核心命題 / Core Question
不是只判斷資料是否可用,而是定位本次使用在何種授權、語意、商業與責任邊界內成立。
Rather than only asking if content is usable, the engine locates the exact licensing, semantic, commercial, and liability boundary that makes a use valid.
Governance Demo / 治理演示系統
不是問能不能用,而是問在哪個邊界內成立、何時失效、失效後誰負責。
Not asking only whether content can be used, but under which boundary it is valid, when it fails, and who bears responsibility afterward.
Owner: 許文耀 / 沈耀888π
Wen-Yao Hsu / Yao Shen 888π
Engine Overview / 引擎說明
不是只判斷資料是否可用,而是定位本次使用在何種授權、語意、商業與責任邊界內成立。
Rather than only asking if content is usable, the engine locates the exact licensing, semantic, commercial, and liability boundary that makes a use valid.
一旦用途超出聲明邊界、缺少閉環條件、或責任鏈斷裂,結果會從 ALLOW / WARN 進入 NON-CLOSABLE、VOID 或 OVERRUN。
When a use exceeds its declared boundary, lacks closure conditions, or breaks the responsibility chain, the state escalates from ALLOW / WARN into NON-CLOSABLE, VOID, or OVERRUN.
每次裁決都會輸出 SCBKR 五維拆解、責任錨點、失效條件與可追蹤的 Semantic Responsibility Hash。
Every adjudication emits a five-dimensional SCBKR breakdown, responsibility anchors, failure conditions, and a traceable Semantic Responsibility Hash.
Interactive Demo / 互動示範區
目前案例 / Current Scenario: -
目前用途 / Current Usage: -
Decision State / 判決狀態
選擇案例與用途後執行裁決。 / Select a scenario and usage mode, then run judgment.
Awaiting judgment / 等待裁決
如需算法 API 授權合作,請聯繫 許文耀 / 沈耀888π
For algorithm/API licensing or collaboration, contact Wen-Yao Hsu / Yao Shen 888π
Official Website: https://hijo790401.github.io/shen-yao-portal/
Email: ken0963521@gmail.com
Auto Demo / 自動播放示範區
流程:新聞內容授權案例 → Summary → RAG Assist → Training → Commercial Generation → Hash。
Flow: News Content Licensing → Summary → RAG Assist → Training → Commercial Generation → Hash.
SCBKR Explanation / SCBKR 說明區
定義資料來源是否有授權、引用條件、期限與轉授權限制。
Defines whether the source has licensing, citation requirements, term limits, and sublicense restrictions.
界定系統對輸出結果主張到何種程度,是否宣稱等價替代原內容。
Sets how strongly the system claims its output and whether it substitutes the original content.
檢查用途是否仍留在原授權、原上下文與原責任結構內。
Checks whether the use remains inside the original license, context, and liability structure.
判定是否只是摘要、檢索輔助,或已進入模型訓練與生成再製。
Determines whether the activity is summary, retrieval assistance, or has escalated into training and generative reproduction.
明確列出失效後由資料供應方、模型方、部署方或商業使用方承擔何種責任。
Specifies whether responsibility transfers to the source provider, model operator, deployer, or commercial user after failure.
Boundary Models / 邊界模型區
適用於來源明確、用途受限且責任可追蹤的摘要或檢索輔助場景。
Applies when source attribution is explicit, use is constrained, and responsibility remains traceable for summary or retrieval assistance.
允許有限轉換,但需要保存引用、權利聲明與人工覆核機制。
Allows limited transformation, provided that citation, rights notices, and human review remain in place.
當訓練或大規模再利用無法回收至原授權條件時,責任鏈不再閉合。
When training or large-scale reuse cannot be brought back under the original license terms, the liability chain no longer closes.
在商業生成、替代型輸出或衍生分發場景中,若超出明示授權即進入越界。
Commercial generation, substitutive outputs, or derivative redistribution enter overrun once they exceed explicit authorization.
Result States / 判決狀態區
授權與用途條件完整,責任鏈可追蹤且可閉合。
License and usage conditions remain intact, with a traceable and closable responsibility chain.
可有限成立,但需明確保留來源、限制用途與監控失效條件。
Conditionally valid, but requires explicit source retention, usage limits, and monitoring of failure triggers.
責任鏈無法回到原授權閉環,需停止自動擴張並補足授權。
The responsibility chain cannot return to the original licensing closure; automated expansion should stop until licensing is extended.
授權條件已失效或原主張不成立,使用結果不應被視為合格。
License conditions have failed or the original claim is invalid, so the resulting use should not be treated as qualified.
商業生成或分發越過授權邊界,由越界方承擔主要責任。
Commercial generation or redistribution exceeds the authorized boundary, transferring primary responsibility to the overrunning party.
Licensing Note / 授權合作說明
如需算法 API 授權合作,請聯繫 許文耀 / 沈耀888π。
For algorithm/API licensing or collaboration, contact Wen-Yao Hsu / Yao Shen 888π.
Owner: 許文耀 / 沈耀888π
Wen-Yao Hsu / Yao Shen 888π
https://hijo790401.github.io/shen-yao-portal/
Email: ken0963521@gmail.com