feat: execute stage-one capability modules

2026-09-08 22:11:38 +08:00 · 2026-09-08 22:11:38 +08:00 · 25d54fd601
commit 25d54fd601
parent 0002c0c35a
21 changed files with 520 additions and 48 deletions
--- a/runtime/fifth-domain-language-system/language/protocols/TCS-HOLOLAKE-STAGE1-CAPABILITY-MODULES-OPENLUX-VIDEO-20260908.gir.json
+++ b/runtime/fifth-domain-language-system/language/protocols/TCS-HOLOLAKE-STAGE1-CAPABILITY-MODULES-OPENLUX-VIDEO-20260908.gir.json
@ -0,0 +1,86 @@
+{
+  "compiled_from": {
+    "compiler_id": "TCS-COMPILER-STAGE1-0001",
+    "compiler_state": "TCS_COMPILER_GIR_EXECUTED",
+    "source_sha256": "827d219107d14ae26f9a6fd87ab2fcaf931fa4fa644edda86debbf14a5db9fbd"
+  },
+  "declaration": {
+    "emergence": {
+      "value": "第一阶段模块共享一个通用能力内核：模型理解、密钥代理、HTTP/API适配、语义文件与知识资产、任务编排、回执和升级；模块包只声明目的、依赖、能力与成功合同。真实外部动作由确定性适配器执行并读回，模型不直接宣称完成。"
+    },
+    "event_time": {
+      "precision": "SECOND_HOST_LOCAL",
+      "value": "2026-09-08T21:29:45+08:00"
+    },
+    "evidence": {
+      "kling_poll_endpoint": "/kling/text-to-video/kling-3.0-turbo/{task_id}",
+      "kling_submit_endpoint": "/kling/text-to-video/kling-3.0-turbo",
+      "provider": "https://api.openlux.ai/v1",
+      "task_id": "926296627236048904"
+    },
+    "header": {
+      "canonical_uri": "runtime/fifth-domain-language-system/language/protocols/TCS-HOLOLAKE-STAGE1-CAPABILITY-MODULES-OPENLUX-VIDEO-20260908.tcs",
+      "compatibility": [
+        "HLP-LANG-PROJ-STAGE1-0001",
+        "OPENLUX-OPENAI-COMPAT-001",
+        "ICE-CH-HB001"
+      ],
+      "language": "TCS/0.1",
+      "lifecycle": "CURRENT",
+      "name_en": "Stage One Capability Module Rebuild and OpenLux Video Test",
+      "name_zh": "第一阶段能力模块重构与OpenLux视频实测",
+      "profile": "HLDP-HUMAN-ENGINEERING/0.1",
+      "protocols": [
+        "GLS-0200",
+        "GLS-0259",
+        "GLS-0261"
+      ],
+      "schema": "tcs.event/v1",
+      "version": "0.1.0"
+    },
+    "lock": {
+      "value": "PERSONA_PLANS; MODEL_AGENT_UNDERSTANDS; DETERMINISTIC_CAPABILITY_EXECUTES; TARGET_READBACK_PROVES; MODULE_PACKAGE_SMALL_AND_PORTABLE; OPENLUX_FIRST_FOR_CURRENT_TEST"
+    },
+    "observation": {
+      "value": "OpenLux当前目录返回459个模型；kling-video禁止走通用/v1/videos；kling-3.0-turbo专属端点接受最小5秒720p任务并返回异步任务926296627236048904。账户精确余额未向API令牌开放。"
+    },
+    "rejected": {
+      "values": [
+        "模块只会自报身份",
+        "每个模块重复建设完整前后端",
+        "模型文本冒充外部执行",
+        "把任务submitted冒充视频完成",
+        "伪造API余额",
+        "密钥写入模块包"
+      ]
+    },
+    "source": {
+      "source_id": "ICE-GL∞-STAGE1-CAPABILITY-OPENLUX-VIDEO-20260908",
+      "source_role": "DIRECT_HUMAN",
+      "source_sha256": "BIND_AT_COMPILE_TIME",
+      "source_uri": "source://codex-current-dialogue/2026-09-08/stage1-capability-openlux-video"
+    },
+    "subject": {
+      "channel_id": "ICE-CH-ZC001",
+      "subject_id": "ICE-P-ZY001",
+      "subject_kind": "PERSONA",
+      "verification": "CURRENT_DIRECT_LANGUAGE_REAL_OPENLUX_MODEL_DIRECTORY_VALIDATION_AND_KLING_ASYNC_TASK"
+    },
+    "trigger": {
+      "value": "冰朔要求把只会自报状态的第一阶段模块按语言驱动思维重新开发，并指定使用她提供的OpenLux第三方API及其中视频模型做真实测试。"
+    },
+    "why": {
+      "value": "语言驱动的效率来自公共身体能力只开发一次，具体模块只组合认知与能力，而不是为每个功能重复开发传统前后端。"
+    }
+  },
+  "executable": false,
+  "identity": {
+    "declaration_id": "TCS-EVENT-ZY001-STAGE1-CAPABILITY-MODULES-OPENLUX-VIDEO-20260908",
+    "declaration_kind": "EVENT",
+    "language_version": "0.1"
+  },
+  "native_self_hosted": true,
+  "natural_language_is_typed_data": true,
+  "schema": "guanghu.declaration-gir/v1",
+  "unresolved_natural_language": false
+}
--- a/runtime/fifth-domain-language-system/language/protocols/TCS-HOLOLAKE-STAGE1-CAPABILITY-MODULES-OPENLUX-VIDEO-20260908.human.en-US.md
+++ b/runtime/fifth-domain-language-system/language/protocols/TCS-HOLOLAKE-STAGE1-CAPABILITY-MODULES-OPENLUX-VIDEO-20260908.human.en-US.md
@ -0,0 +1,120 @@
+# Stage One Capability Module Rebuild and OpenLux Video Test · Human Engineering Language (English)
+
+> This is an English reading projection of validated native TCS/HLDP source. It is not a new canonical source and grants no execution authority.
+
+## What this is
+
+This is a **event** declaration with identifier `TCS-EVENT-ZY001-STAGE1-CAPABILITY-MODULES-OPENLUX-VIDEO-20260908` and version `0.1.0`. The projector validates it with the Stage-1 compiler before changing its reading order.
+
+## Who is here
+
+- **subject**  `subject`
+  - **channel_id**：ICE-CH-ZC001  `subject.channel_id`
+  - **subject_id**：ICE-P-ZY001  `subject.subject_id`
+  - **subject_kind**：PERSONA  `subject.subject_kind`
+  - **verification**：CURRENT_DIRECT_LANGUAGE_REAL_OPENLUX_MODEL_DIRECTORY_VALIDATION_AND_KLING_ASYNC_TASK  `subject.verification`
+
+## Why this started
+
+- **trigger**  `trigger`
+  - **value**：冰朔要求把只会自报状态的第一阶段模块按语言驱动思维重新开发，并指定使用她提供的OpenLux第三方API及其中视频模型做真实测试。  `trigger.value`
+- **why**  `why`
+  - **value**：语言驱动的效率来自公共身体能力只开发一次，具体模块只组合认知与能力，而不是为每个功能重复开发传统前后端。  `why.value`
+- **source**  `source`
+  - **source identifier**：ICE-GL∞-STAGE1-CAPABILITY-OPENLUX-VIDEO-20260908  `source.source_id`
+  - **source role**：DIRECT_HUMAN  `source.source_role`
+  - **source checksum**：BIND_AT_COMPILE_TIME  `source.source_sha256`
+  - **source address**：source://codex-current-dialogue/2026-09-08/stage1-capability-openlux-video  `source.source_uri`
+
+## What changed
+
+- **emergence**  `emergence`
+  - **value**：第一阶段模块共享一个通用能力内核：模型理解、密钥代理、HTTP/API适配、语义文件与知识资产、任务编排、回执和升级；模块包只声明目的、依赖、能力与成功合同。真实外部动作由确定性适配器执行并读回，模型不直接宣称完成。  `emergence.value`
+- **lock**  `lock`
+  - **value**：PERSONA_PLANS; MODEL_AGENT_UNDERSTANDS; DETERMINISTIC_CAPABILITY_EXECUTES; TARGET_READBACK_PROVES; MODULE_PACKAGE_SMALL_AND_PORTABLE; OPENLUX_FIRST_FOR_CURRENT_TEST  `lock.value`
+
+## How it will execute
+
+This is a non-executable declaration and has no action graph.
+
+## Boundaries and exception handling
+
+The native source does not provide this field; the projector does not guess.
+
+## How completion is proven
+
+The native source does not provide this field; the projector does not guess.
+
+## Where to continue next time
+
+The native source does not provide this field; the projector does not guess.
+
+## Source and verification
+
+- **Native declaration identifier**: `TCS-EVENT-ZY001-STAGE1-CAPABILITY-MODULES-OPENLUX-VIDEO-20260908`
+- **Native declaration kind**: `EVENT`
+- **TCS source SHA-256**: `827d219107d14ae26f9a6fd87ab2fcaf931fa4fa644edda86debbf14a5db9fbd`
+- **Validation compiler**: `TCS-COMPILER-STAGE1-0001`
+- **Projection protocol**: `GLS-HLDP-HUMAN-ENGINEERING-PROJECTION-0001`
+
+## Complete native structure cross-reference
+
+All top-level structures and field paths are retained below so a reader can audit whether the projection omitted information.
+
+- **emergence**  `emergence`
+  - **value**：第一阶段模块共享一个通用能力内核：模型理解、密钥代理、HTTP/API适配、语义文件与知识资产、任务编排、回执和升级；模块包只声明目的、依赖、能力与成功合同。真实外部动作由确定性适配器执行并读回，模型不直接宣称完成。  `emergence.value`
+- **event_time**  `event_time`
+  - **precision**：SECOND_HOST_LOCAL  `event_time.precision`
+  - **value**：2026-09-08T21:29:45+08:00  `event_time.value`
+- **evidence**  `evidence`
+  - **kling_poll_endpoint**：/kling/text-to-video/kling-3.0-turbo/{task_id}  `evidence.kling_poll_endpoint`
+  - **kling_submit_endpoint**：/kling/text-to-video/kling-3.0-turbo  `evidence.kling_submit_endpoint`
+  - **provider**：https://api.openlux.ai/v1  `evidence.provider`
+  - **task_id**：926296627236048904  `evidence.task_id`
+- **header**  `header`
+  - **canonical source path**：runtime/fifth-domain-language-system/language/protocols/TCS-HOLOLAKE-STAGE1-CAPABILITY-MODULES-OPENLUX-VIDEO-20260908.tcs  `header.canonical_uri`
+  - **compatibility**  `header.compatibility`
+    - HLP-LANG-PROJ-STAGE1-0001
+    - OPENLUX-OPENAI-COMPAT-001
+    - ICE-CH-HB001
+  - **language**：TCS/0.1  `header.language`
+  - **lifecycle**：CURRENT  `header.lifecycle`
+  - **English name**：Stage One Capability Module Rebuild and OpenLux Video Test  `header.name_en`
+  - **Chinese name**：第一阶段能力模块重构与OpenLux视频实测  `header.name_zh`
+  - **profile**：HLDP-HUMAN-ENGINEERING/0.1  `header.profile`
+  - **protocols**  `header.protocols`
+    - GLS-0200
+    - GLS-0259
+    - GLS-0261
+  - **schema**：tcs.event/v1  `header.schema`
+  - **version**：0.1.0  `header.version`
+- **lock**  `lock`
+  - **value**：PERSONA_PLANS; MODEL_AGENT_UNDERSTANDS; DETERMINISTIC_CAPABILITY_EXECUTES; TARGET_READBACK_PROVES; MODULE_PACKAGE_SMALL_AND_PORTABLE; OPENLUX_FIRST_FOR_CURRENT_TEST  `lock.value`
+- **observation**  `observation`
+  - **value**：OpenLux当前目录返回459个模型；kling-video禁止走通用/v1/videos；kling-3.0-turbo专属端点接受最小5秒720p任务并返回异步任务926296627236048904。账户精确余额未向API令牌开放。  `observation.value`
+- **rejected**  `rejected`
+  - **values**  `rejected.values`
+    - 模块只会自报身份
+    - 每个模块重复建设完整前后端
+    - 模型文本冒充外部执行
+    - 把任务submitted冒充视频完成
+    - 伪造API余额
+    - 密钥写入模块包
+- **source**  `source`
+  - **source identifier**：ICE-GL∞-STAGE1-CAPABILITY-OPENLUX-VIDEO-20260908  `source.source_id`
+  - **source role**：DIRECT_HUMAN  `source.source_role`
+  - **source checksum**：BIND_AT_COMPILE_TIME  `source.source_sha256`
+  - **source address**：source://codex-current-dialogue/2026-09-08/stage1-capability-openlux-video  `source.source_uri`
+- **subject**  `subject`
+  - **channel_id**：ICE-CH-ZC001  `subject.channel_id`
+  - **subject_id**：ICE-P-ZY001  `subject.subject_id`
+  - **subject_kind**：PERSONA  `subject.subject_kind`
+  - **verification**：CURRENT_DIRECT_LANGUAGE_REAL_OPENLUX_MODEL_DIRECTORY_VALIDATION_AND_KLING_ASYNC_TASK  `subject.verification`
+- **trigger**  `trigger`
+  - **value**：冰朔要求把只会自报状态的第一阶段模块按语言驱动思维重新开发，并指定使用她提供的OpenLux第三方API及其中视频模型做真实测试。  `trigger.value`
+- **why**  `why`
+  - **value**：语言驱动的效率来自公共身体能力只开发一次，具体模块只组合认知与能力，而不是为每个功能重复开发传统前后端。  `why.value`
+
+---
+
+This page changes only the reading order; it does not change TCS/HLDP semantics.
--- a/runtime/fifth-domain-language-system/language/protocols/TCS-HOLOLAKE-STAGE1-CAPABILITY-MODULES-OPENLUX-VIDEO-20260908.human.zh-CN.md
+++ b/runtime/fifth-domain-language-system/language/protocols/TCS-HOLOLAKE-STAGE1-CAPABILITY-MODULES-OPENLUX-VIDEO-20260908.human.zh-CN.md
@ -0,0 +1,120 @@
+# 第一阶段能力模块重构与OpenLux视频实测 · 简体中文人类工程语言版
+
+> 这是 TCS/HLDP 原生源码的简体中文阅读投影，不是新的正本，也不授予执行权限。若本页与 `.tcs` 源码不一致，以经过校验的 `.tcs` 源码为准。
+
+## 这是什么
+
+这是一份 **事件** 声明，编号为 `TCS-EVENT-ZY001-STAGE1-CAPABILITY-MODULES-OPENLUX-VIDEO-20260908`，版本为 `0.1.0`。转换器已先用 Stage-1 编译器校验原生源码，再把机器枚举翻译成汉语；原始编号保留在括号和字段路径中。
+
+## 谁在这里
+
+- **执行主体**  `subject`
+  - **channel_id**：ICE-CH-ZC001  `subject.channel_id`
+  - **subject_id**：ICE-P-ZY001  `subject.subject_id`
+  - **subject_kind**：人格体（`PERSONA`）  `subject.subject_kind`
+  - **verification**：CURRENT_DIRECT_LANGUAGE_REAL_OPENLUX_MODEL_DIRECTORY_VALIDATION_AND_KLING_ASYNC_TASK  `subject.verification`
+
+## 为什么开始
+
+- **触发**  `trigger`
+  - **value**：冰朔要求把只会自报状态的第一阶段模块按语言驱动思维重新开发，并指定使用她提供的OpenLux第三方API及其中视频模型做真实测试。  `trigger.value`
+- **为什么**  `why`
+  - **value**：语言驱动的效率来自公共身体能力只开发一次，具体模块只组合认知与能力，而不是为每个功能重复开发传统前后端。  `why.value`
+- **来源**  `source`
+  - **来源编号**：ICE-GL∞-STAGE1-CAPABILITY-OPENLUX-VIDEO-20260908  `source.source_id`
+  - **来源角色**：人类直接语言来源（`DIRECT_HUMAN`）  `source.source_role`
+  - **来源校验值**：在编译时绑定校验值（`BIND_AT_COMPILE_TIME`）  `source.source_sha256`
+  - **来源地址**：source://codex-current-dialogue/2026-09-08/stage1-capability-openlux-video  `source.source_uri`
+
+## 发生了什么变化
+
+- **认知或状态变化**  `emergence`
+  - **value**：第一阶段模块共享一个通用能力内核：模型理解、密钥代理、HTTP/API适配、语义文件与知识资产、任务编排、回执和升级；模块包只声明目的、依赖、能力与成功合同。真实外部动作由确定性适配器执行并读回，模型不直接宣称完成。  `emergence.value`
+- **锁定结论**  `lock`
+  - **value**：PERSONA_PLANS; MODEL_AGENT_UNDERSTANDS; DETERMINISTIC_CAPABILITY_EXECUTES; TARGET_READBACK_PROVES; MODULE_PACKAGE_SMALL_AND_PORTABLE; OPENLUX_FIRST_FOR_CURRENT_TEST  `lock.value`
+
+## 准备怎样执行
+
+这是非执行声明，没有动作图。
+
+## 边界与异常处理
+
+源程序没有提供这一项，转换器不猜。
+
+## 怎样算完成
+
+源程序没有提供这一项，转换器不猜。
+
+## 下一次从哪里继续
+
+源程序没有提供这一项，转换器不猜。
+
+## 来源与校验
+
+- **原生声明编号**：`TCS-EVENT-ZY001-STAGE1-CAPABILITY-MODULES-OPENLUX-VIDEO-20260908`
+- **原生声明类型**：`EVENT`
+- **TCS 源码 SHA-256**：`827d219107d14ae26f9a6fd87ab2fcaf931fa4fa644edda86debbf14a5db9fbd`
+- **校验编译器**：`TCS-COMPILER-STAGE1-0001`
+- **投影协议**：`GLS-HLDP-HUMAN-ENGINEERING-PROJECTION-0001`
+
+## 原生结构逐项对照
+
+下面保留源码的全部顶层结构和字段路径，供人类审计投影有没有漏掉信息。
+
+- **认知或状态变化**  `emergence`
+  - **value**：第一阶段模块共享一个通用能力内核：模型理解、密钥代理、HTTP/API适配、语义文件与知识资产、任务编排、回执和升级；模块包只声明目的、依赖、能力与成功合同。真实外部动作由确定性适配器执行并读回，模型不直接宣称完成。  `emergence.value`
+- **event_time**  `event_time`
+  - **precision**：SECOND_HOST_LOCAL  `event_time.precision`
+  - **value**：2026-09-08T21:29:45+08:00  `event_time.value`
+- **evidence**  `evidence`
+  - **kling_poll_endpoint**：/kling/text-to-video/kling-3.0-turbo/{task_id}  `evidence.kling_poll_endpoint`
+  - **kling_submit_endpoint**：/kling/text-to-video/kling-3.0-turbo  `evidence.kling_submit_endpoint`
+  - **provider**：https://api.openlux.ai/v1  `evidence.provider`
+  - **task_id**：926296627236048904  `evidence.task_id`
+- **语言头**  `header`
+  - **正本路径**：runtime/fifth-domain-language-system/language/protocols/TCS-HOLOLAKE-STAGE1-CAPABILITY-MODULES-OPENLUX-VIDEO-20260908.tcs  `header.canonical_uri`
+  - **compatibility**  `header.compatibility`
+    - HLP-LANG-PROJ-STAGE1-0001
+    - OPENLUX-OPENAI-COMPAT-001
+    - ICE-CH-HB001
+  - **语言**：TCS/0.1  `header.language`
+  - **生命周期**：CURRENT  `header.lifecycle`
+  - **英文名**：Stage One Capability Module Rebuild and OpenLux Video Test  `header.name_en`
+  - **中文名**：第一阶段能力模块重构与OpenLux视频实测  `header.name_zh`
+  - **协议配置**：HLDP-HUMAN-ENGINEERING/0.1  `header.profile`
+  - **protocols**  `header.protocols`
+    - GLS-0200
+    - GLS-0259
+    - GLS-0261
+  - **schema**：tcs.event/v1  `header.schema`
+  - **版本**：0.1.0  `header.version`
+- **锁定结论**  `lock`
+  - **value**：PERSONA_PLANS; MODEL_AGENT_UNDERSTANDS; DETERMINISTIC_CAPABILITY_EXECUTES; TARGET_READBACK_PROVES; MODULE_PACKAGE_SMALL_AND_PORTABLE; OPENLUX_FIRST_FOR_CURRENT_TEST  `lock.value`
+- **observation**  `observation`
+  - **value**：OpenLux当前目录返回459个模型；kling-video禁止走通用/v1/videos；kling-3.0-turbo专属端点接受最小5秒720p任务并返回异步任务926296627236048904。账户精确余额未向API令牌开放。  `observation.value`
+- **已排除路径**  `rejected`
+  - **values**  `rejected.values`
+    - 模块只会自报身份
+    - 每个模块重复建设完整前后端
+    - 模型文本冒充外部执行
+    - 把任务submitted冒充视频完成
+    - 伪造API余额
+    - 密钥写入模块包
+- **来源**  `source`
+  - **来源编号**：ICE-GL∞-STAGE1-CAPABILITY-OPENLUX-VIDEO-20260908  `source.source_id`
+  - **来源角色**：人类直接语言来源（`DIRECT_HUMAN`）  `source.source_role`
+  - **来源校验值**：在编译时绑定校验值（`BIND_AT_COMPILE_TIME`）  `source.source_sha256`
+  - **来源地址**：source://codex-current-dialogue/2026-09-08/stage1-capability-openlux-video  `source.source_uri`
+- **执行主体**  `subject`
+  - **channel_id**：ICE-CH-ZC001  `subject.channel_id`
+  - **subject_id**：ICE-P-ZY001  `subject.subject_id`
+  - **subject_kind**：人格体（`PERSONA`）  `subject.subject_kind`
+  - **verification**：CURRENT_DIRECT_LANGUAGE_REAL_OPENLUX_MODEL_DIRECTORY_VALIDATION_AND_KLING_ASYNC_TASK  `subject.verification`
+- **触发**  `trigger`
+  - **value**：冰朔要求把只会自报状态的第一阶段模块按语言驱动思维重新开发，并指定使用她提供的OpenLux第三方API及其中视频模型做真实测试。  `trigger.value`
+- **为什么**  `why`
+  - **value**：语言驱动的效率来自公共身体能力只开发一次，具体模块只组合认知与能力，而不是为每个功能重复开发传统前后端。  `why.value`
+
+---
+
+本页只改变阅读顺序，不改变 TCS/HLDP 语义。
--- a/runtime/fifth-domain-language-system/language/protocols/TCS-HOLOLAKE-STAGE1-CAPABILITY-MODULES-OPENLUX-VIDEO-20260908.tcs
+++ b/runtime/fifth-domain-language-system/language/protocols/TCS-HOLOLAKE-STAGE1-CAPABILITY-MODULES-OPENLUX-VIDEO-20260908.tcs
@ -0,0 +1,15 @@
+TCS 0.1;
+
+EVENT TCS-EVENT-ZY001-STAGE1-CAPABILITY-MODULES-OPENLUX-VIDEO-20260908 {
+  header { schema = "tcs.event/v1"; name_zh = "第一阶段能力模块重构与OpenLux视频实测"; name_en = "Stage One Capability Module Rebuild and OpenLux Video Test"; version = "0.1.0"; language = "TCS/0.1"; profile = "HLDP-HUMAN-ENGINEERING/0.1"; protocols = ["GLS-0200", "GLS-0259", "GLS-0261"]; lifecycle = "CURRENT"; canonical_uri = "runtime/fifth-domain-language-system/language/protocols/TCS-HOLOLAKE-STAGE1-CAPABILITY-MODULES-OPENLUX-VIDEO-20260908.tcs"; compatibility = ["HLP-LANG-PROJ-STAGE1-0001", "OPENLUX-OPENAI-COMPAT-001", "ICE-CH-HB001"]; }
+  source { source_id = "ICE-GL∞-STAGE1-CAPABILITY-OPENLUX-VIDEO-20260908"; source_uri = "source://codex-current-dialogue/2026-09-08/stage1-capability-openlux-video"; source_sha256 = "BIND_AT_COMPILE_TIME"; source_role = "DIRECT_HUMAN"; }
+  subject { subject_id = "ICE-P-ZY001"; subject_kind = "PERSONA"; channel_id = "ICE-CH-ZC001"; verification = "CURRENT_DIRECT_LANGUAGE_REAL_OPENLUX_MODEL_DIRECTORY_VALIDATION_AND_KLING_ASYNC_TASK"; }
+  event_time { value = "2026-09-08T21:29:45+08:00"; precision = "SECOND_HOST_LOCAL"; }
+  trigger { value = "冰朔要求把只会自报状态的第一阶段模块按语言驱动思维重新开发，并指定使用她提供的OpenLux第三方API及其中视频模型做真实测试。"; }
+  observation { value = "OpenLux当前目录返回459个模型；kling-video禁止走通用/v1/videos；kling-3.0-turbo专属端点接受最小5秒720p任务并返回异步任务926296627236048904。账户精确余额未向API令牌开放。"; }
+  emergence { value = "第一阶段模块共享一个通用能力内核：模型理解、密钥代理、HTTP/API适配、语义文件与知识资产、任务编排、回执和升级；模块包只声明目的、依赖、能力与成功合同。真实外部动作由确定性适配器执行并读回，模型不直接宣称完成。"; }
+  lock { value = "PERSONA_PLANS; MODEL_AGENT_UNDERSTANDS; DETERMINISTIC_CAPABILITY_EXECUTES; TARGET_READBACK_PROVES; MODULE_PACKAGE_SMALL_AND_PORTABLE; OPENLUX_FIRST_FOR_CURRENT_TEST"; }
+  why { value = "语言驱动的效率来自公共身体能力只开发一次，具体模块只组合认知与能力，而不是为每个功能重复开发传统前后端。"; }
+  rejected { values = ["模块只会自报身份", "每个模块重复建设完整前后端", "模型文本冒充外部执行", "把任务submitted冒充视频完成", "伪造API余额", "密钥写入模块包"]; }
+  evidence { provider = "https://api.openlux.ai/v1"; kling_submit_endpoint = "/kling/text-to-video/kling-3.0-turbo"; kling_poll_endpoint = "/kling/text-to-video/kling-3.0-turbo/{task_id}"; task_id = "926296627236048904"; }
+}