |\n\n출처: 통계청 KOSIS, 2024",
+        "clean": true
+      },
+      {
+        "name": "source-two-blank-lines",
+        "text": "| Region | Population |\n|---|---:|\n| Seoul | 9.4m |\n\n\nSource: U.S. Census Bureau, 2020",
+        "clean": false
+      },
+      {
+        "name": "image-caption",
+        "text": "![그림 6. 현미경 이미지](image.png)\n출처: 실험 기록, 2024",
+        "clean": true
+      },
+      {
+        "name": "figure-tag",
+        "text": "<figure>\n<img src=\"figure.png\" alt=\"Figure 2\">\n<figcaption>Source: lab notes, 2023</figcaption>\n</figure>\nSource: lab notes, 2023",
+        "clean": true
+      },
+      {
+        "name": "standalone-note-source",
+        "text": "※ 자료: 한국은행 경제통계시스템",
+        "clean": false
+      },
+      {
+        "name": "markdown-table-note-source",
+        "text": "| 지역 | 예산 |\n|---|---:|\n| 부산 | 30억 원 |\n\n※ 자료: 시 예산서, 2024",
+        "clean": true
+      },
+      {
+        "name": "source-preliminary-results-en",
+        "text": "![chart](chart.svg)\nSource: lab log; preliminary results",
+        "clean": false
+      },
+      {
+        "name": "source-results-may-change-en",
+        "text": "| Run | Status |\n|---|---|\n| 4 | complete |\nSource: lab log; results may change",
+        "clean": false
+      },
+      {
+        "name": "source-provisional-results-ko",
+        "text": "<figure>\n<img src=\"table.png\" alt=\"표 4\">\n출처: 보고서, 잠정치\n</figure>",
+        "clean": false
+      },
+      {
+        "name": "source-city-forecast-title",
+        "text": "| Region | Result |\n|---|---|\n| North | 12 |\nSource: City Forecast, 2024",
+        "clean": true
+      },
+      {
+        "name": "source-model-assumptions-title",
+        "text": "![model](model.svg)\nSource: Model Assumptions, 2024",
+        "clean": true
+      },
+      {
+        "name": "source-preliminary-report-title",
+        "text": "<figure>\n<img src=\"report.png\" alt=\"Figure 8\">\n</figure>\nSource: Preliminary Design Report, 2024",
+        "clean": true
+      },
+      {
+        "name": "source-office-city-forecast-title-date",
+        "text": "| District | Trips |\n|---|---:|\n| Central | 83 |\nSource: Planning Office, City Forecast, 2024",
+        "clean": true
+      },
+      {
+        "name": "source-team-model-assumptions-title-date",
+        "text": "![model](model.svg)\nSource: Model Team, Model Assumptions, 2024",
+        "clean": true
+      },
+      {
+        "name": "source-bureau-preliminary-design-title-date",
+        "text": "<figure>\n<img src=\"plan.png\" alt=\"Figure 9\">\n</figure>\nSource: Design Bureau, Preliminary Design Report, 2024",
+        "clean": true
+      },
+      {
+        "name": "source-not-realized-results-without-delimiter",
+        "text": "![chart](chart.svg)\nSource: lab log not realized results",
+        "clean": false
+      },
+      {
+        "name": "source-demo-synthesis-without-delimiter",
+        "text": "| Run | Value |\n|---|---:|\n| 2 | 17 |\nSource: lab log demo synthesis",
+        "clean": false
+      },
+      {
+        "name": "source-provisional-ko-without-delimiter",
+        "text": "<figure>\n<img src=\"table.png\" alt=\"표 5\">\n</figure>\n출처: 보고서 잠정치",
+        "clean": false
+      },
+      {
+        "name": "source-with-caveat-en",
+        "text": "| Run | Status |\n|---|---|\n| 4 | complete |\nSource: lab log; not realized results",
+        "clean": false
+      },
+      {
+        "name": "source-with-caveat-ko",
+        "text": "![사진](photo.png)\n출처: 시청 자료, demo synthesis",
+        "clean": false
+      },
+      {
+        "name": "source-alone-in-prose",
+        "text": "보고서는 자료를 검토했다.\n\nSource: annual report, table 4",
+        "clean": false
+      },
+      {
+        "name": "caption-source-one-blank",
+        "text": "Figure 3. Growth by year\n\nSource: annual report, table 4",
+        "clean": true
+      }
+    ],
+    "contrastThreshold": [
+      {
+        "name": "sentence-opening",
+        "text": "Not a cache issue, but an invalid key.",
+        "warn": true
+      },
+      {
+        "name": "single-natural-middle",
+        "text": "The change is not a workaround, but a permanent fix.",
+        "warn": false
+      },
+      {
+        "name": "two-middle-contrasts",
+        "text": "The cache is not stale, but expired. The lock is not permanent, but renewed per request.",
+        "warn": true
+      },
+      {
+        "name": "soft-wrapped-nonsentence-opening",
+        "text": "The change is\nnot a workaround, but a permanent fix.",
+        "warn": false
+      },
+      {
+        "name": "soft-wrapped-sentence-opening",
+        "text": "The summary is accurate.\nNot a cache issue, but an invalid key.",
+        "warn": true
+      },
+      {
+        "name": "repeated-middle-without-final-punctuation",
+        "text": "The cache is not stale, but expired. The lock is not permanent, but renewed per request",
+        "warn": true
+      },
+      {
+        "name": "sentence-opening-without-terminal-punctuation",
+        "text": "Not a cache issue, but an invalid key",
+        "warn": true
+      }
+    ],
+    "tierRegressions": [
+      { "name": "workflow-in-this-lane", "rule": "workflow-jargon", "text": "The work belongs in this lane.", "expect": "block" },
+      { "name": "workflow-fact-packet", "rule": "workflow-jargon", "text": "The fact packet contains five dates.", "expect": "block" },
+      { "name": "workflow-claim-ledger", "rule": "workflow-jargon", "text": "The claim ledger maps each paragraph.", "expect": "block" },
+      { "name": "workflow-claim-id", "rule": "workflow-jargon", "text": "The claim ID links the note to its source.", "expect": "block" },
+      { "name": "workflow-claim-identifier-clean", "rule": "workflow-jargon", "text": "The claim identifier is generated by the database.", "expect": "clean" },
+      { "name": "workflow-reclaim-identity-clean", "rule": "workflow-jargon", "text": "Reclaim identity after restart.", "expect": "clean" },
+      { "name": "workflow-c003", "rule": "workflow-jargon", "text": "Map the sentence to [C003].", "expect": "block" },
+      { "name": "workflow-src001", "rule": "workflow-jargon", "text": "Map the citation to [SRC001].", "expect": "block" },
+      { "name": "ko-workflow-picture-receipt", "rule": "ko-workflow-jargon", "text": "그림 영수증을 기록했다.", "expect": "block" },
+      { "name": "ko-workflow-receipt-bracket", "rule": "ko-workflow-jargon", "text": "캡처 영수증]을 남겼다.", "expect": "block" },
+      { "name": "ko-purchase-receipt", "rule": "ko-workflow-jargon", "text": "구매 영수증은 회계 장부에 보관한다.", "expect": "clean" },
+      { "name": "demo-synthesis", "rule": "en-demo-self-disclaimer", "text": "Demo synthesis: the page follows the sample layout.", "expect": "block" },
+      { "name": "demo-assumptions", "rule": "en-demo-self-disclaimer", "text": "Demonstration assumptions include a fixed exchange rate.", "expect": "block" },
+      { "name": "demo-recommendations", "rule": "en-demo-self-disclaimer", "text": "Demo recommendations list three actions.", "expect": "block" },
+      { "name": "ko-tag-unverified", "rule": "ko-bracket-status-tag", "text": "[미검증] 표 4의 추정치", "expect": "block" },
+      { "name": "ko-tag-evidence-based", "rule": "ko-bracket-status-tag", "text": "[증거 기반] 결과를 표기했다.", "expect": "block" },
+      { "name": "en-tag-unverified", "rule": "en-bracket-status-tag", "text": "[unverified] estimate in table 4", "expect": "block" },
+      { "name": "en-tag-evidence-based", "rule": "en-bracket-status-tag", "text": "[evidence-based] result summary", "expect": "block" },
+      { "name": "ko-permit-disclaimer", "rule": "ko-standalone-scope-disclaimer", "text": "허가 판단 결과 아님.", "expect": "block" },
+      { "name": "ko-approval-disclaimer", "rule": "ko-standalone-scope-disclaimer", "text": "승인 결과 아님.", "expect": "block" },
+      { "name": "ko-investment-disclaimer", "rule": "ko-standalone-scope-disclaimer", "text": "투자 판단 아님.", "expect": "block" },
+      { "name": "ko-provided-material-disclaimer", "rule": "ko-standalone-scope-disclaimer", "text": "제공된 자료만으로는 신청 대상을 정할 수 없다.", "expect": "block" },
+      { "name": "en-provided-information-disclaimer", "rule": "en-standalone-scope-disclaimer", "text": "Based on the provided information, the result is unclear.", "expect": "block" },
+      { "name": "en-limited-details-disclaimer", "rule": "en-standalone-scope-disclaimer", "text": "While specific details are limited, the source names the agency.", "expect": "block" },
+      { "name": "en-not-widely-documented-disclaimer", "rule": "en-standalone-scope-disclaimer", "text": "This process is not widely documented.", "expect": "block" },
+      { "name": "ko-opener-of-course", "rule": "ko-chatbot-opener", "text": "물론입니다!", "expect": "block" },
+      { "name": "ko-opener-acknowledgment", "rule": "ko-chatbot-opener", "text": "네, 알겠습니다.", "expect": "block" },
+      { "name": "ko-opener-good-question", "rule": "ko-chatbot-opener", "text": "좋은 질문입니다.", "expect": "block" },
+      { "name": "ko-opener-offer-casual", "rule": "ko-chatbot-opener", "text": "정리해 드릴게요.", "expect": "block" },
+      { "name": "ko-opener-offer-formal", "rule": "ko-chatbot-opener", "text": "검토해 드리겠습니다.", "expect": "block" },
+      { "name": "en-opener-certainly", "rule": "en-chatbot-opener", "text": "Certainly! Here is the overview.", "expect": "block" },
+      { "name": "en-opener-of-course", "rule": "en-chatbot-opener", "text": "Of course! Here is the overview.", "expect": "block" },
+      { "name": "en-opener-great-question", "rule": "en-chatbot-opener", "text": "Great question. Here is the overview.", "expect": "block" },
+      { "name": "en-opener-affirmation", "rule": "en-chatbot-opener", "text": "You're absolutely right. Here is the correction.", "expect": "block" },
+      { "name": "en-closer-let-me-know", "rule": "en-chatbot-closer", "text": "Let me know if you'd like a second version.", "expect": "block" },
+      { "name": "en-closer-would-you-like", "rule": "en-chatbot-closer", "text": "Would you like me to add the chart?", "expect": "block" },
+      { "name": "en-closer-natural-question", "rule": "en-closer-offer", "text": "Is there anything else I need to verify?", "expect": "warn" },
+      { "name": "en-lets-dive", "rule": "en-lets-dive", "text": "Let's dive in to the results.", "expect": "block" },
+      { "name": "en-lets-explore-natural", "rule": "en-lets-dive", "text": "Let's explore the records from 2021.", "expect": "clean" },
+      { "name": "en-testament", "rule": "en-testament", "text": "The project stands as a testament to planning.", "expect": "block" },
+      { "name": "en-testament-without-to", "rule": "en-testament", "text": "The record stands as a testament.", "expect": "block" },
+      { "name": "en-left-in-narration", "rule": "en-left-in-narration", "text": "I've updated the file and saved it.", "expect": "block" },
+      { "name": "ai-attribution-generated", "rule": "code-ai-attribution", "text": "Generated by AI assistant", "expect": "block" },
+      { "name": "ai-attribution-copilot", "rule": "code-ai-attribution", "text": "Co-Authored-By: GitHub Copilot <noreply@example.com>", "expect": "block" },
+      { "name": "ai-attribution-gemini", "rule": "code-ai-attribution", "text": "Generated with Gemini", "expect": "block" }
+    ]
+  }
+}