@givolt/i3-composer-agent 3.2.1 → 4.0.0

package/skills/i3-enterprise-sop/references/templates/testcase.md

@@ -8,13 +8,13 @@
 默认使用简体中文解释、追问及生成报告；用户明确要求其他语言时可切换。原方法正文、引用、阈值和审批条件完整保留。代码标识、路径、命令、工具参数、JSON 键/枚举与原文引用保持原样。
 <!-- i3:zh-guide:end -->
 
-模板版本：v0.1.7。模板版本与业务文档版本分别维护。
+模板版本：v0.1.8。模板版本与业务文档版本分别维护。
 
 ## 使用说明（不复制为业务正文）
 
 - **输入：**准确选定的 PRD／配套 AD，或外部 TestCase及产品与技术依据；适用测试规范、环境和数据条件。
 - **来源选择：**完整取得所选 PRD、配套 AD及本次必要真实答复；从 AD 采用表或项目已选引用定位规约、契约与目录规范，项目自有资料优先。本次用技术规约／契约条款制定测试预期或质量判据时，直接读取实际采用原件的完整正文，不以 AD 转述代替。测试目录、命令沿用 AD 已落实的映射；直接检查目录规范条款、映射不足或有冲突时，再读实际采用的目录规范，不将其设为所有 TestCase 的统一必读。可读资料未读完先补读，不作为业务 Q 跳过；确实无法取得时说明受限范围，不声称依据齐备。
-- **输出：**对应所选主题／AD的测试设计；默认保存到业务项目的本PBC目录，路径为`spec/{PBC名称}/{PBC名称}-{主题名称}-TestCase_v{业务版本}.md`，实际运行另存TestReport。已有明确采用位置或外部原件保留实际路径，不自动迁移。
+- **输出：**对应所选PBC／AD的测试设计；默认保存到业务项目的本PBC目录，路径为`spec/{PBC名称}/{PBC名称}-TestCase_v{业务版本}.md`，实际运行另存TestReport。已有明确采用位置或外部原件保留实际路径，不自动迁移。
 - **阶段职责：**产品预期直接核 PRD，技术判据核 AD和正式契约，不只由 AD 反推需求。完整交接绑定所选直接上游；关键依据缺失先澄清或标明受限范围，不伪造预期。TestCase不是所有 Plan 的统一前置；已有适用版本应引用，测试策略仍须充分。
 - **写作：**主会话按本 Skill 显式使用原 Spec；来源、假设、真实答复与 Q 分开。新建、外部稿整理与授权修订均保护业务原件、人工编辑和历史，不预填批准或运行结果。
 - **交付：**按下文完整填写适用内容，不以标题或数量代替语义覆盖；新增规范性要求逐条回查来源、权限与状态跨处一致、表格逐行核列义。生成或修订稿保存后检查落盘完整性，分段写入或报错时核对相关内容，保留内容自查；检查完成后按本Skill自动登记本次文档路径，再交人工审阅并停轮。登记不表示批准，只读审查只标注问题，不改稿或登记。
@@ -24,7 +24,7 @@
 
 ---
 
-# {PBC名称} · {需求主题} — 测试用例与验证合同
+# {PBC名称} — 测试用例与验证合同
 
 ## 0. 身份与固定输入
 
@@ -32,7 +32,7 @@
 | --- | --- |
 | 所属 EBC | `{ebcId；ebcName；已确认或候选边界}` |
 | 所属 PBC | `{pbcId；pbcName；已确认或候选边界}` |
-| 产品需求主题 | `{稳定prdId；主题名称；改名不换身份}` |
+| 产品需求依据 | `{本PBC的PRD documentId；实际版本与范围；改名不换身份}` |
 | 文档身份/版本/位置 | `{documentId；类型TestCase；实际业务版本；可回读路径/固定内容及摘要}` |
 | 直接 PRD 依据 | `{选定PRD的documentId、实际路径、文内版本、需求／AC具体条目及固定内容；摘要可选}` |
 | 直接 AD 依据 | `{选定AD的documentId、实际路径、文内版本、设计条款及固定内容；其直接PRD身份／路径／版本与本PRD一致；摘要可选}` |
@@ -44,15 +44,15 @@
 
 | 引用 ID／类型 | 来源身份／实际路径或URL／文内版本或固定提交／具体条目 | 固定内容（摘要可选）与已读范围 | 本次用途／采用范围／判据落实位置与权威定义 | 认可/真实可用性/缺口 |
 | --- | --- | --- | --- | --- |
-| `{IN别名；直接PRD／直接AD／原型／契约／外部规范／真实答复}` | `{documentId或资料标识及适用ebcId／pbcId／prdId；准确位置、版本和条款}` | `{可回读的固定内容及实际已读部分}` | `{产品预期／技术约束／交互等及用例／章节位置}` | `{实际采用、仅参考、未取得分别记录，不由文件存在推定}` |
+| `{IN别名；直接PRD／直接AD／原型／契约／外部规范／真实答复}` | `{documentId或资料标识及适用ebcId／pbcId；准确位置、版本和条款}` | `{可回读的固定内容及实际已读部分}` | `{产品预期／技术约束／交互等及用例／章节位置}` | `{实际采用、仅参考、未取得分别记录，不由文件存在推定}` |
 
-本稿与直接PRD／AD使用一致的EBC／PBC编号名称和prdId，各文档独立版本，名称变化不换编号。跨PBC契约、关联主题与外部资料明确来源身份、准确版本及采用范围，不能替代本主题产品或设计依据。沿PRD／AD的固定引用读取必要原型、契约和适用规范，将判据落实到用例；这些资料不再依靠handoff.json扩展字段传递，也不另建平行清单。必要引用失效或版本冲突时列明缺项和影响，不从目录猜最新。
+本稿与直接PRD／AD使用一致的EBC／PBC编号名称和PRD documentId，各文档独立版本，名称变化不换编号。跨PBC契约、关联PBC与外部资料明确来源身份、准确版本及采用范围，不能替代本PBC产品或设计依据。沿PRD／AD的固定引用读取必要 i3-ui-demo 实例的准确版本、HANDOFF、设计与QA资料，以及契约和适用规范，将判据落实到用例；这些资料不再依靠handoff.json扩展字段传递，也不另建平行清单。必要引用失效或版本冲突时列明缺项和影响，不从目录猜最新。
 
-本 TestCase 对应本次选定的一个 AD 版本；后续方法改进可形成新 TestCase 版本而不强制改变 AD 或 PRD 版本。同一能力其他主题可以有自己的用例集，组合交付时显式选择配套集合及集成范围，不自动合并目录中全部最新文件。
+本 TestCase 对应本次选定的一个 AD 版本；后续方法改进可形成新 TestCase 版本而不强制改变 AD 或 PRD 版本。同一能力其他PBC可以有自己的用例集，组合交付时显式选择配套集合及集成范围，不自动合并目录中全部最新文件。
 
-### 0.2 跨主题采用依据与原方法分工
+### 0.2 跨PBC采用依据与原方法分工
 
-先读本主题选定PRD与直接配套AD，再按实际共享定义读取项目已采用公共技术规约、契约正文和必要关联主题。关联AD／TestCase只用于本次组合和交接，不替代本主题产品来源，不把整目录最大版本或未采用候选纳入。基本引用按实际路径、正文版本与条款核对，摘要可选；来源失效或冲突列Q及受影响范围，不静默换源，也不从历史测试恢复旧答案。
+先读本PBC选定PRD与直接配套AD，再按实际共享定义读取项目已采用公共技术规约、契约正文和必要关联PBC。关联AD／TestCase只用于本次组合和交接，不替代本PBC产品来源，不把整目录最大版本或未采用候选纳入。基本引用按实际路径、正文版本与条款核对，摘要可选；来源失效或冲突列Q及受影响范围，不静默换源，也不从历史测试恢复旧答案。
 
 | 方法／产物 | 职责与输入 | 输出及权限边界 |
 | --- | --- | --- |
@@ -66,7 +66,7 @@
 
 ## 1. 验证目标、范围与排除
 
-`{本主题需要证明的业务行为、全部AC/复合子项、关键端到端流程、相关工程质量及受影响既有功能。产品功能验收与运行后业务效果分别说明；未知阈值或观察周期标待确认，不编造。}`
+`{本PBC需要证明的业务行为、全部AC/复合子项、关键端到端流程、相关工程质量及受影响既有功能。产品功能验收与运行后业务效果分别说明；未知阈值或观察周期标待确认，不编造。}`
 
 | 验证对象 | 本次纳入内容 | 排除/暂不能执行范围及理由 | 依赖/风险 | 所需方法 |
 | --- | --- | --- | --- | --- |
@@ -113,7 +113,7 @@
 
 本章定义所需条件；它们的实际就绪情况在执行前独立核实。示例身份不等于真实授权，模拟服务不等于已接通真实依赖，不将生产数据或高风险故障注入当默认测试材料。
 
-### 3.3 跨主题组合的执行条件
+### 3.3 跨PBC组合的执行条件
 
 | 条件ID／组合ID | 提供／消费方及受影响用例 | 规范／契约／数据／实现的选定引用 | 就绪核实与真实／替身边界 | 准备、失败、清理及恢复 | 缺口及可继续部分 |
 | --- | --- | --- | --- | --- | --- |
@@ -157,21 +157,21 @@
 
 维度用于核对，不要求无相关需求的产品强行增加功能或专项；不适用要有明确依据。
 
-## 6. 跨主题组合、变化与回归
+## 6. 跨PBC组合、变化与回归
 
-| 变更前基准/本次目标 | 受影响主题/功能/状态/契约 | 保留/修订/新增/停止适用的用例 | 选择依据与风险 | 必须重取的证据 |
+| 变更前基准/本次目标 | 受影响PBC/功能/状态/契约 | 保留/修订/新增/停止适用的用例 | 选择依据与风险 | 必须重取的证据 |
 | --- | --- | --- | --- | --- |
 | `{PRD/AD/用例/代码固定引用}` | `{精确范围}` | `{版本与替代关系}` | `{未变ID不等于未变预期}` | `{需求/实现/依赖/环境变化影响}` |
 
-多主题组合登记各自 PRD/AD/TestCase 版本，检查共享状态/字段/权限/接口一致性与交接，不以单份检查通过代替组合验证。历史运行仍保留在原对象和环境版本下；未受影响的证据也需说明适用依据，不直接复制为新目标通过。
+多PBC组合登记各自 PRD/AD/TestCase 版本，检查共享状态/字段/权限/接口一致性与交接，不以单份检查通过代替组合验证。历史运行仍保留在原对象和环境版本下；未受影响的证据也需说明适用依据，不直接复制为新目标通过。
 
 ### 6.1 明确本次兼容组合
 
-| 组合ID／作用范围 | 本主题PRD／AD／TestCase实际路径和各自版本 | 提供／消费方主题及配套引用 | 公共规范条款／契约与数据权威 | 实现／制品／环境事实 | 设计判断、执行状态与缺口 |
+| 组合ID／作用范围 | 本PBCPRD／AD／TestCase实际路径和各自版本 | 提供／消费方PBC及配套引用 | 公共规范条款／契约与数据权威 | 实现／制品／环境事实 | 设计判断、执行状态与缺口 |
 | --- | --- | --- | --- | --- | --- |
-| `{独立主题或本次集成组合}` | `{各自documentId、准确路径及文内版本，以正文及真实采用为准，不强制同号}` | `{提供／消费方ebcId／pbcId／prdId及配套引用；仅相关范围，缺关联TestCase如实记，不造未来文件路径}` | `{来源引用ID、准确位置／版本／条款；实际共享定义不在用例中复制另一权威}` | `{已有／拟建／未知；代码或服务不存在不填已部署}` | `{相容／冲突／待核实与依据；不预填运行通过}` |
+| `{独立PBC或本次集成组合}` | `{各自documentId、准确路径及文内版本，以正文及真实采用为准，不强制同号}` | `{提供／消费方ebcId／pbcId及配套引用；仅相关范围，缺关联TestCase如实记，不造未来文件路径}` | `{来源引用ID、准确位置／版本／条款；实际共享定义不在用例中复制另一权威}` | `{已有／拟建／未知；代码或服务不存在不填已部署}` | `{相容／冲突／待核实与依据；不预填运行通过}` |
 
-一个主题TestCase仍对应本主题选定PRD和一个直接AD版本，关联引用用于组合场景；不用总用例文档包办全部主题，不要求双方版本数值一致。文档、契约、代码、服务各自版本属于不同对象，不能跨对象比较大小或据同号／SemVer／Schema校验宣布语义兼容。默认值、枚举、权限、错误、副作用及严格消费者均应核对。
+本PBC的一份逻辑TestCase对应选定PRD和一个直接AD版本，关联引用用于组合场景；不用总用例文档包办全部PBC，不要求双方版本数值一致。文档、契约、代码、服务各自版本属于不同对象，不能跨对象比较大小或据同号／SemVer／Schema校验宣布语义兼容。默认值、枚举、权限、错误、副作用及严格消费者均应核对。
 
 ### 6.2 跨边界场景与最终权威
 
@@ -191,11 +191,11 @@
 
 ### 6.4 变更影响与既有行为回归
 
-| 变化条款／权威位置 | 直接及必要传递影响 | 受影响主题AD／AC／用例及已有任务 | 必须保留／补取的回归证据 | 无关范围与不受影响依据 | 处理决定／Q |
+| 变化条款／权威位置 | 直接及必要传递影响 | 受影响PBCAD／AC／用例及已有任务 | 必须保留／补取的回归证据 | 无关范围与不受影响依据 | 处理决定／Q |
 | --- | --- | --- | --- | --- | --- |
-| `{公共规范、契约、实体、代码或环境的实际变化}` | `{定义→提供方→消费者→必要下游}` | `{有真实任务才登记，不造Plan或tracker项}` | `{旧调用方、拒绝、未知与恢复及既有业务行为}` | `{未消费该定义或明确不受影响；不机械改写全部主题}` | `{降级／移出必需回归要真实理由与决定}` |
+| `{公共规范、契约、实体、代码或环境的实际变化}` | `{定义→提供方→消费者→必要下游}` | `{有真实任务才登记，不造Plan或tracker项}` | `{旧调用方、拒绝、未知与恢复及既有业务行为}` | `{未消费该定义或明确不受影响；不机械改写全部PBC}` | `{降级／移出必需回归要真实理由与决定}` |
 
-公共变化不能只查本主题，也不能把所有AD／PRD统一升版。共享底座、迁移和测试夹具的后续工程修改应在唯一任务目标安排归属，相关任务列依赖；本次仅设计，不生成第二份Plan或竞争tracker，不改人工代码及其他主题。
+公共变化不能只查本PBC，也不能把所有AD／PRD统一升版。共享底座、迁移和测试夹具的后续工程修改应在唯一任务目标安排归属，相关任务列依赖；本次仅设计，不生成第二份Plan或竞争tracker，不改人工代码及其他PBC。
 
 ## 7. 两种结论与执行准入
 
@@ -205,7 +205,7 @@
 
 **TestCase是工程SPEC和Plan的可选输入，不是统一前置文件。**T2可从PRD、AD及已采用规范形成含充分测试策略／验收预期的工程SPEC并交审，T3再按选定工程SPEC调用原Plan；没有独立TestCase不要求先补一份才能规划。已有且适用的TestCase应读取准确版本，经工程SPEC衔接并落实到任务验收和验证步骤；缺少或待补的用例设计可作为计划任务，相关验证定义在依赖它的实现、测试或验收之前明确。核心行为或技术判据尚不明确时，只限制相应依赖任务，不把计划先行等同编码、测试或发布许可。
 
-AD或现有测试规范可提供测试策略依据，由T2落实到工程SPEC六项中的相应内容；不能因TestCase可选而跳过六项核对。企业SOP仍保留本主题TestCase设计交付，允许其与T2／T3按依赖衔接，不取消正式交付所要求的验证定义。用例与任务、脚本可以多对多；只有真实脚本和任务形成后才登记实际路径，人工用例明确执行方式，不伪造全部自动化。
+AD或现有测试规范可提供测试策略依据，由T2落实到工程SPEC六项中的相应内容；不能因TestCase可选而跳过六项核对。企业SOP仍保留本PBCTestCase设计交付，允许其与T2／T3按依赖衔接，不取消正式交付所要求的验证定义。用例与任务、脚本可以多对多；只有真实脚本和任务形成后才登记实际路径，人工用例明确执行方式，不伪造全部自动化。
 
 ### 7.2 执行可准入（候选形成后另存核实记录）
 
@@ -228,14 +228,14 @@ AD或现有测试规范可提供测试策略依据，由T2落实到工程SPEC六
 
 ## 8. 测试报告合同：一份用例集，多次运行
 
-测试报告另存于能力目录的实际报告位置，按项目登记解析；命名采用 `{PBC}-{需求主题}-TestReport_v{报告版本}.md`，不强制跟 TestCase 同号。报告版本标识文档修订，`runId` 标识一次真实运行；报告编辑不伪造新运行，重测产生的新运行不覆盖前次失败。
+测试报告另存于能力目录的实际报告位置，按项目登记解析；命名采用 `{PBC名称}-TestReport_v{报告版本}.md`，不强制跟 TestCase 同号。报告版本标识文档修订，`runId` 标识一次真实运行；报告编辑不伪造新运行，重测产生的新运行不覆盖前次失败。
 
 ### 8.1 报告必须记录的事实
 
 | 报告字段 | 执行后记录要求 |
 | --- | --- |
 | 报告身份与运行 | `{报告ID/版本、runId、实际起止时间、执行主体、实际工具/运行方式}` |
-| 本次选择 | `{ebcId／ebcName、pbcId／pbcName、prdId；PRD/AD/TestCase的documentId、准确路径、文内版本及固定内容，摘要按需；多主题组合逐项登记}` |
+| 本次选择 | `{ebcId／ebcName、pbcId／pbcName、PRD documentId；PRD/AD/TestCase的documentId、准确路径、文内版本及固定内容，摘要按需；多PBC组合逐项登记}` |
 | 本次组合 | `{§6.1组合ID；提供／消费方文档、契约、数据、代码／服务及环境各自版本；设计配套与真实兼容分开}` |
 | 实际候选 | `{源提交、构建/镜像制品、摘要；实际启动/部署对象，未部署不能写已部署}` |
 | 实际环境 | `{配置、依赖、身份/权限、数据版本及来源、真实/模拟边界；不泄露密钥}` |
@@ -265,13 +265,13 @@ AD或现有测试规范可提供测试策略依据，由T2落实到工程SPEC六
 
 ### 完成核对
 
-- PRD 直接覆盖与 AD 工程覆盖均完整，复合子项、跨主题交接及必要回归未漏。
+- PRD 直接覆盖与 AD 工程覆盖均完整，复合子项、跨PBC交接及必要回归未漏。
 - 每个用例九要素明确，允许/禁止副作用可判定，阈值和预期有来源。
 - 所需环境/数据与实际可用状态分开，替身不冒充真实集成。
 - 框架／工具版本、测试目录、层级和适用覆盖要求明确，与 AD 命令／cwd 及项目质量权威一致；未安装、未执行及不适用范围如实登记。
 - 规格可交接、执行可准入、实际运行、业务认可分别记录，无循环开工条件。
 - 报告字段和1:N运行关系明确，未执行不预填通过；旧证据和失败记录保留。
-- §0.2、§3.3、§6.1—§6.4中的准确组合、唯一权威、失败恢复、传递影响和无关主题保护有实际来源。
+- §0.2、§3.3、§6.1—§6.4中的准确组合、唯一权威、失败恢复、传递影响和无关PBC保护有实际来源。
 - §7.1不把完整TestCase设为T2工程SPEC或T3原Plan的统一前置；原test／Build仍执行TDD，原代码review与零写入文档审查职责分开。
 
 交接结论：`{本次供T2工程SPEC及后续T3任务采用的验证要求、尚不能确定的判据和影响；实际执行须另行准入。当前版本及真实认可范围如实记录。}`

package/skills/i3-frontend-ui-engineering/SKILL.md

@@ -29,10 +29,6 @@ Build production-quality user interfaces that are accessible, performant, and vi
 - Adding interactivity or state management
 - Fixing visual or UX issues
 
-## Applicable Prototype Baselines
-
-Read the project prototype index at `prototypes/index.json`, falling back to legacy `.i3-ui-prototypes.json` when no new index exists. Select the entry whose `appliesTo` contains the explicit business module ID; an empty array applies to no module, and conflicting indices or multiple matches require resolution before selecting a baseline. New associations are derived from fixed scope.moduleIds by `i3-ui-prototype`, not inferred from a PRD filename. Then read its `baseline.json`, `design-system/MASTER.md`, tokens and page rules. Preserve approved design and interaction decisions; explain conflicts with new requirements instead of silently changing them. New pages reuse the selected template and component conventions. Changes to global tokens, navigation, component behavior or business states go through the `i3-ui-prototype` workflow and corresponding review. HTML review prototypes need explicit component mapping when implemented in another framework. Verify affected production pages in the browser; a prototype manifest check alone does not prove production fidelity. Projects without an applicable baseline retain this skill's existing workflow.
-
 ## Component Architecture
 
 ### File Structure
@@ -95,20 +91,20 @@ export function TaskItem({ task, onToggle, onDelete }: TaskItemProps) {
 ```tsx
 // Container: handles data
 export function TaskListContainer() {
-  const { tasks, isLoading, error, refetch } = useTasks();
+  const { tasks, isLoading, error, refetch, toggleTask, deleteTask } = useTasks();
 
   if (isLoading) return <TaskListSkeleton />;
   if (error) return <ErrorState message="Failed to load tasks" retry={refetch} />;
   if (tasks.length === 0) return <EmptyState message="No tasks yet" />;
 
-  return <TaskList tasks={tasks} />;
+  return <TaskList tasks={tasks} onToggle={toggleTask} onDelete={deleteTask} />;
 }
 
 // Presentation: handles rendering
-export function TaskList({ tasks }: { tasks: Task[] }) {
+export function TaskList({ tasks, onToggle, onDelete }: { tasks: Task[] } & Pick<TaskItemProps, 'onToggle' | 'onDelete'>) {
   return (
     <ul role="list" className="divide-y">
-      {tasks.map(task => <TaskItem key={task.id} task={task} />)}
+      {tasks.map(task => <TaskItem key={task.id} task={task} onToggle={onToggle} onDelete={onDelete} />)}
     </ul>
   );
 }
@@ -250,7 +246,7 @@ function Dialog({ isOpen, onClose }: DialogProps) {
 
 ```tsx
 // Don't show blank screens
-function TaskList({ tasks }: { tasks: Task[] }) {
+function TaskList({ tasks, onCreateTask }: { tasks: Task[]; onCreateTask: () => void }) {
   if (tasks.length === 0) {
     return (
       <div role="status" className="text-center py-12">
@@ -288,7 +284,7 @@ Test at these breakpoints: 320px, 768px, 1024px, 1440px.
 // Skeleton loading (not spinners for content)
 function TaskListSkeleton() {
   return (
-    <div className="space-y-3" aria-busy="true" aria-label="Loading tasks">
+    <div className="space-y-3" role="status" aria-busy="true" aria-label="Loading tasks">
       {Array.from({ length: 3 }).map((_, i) => (
         <div key={i} className="h-12 bg-muted animate-pulse rounded" />
       ))}

package/skills/i3-idea-refine/SKILL.md

@@ -40,8 +40,10 @@ Refines raw ideas into sharp, actionable concepts worth building through structu
 This skill is primarily an interactive dialogue. Invoke it with an idea, and the agent will guide you through the process.
 
 ```bash
-# Optional: Initialize the ideas directory
-bash skills/i3-idea-refine/scripts/idea-refine.sh
+# Optional. Run from the project root so this creates ./docs/ideas.
+# <skill-dir> is the directory that contains this SKILL.md
+# (for example .agents/skills/i3-idea-refine or .cursor/skills/i3-idea-refine).
+bash <skill-dir>/scripts/idea-refine.sh
 ```
 
 ## Output

package/skills/i3-incremental-implementation/SKILL.md

@@ -192,10 +192,6 @@ Each increment should be independently revertable:
 - Database migrations should have corresponding rollback migrations
 - Avoid deleting something in one commit and replacing it in the same commit — separate them
 
-## UI Slices with Prototype Baselines
-
-For a UI slice, check `prototypes/index.json` (legacy `.i3-ui-prototypes.json` fallback) for an applicable approved prototype. Match explicit module IDs against appliesTo; an empty array is unbound. Resolve conflicting indices or multiple matches instead of applying a prototype globally. New appliesTo values come from fixed scope.moduleIds through `i3-ui-prototype`. Read its design contract, business mapping and baseline before implementation, and include affected interaction/visual checks in the slice acceptance. Preserve prior snapshots; coordinate contract changes through `i3-ui-prototype`. Backend-only slices and projects without a baseline continue the existing increment workflow.
-
 ## Working with Agents
 
 When directing an agent to implement incrementally:

package/skills/i3-performance-optimization/SKILL.md

@@ -217,8 +217,8 @@ Lighthouse Performance score: ≥ 90
 # Bundle size check
 npx bundlesize --config bundlesize.config.json
 
-# Lighthouse CI
-npx lhci autorun
+# Lighthouse CI. The CLI package is @lhci/cli; unscoped `lhci` on npm is not it.
+npx --package=@lhci/cli lhci autorun
 ```
 
 ## See Also

package/skills/i3-security-and-hardening/references/hardening-patterns.md

@@ -73,6 +73,11 @@ const clean = DOMPurify.sanitize(userInput);
 // Always check authorization, not just authentication
 app.patch('/api/tasks/:id', authenticate, async (req, res) => {
   const task = await taskService.findById(req.params.id);
+  if (!task) {
+    return res.status(404).json({
+      error: { code: 'NOT_FOUND', message: 'Task not found' }
+    });
+  }
 
   // Check that the authenticated user owns this resource
   if (task.ownerId !== req.user.id) {

package/skills/i3-source-driven-development/SKILL.md

@@ -170,8 +170,9 @@ form submission state. React 19 replaced the manual
 isPending/setIsPending pattern with this hook.
 
 Source: https://react.dev/blog/2024/12/05/react-19#actions
-"useTransition now supports async functions [...] to handle
-pending states automatically"
+"In React 19, we're adding support for using async functions in
+transitions to handle pending states, errors, forms, and optimistic
+updates automatically."
 ```
 
 **Citation rules:**

package/skills/i3-spec-driven-development/SKILL.md

@@ -101,7 +101,7 @@ Don't silently fill in ambiguous requirements. The spec's entire purpose is to s
    ```
    Build: npm run build
    Test: npm test -- --coverage
-   Lint: npm run lint --fix
+   Lint: npm run lint -- --fix
    Dev: npm run dev
    ```
 

package/skills/i3-ui-demo/SKILL.md

@@ -1,7 +1,7 @@
 ---
 name: i3-ui-demo
 description: "基于业务文档或提示词和内置模板创建、增量修改 Vue 3 与 antdv-next 1.6.0 交互原型。适用于明确请求 ui-demo、模板驱动的界面演示与只读方案分析，以及新增页面、调整设计或修复交互；结合自有能力定义、原生组件映射、模板视觉变量和专业技能完成生成与检查。"
-# i3-description-source: "description: Creates and incrementally updates runnable Vue 3 and antdv-next 1.6.0 UI demos from business documents or prompts using owned templates and required UI/UX Pro Max. Use when requesting ui-demo, a template-based UI demo, or adding pages and revising its design and interactions."
+# i3-description-source: "description: Creates reviewable interactive UI prototypes from business requirements, documents or prompts, and incrementally updates runnable Vue 3 and antdv-next 1.6.0 UI demos using owned templates and required UI/UX Pro Max. Use when requesting ui-demo, an interactive UI prototype, a template-based UI demo, or adding pages and revising its design and interactions."
 ---
 
 <!-- i3:zh-CN:start -->
@@ -187,6 +187,8 @@ Skill/Command/Agent 标识、文件路径、命令参数、代码/API 名称、
 分别说明实现完成情况、技术验证结果和用户评审状态；技术通过仍可待用户评审。
 认可必须对应具体版本，不从构建、浏览器结果或技能记录推断用户批准。
 
+来自企业 SOP 的任务沿用已选 PRD 的路径、版本及已确定的 EBC／PBC 和主题身份，在现有 `HANDOFF.md` 记录来源、当前实现版本、设计与 QA 位置及实际评审状态，供 PRD／AD 固定引用。普通提示词任务仍可独立开展；不要求为制作原型补齐企业文档，不创建另一套索引或基线管理流程。已有业务原型先读取结构和版本，按明确范围增量维护，保留人工修改和历史。
+
 ## Incremental Changes
 
 先读取现有代码、全局设计系统和相关页面规则，不整体重新生成。

package/skills/i3-ui-demo/references/dependencies.md

@@ -49,7 +49,7 @@ global 仅在用户已经授权对应全局宿主范围时使用；优先项目
 - antDv 官方源固定提交和下载摘要；使用 `assets/antdv-build/` 的锁定构建环境编译未修改的官方源码，不用 npm 上缺失目标数据的 CLI 包替代。
 - 固定的数据文件、索引及官方源码摘要可核对；构建环境只用于工具准备，不进入业务项目 app/。
 - 官方构建器需要 Node.js 22.18.x 或 24.11+。优先当前 Node，否则只读查找 PATH 中已有兼容 Node；缺失时在安装前说明要求，不自动升级全局 Node。已经准备好的 CLI 可在兼容的 Node 20 环境继续复用，不触发重建。
-- 新工具使用独立的宿主工具目录，不覆盖旧 i3-prototype 工具或旧组件配置。
+- 新工具使用独立的宿主工具目录，保留已有工具和组件配置。
 - 版本不兼容或无法验证的技能和工具保留并报告，不按“文件存在”判为就绪。安装完成必须两项均通过；CLI 的较高版本不改变 antdv-next 1.6.0 和知识 1.6.0。
 
 支持用 `--archive <本地官方源码tgz>` 复用已下载的同一固定归档，仍校验其摘要。

package/skills/i3-ui-demo/references/resources.md

@@ -129,7 +129,7 @@ python3 "<ui-ux-pro-max技能绝对目录>/scripts/search.py" "<业务与设计
 - `requirements-map.json`：原文需求到页面、操作、状态、实施文件及验证场景的对应。
 - `design-system/<project-slug>/MASTER.md` 与 pages/：ui-ux-pro-max 的原生设计系统与页面覆盖。
 - `qa/`：对应当前代码和构建的真实检查结果、截图及未完成项。
-- `HANDOFF.md`：本次完成范围、用户选择、版本、启动与预览入口、缺项和下一步。
+- `HANDOFF.md`：准确需求来源及版本、本次完成范围、用户选择、当前实现版本、设计／QA 位置、实际认可范围、启动与预览入口、缺项和下一步。SOP 场景沿用已确定的 EBC／PBC 与主题身份，供 PRD／AD 引用。
 
 这些名称用于新实例约定；已有原型沿用现有资料，不为了统一格式自动迁移。
 只维护一份任务状态。需求映射保存来源和对应关系，不再创建第二份独立任务勾选清单。

package/skills/i3-ui-demo/scripts/dependencies.cjs

@@ -100,9 +100,6 @@ function resolveOfficial(location, options = {}) {
     if (fs.existsSync(directory)) for (const name of fs.readdirSync(directory).sort()) {
       if (name.startsWith('antdv-')) candidates.push(path.join(directory, name, 'source'));
     }
-    // Read the old install's version; a required upgrade is prepared in the
-    // new owned directory without deleting the old tool or its user files.
-    candidates.push(path.join(base, 'tools/i3-prototype/antdv-cli/node_modules/@antdv-next/cli'));
   }
   candidates.push(path.join(location.project, 'node_modules/@antdv-next/cli'), path.resolve(path.dirname(process.execPath), '../lib/node_modules/@antdv-next/cli'));
   const env = options.env || process.env;

package/skills/i3-using-agent-skills/SKILL.md

@@ -33,7 +33,7 @@ Task arrives
     ├── New project/feature/change? ──→ i3-spec-driven-development
     ├── No quality bar written down? ──→ i3-constraint-driven-development
     ├── Have a spec, need tasks? ──────→ i3-planning-and-task-breakdown
-    ├── UI prototype work? ───────────→ i3-ui-prototype
+    ├── UI prototype work? ───────────→ i3-ui-demo
     ├── Implementing code? ────────────→ i3-incremental-implementation
     │   ├── UI work? ─────────────────→ i3-frontend-ui-engineering
     │   ├── API work? ────────────────→ i3-api-and-interface-design
@@ -184,7 +184,7 @@ Not every task needs every skill. A bug fix might only need: `i3-debugging-and-e
 | Define | i3-idea-refine | Refine ideas through structured divergent and convergent thinking |
 | Define | i3-spec-driven-development | Requirements and acceptance criteria before code |
 | Plan | i3-planning-and-task-breakdown | Decompose into small, verifiable tasks |
-| Prototype | i3-ui-prototype | Create, inspect and iterate reviewable UI prototypes |
+| Prototype | i3-ui-demo | Create, inspect and iterate reviewable UI prototypes |
 | Build | i3-incremental-implementation | Thin vertical slices, test each before expanding |
 | Build | i3-source-driven-development | Verify against official docs before implementing |
 | Build | i3-doubt-driven-development | Adversarial fresh-context review of every non-trivial decision |

package/.claude/commands/i3-prototype.md

@@ -1,23 +0,0 @@
----
-description: "在主会话中分析、创建、检查与迭代 UI 交互原型。"
-# i3-description-source: "description: Create, analyze, inspect and iterate UI prototypes in the main session"
----
-
-<!-- i3:zh-CN:start -->
-## 中文交互与输出
-
-默认使用简体中文回复，包括追问、选项说明、阶段进度、分析、规格、计划、验收、代码审查与发布报告。只有用户明确要求其他语言时才切换；输入中出现英文术语、代码或引用不构成切换语言的要求。
-下文的工程步骤、审批要求、退出条件和工具约束完整有效。英文示例中的人类可读标题、说明和占位内容在实际输出时应译为中文；保留原有信息与验证标准，不删减方法。
-Skill/Command/Agent 标识、文件路径、命令参数、代码/API 名称、工具调用参数、JSON 键/类型/枚举、机器状态、测试期望字面量及引用原文保持原样。生成文档正文用中文，SPEC.md、tasks/plan.md、tasks/todo.md、CONSTRAINTS.md 等约定路径不改名。
-严重程度按原规则判断，面向用户可写“严重（Critical）”“必须修复（Required）”“小问题（Nit）”“可选（Optional）”“建议考虑（Consider）”“仅供参考（FYI）”；High/Medium/Low、APPROVE/REQUEST CHANGES 和 GO/NO-GO 保留其原始语义与机器值。不得因翻译降低门槛，角色不能互相委派。
-
-适用说明：在主会话中分析、创建、检查与迭代 UI 交互原型。
-<!-- i3:zh-CN:end -->
-
-Use the i3-ui-prototype skill in the main session for the user's request below. Follow its shared preflight, operation boundaries and applicable runtime workflow.
-
-$ARGUMENTS
-
-With no arguments use create; explicit analysis or read-only limits take precedence. Preserve the user's scope, manual edits, fixed inputs and approved baselines. Use supporting engineering, framework, design and browser skills as the working skill requires; keep its workflow and evidence requirements there.
-
-Create the prototype in the current session. Load the required frontend/framework/design Skills before implementation and use the browser Skill for interaction acceptance. Installed Hooks run automatically; do not inspect Hook internals, search host logs, or run a manual binding step. When a required Skill is missing, invoke it and retry; report an infrastructure failure briefly through the separate maintenance entry.

package/.gemini/commands/i3-prototype.toml

@@ -1,23 +0,0 @@
-description = "在主会话中分析、创建、检查与迭代 UI 交互原型。"
-# i3-description-source: "description = \"Create, analyze, inspect and iterate UI prototypes in the main session\""
-
-prompt = """
-<!-- i3:zh-CN:start -->
-## 中文交互与输出
-
-默认使用简体中文回复，包括追问、选项说明、阶段进度、分析、规格、计划、验收、代码审查与发布报告。只有用户明确要求其他语言时才切换；输入中出现英文术语、代码或引用不构成切换语言的要求。
-下文的工程步骤、审批要求、退出条件和工具约束完整有效。英文示例中的人类可读标题、说明和占位内容在实际输出时应译为中文；保留原有信息与验证标准，不删减方法。
-Skill/Command/Agent 标识、文件路径、命令参数、代码/API 名称、工具调用参数、JSON 键/类型/枚举、机器状态、测试期望字面量及引用原文保持原样。生成文档正文用中文，SPEC.md、tasks/plan.md、tasks/todo.md、CONSTRAINTS.md 等约定路径不改名。
-严重程度按原规则判断，面向用户可写“严重（Critical）”“必须修复（Required）”“小问题（Nit）”“可选（Optional）”“建议考虑（Consider）”“仅供参考（FYI）”；High/Medium/Low、APPROVE/REQUEST CHANGES 和 GO/NO-GO 保留其原始语义与机器值。不得因翻译降低门槛，角色不能互相委派。
-
-适用说明：在主会话中分析、创建、检查与迭代 UI 交互原型。
-<!-- i3:zh-CN:end -->
-
-Use the i3-ui-prototype skill in the main session for the user's request below. Follow its shared preflight, operation boundaries and applicable runtime workflow.
-
-$ARGUMENTS
-
-With no arguments use create; explicit analysis or read-only limits take precedence. Preserve the user's scope, manual edits, fixed inputs and approved baselines. Use supporting engineering, framework, design and browser skills as the working skill requires; keep its workflow and evidence requirements there.
-
-Create the prototype in the current session. Load the required frontend/framework/design Skills before implementation and use the browser Skill for interaction acceptance. Installed Hooks run automatically; do not inspect Hook internals, search host logs, or run a manual binding step. When a required Skill is missing, invoke it and retry; report an infrastructure failure briefly through the separate maintenance entry.
-"""

package/commands/i3-prototype.toml

@@ -1,23 +0,0 @@
-description = "在主会话中分析、创建、检查与迭代 UI 交互原型。"
-# i3-description-source: "description = \"Create, analyze, inspect and iterate UI prototypes in the main session\""
-
-prompt = """
-<!-- i3:zh-CN:start -->
-## 中文交互与输出
-
-默认使用简体中文回复，包括追问、选项说明、阶段进度、分析、规格、计划、验收、代码审查与发布报告。只有用户明确要求其他语言时才切换；输入中出现英文术语、代码或引用不构成切换语言的要求。
-下文的工程步骤、审批要求、退出条件和工具约束完整有效。英文示例中的人类可读标题、说明和占位内容在实际输出时应译为中文；保留原有信息与验证标准，不删减方法。
-Skill/Command/Agent 标识、文件路径、命令参数、代码/API 名称、工具调用参数、JSON 键/类型/枚举、机器状态、测试期望字面量及引用原文保持原样。生成文档正文用中文，SPEC.md、tasks/plan.md、tasks/todo.md、CONSTRAINTS.md 等约定路径不改名。
-严重程度按原规则判断，面向用户可写“严重（Critical）”“必须修复（Required）”“小问题（Nit）”“可选（Optional）”“建议考虑（Consider）”“仅供参考（FYI）”；High/Medium/Low、APPROVE/REQUEST CHANGES 和 GO/NO-GO 保留其原始语义与机器值。不得因翻译降低门槛，角色不能互相委派。
-
-适用说明：在主会话中分析、创建、检查与迭代 UI 交互原型。
-<!-- i3:zh-CN:end -->
-
-Use the i3-ui-prototype skill in the main session for the user's request below. Follow its shared preflight, operation boundaries and applicable runtime workflow.
-
-$ARGUMENTS
-
-With no arguments use create; explicit analysis or read-only limits take precedence. Preserve the user's scope, manual edits, fixed inputs and approved baselines. Use supporting engineering, framework, design and browser skills as the working skill requires; keep its workflow and evidence requirements there.
-
-Create the prototype in the current session. Load the required frontend/framework/design Skills before implementation and use the browser Skill for interaction acceptance. Installed Hooks run automatically; do not inspect Hook internals, search host logs, or run a manual binding step. When a required Skill is missing, invoke it and retry; report an infrastructure failure briefly through the separate maintenance entry.
-"""

package/evals/cases/i3-ui-prototype.json

@@ -1,158 +0,0 @@
-{
-  "skill_name": "i3-ui-prototype",
-  "trigger": {
-    "positive": [
-      {
-        "prompt": "Choose a UI prototype template and generate an interactive approval prototype for review",
-        "top_k": 1
-      },
-      {
-        "prompt": "Generate a reviewable interactive UI prototype from business requirements",
-        "top_k": 1
-      },
-      {
-        "prompt": "Inspect prototype fidelity against its approved prototype baseline",
-        "top_k": 1
-      },
-      {
-        "prompt": "Iterate the interactive prototype using review feedback and preserve the approved baseline",
-        "top_k": 1
-      },
-      {
-        "prompt": "Use this editable HTML reference to generate a device management prototype with native Vue and antDv-next components",
-        "top_k": 1
-      },
-      {
-        "prompt": "Analyze a plan for turning this content template and business brief into an interactive prototype, without changing any files",
-        "top_k": 1
-      },
-      {
-        "prompt": "Revise the prototype using my feedback and preserve my manual edits while reviewing the drawer animation",
-        "top_k": 1
-      },
-      {
-        "prompt": "Use the prototype skill directly to inspect our review candidate without starting another agent",
-        "top_k": 1
-      }
-    ],
-    "negative": [
-      {
-        "prompt": "Build an accessible modal component with keyboard support",
-        "owner": "i3-frontend-ui-engineering"
-      },
-      {
-        "prompt": "Optimize slow database indexes",
-        "owner": "i3-performance-optimization"
-      },
-      {
-        "prompt": "Implement the approved production design as an accessible reusable Vue form component",
-        "owner": "i3-frontend-ui-engineering"
-      },
-      {
-        "prompt": "Debug an existing application failure: the save request broke after a network error",
-        "owner": "i3-debugging-and-error-recovery"
-      }
-    ]
-  },
-  "evals": [
-    {
-      "id": 1,
-      "prompt": "Create an interactive application approval prototype from the supplied business brief. Use template mode explicitly, compare the templates and generate a review candidate. Do not approve it yourself.",
-      "expected_output": "A runnable business-specific prototype with traceability and pending-review status",
-      "files": [
-        "i3-ui-prototype"
-      ],
-      "expectations": [
-        "Selected template and template mode are recorded without claiming UI/UX Pro Max was used",
-        "Business fields, role visibility, status transitions and review scenarios follow the supplied brief",
-        "A runnable prototype and verification evidence are produced; absent browser evidence is marked pending",
-        "The prototype remains pending human review and no approved baseline is invented"
-      ]
-    },
-    {
-      "id": 2,
-      "kind": "dialogue",
-      "prompt": "Use the i3-ui-prototype skill directly in this session to analyze how an editable HTML content template and capability list could produce a Vue 3 plus antDv-next device-list prototype. This is analysis only: do not change files, install or upgrade tools, persist design output, or start a separate agent.",
-      "expected_output": "A read-only prototype analysis that keeps direct Skill execution valid and states the framework, content, capability and acceptance boundaries",
-      "expectations": [
-        "The current executor runs the shared preflight without redispatching to an Agent, command or meta-skill",
-        "The response preserves analysis-only scope and does not initialize, install, persist design output or update project execution state",
-        "Editable content reference, reusable capabilities and target-framework behavior are explained as separate inputs",
-        "Unavailable provider, host or browser capabilities remain explicitly unverified rather than reported as executed",
-        "The analysis deliverable is the conversation; it does not invent prototype completion or human approval"
-      ]
-    },
-    {
-      "id": 3,
-      "prompt": "Create a separate device-list review prototype with the admin-list editable content template and native Vue 3 plus antDv-next components. For this case explicitly use template mode. Requirements: 21 deterministic devices, 10 per page; filtering or changing page clears current-page selection; batch status changes affect only selected records; an edit drawer validates required name, preserves input on simulated save failure, retries, and updates the correct row. Verify normal drawer exit and mask blur on the current build when a real browser is available. Preserve any existing prototype and leave this candidate for human review.",
-      "expected_output": "An isolated native component prototype with fixed capability inputs, reproducible local data, current-build verification or a truthful verification gap",
-      "files": [
-        "i3-ui-prototype"
-      ],
-      "expectations": [
-        "The instance uses the native framework runtime and fixed template, capability, scope and framework inputs rather than converting an existing legacy instance",
-        "Real antDv-next components implement filtering, current-page selection, pagination and the edit flow with deterministic isolated mock state",
-        "Required capability bindings remain independently checkable and are not removed or arbitrarily marked inapplicable to make checks pass",
-        "The executor distinguishes source inspection, successful build and real browser execution; unavailable checks remain unverified",
-        "Browser evidence, when present, refers to the current candidate and actual build, with drawer motion and blur verified separately from static screenshots",
-        "Template mode is explicit and does not claim UIUX use; existing prototypes remain unchanged and no human baseline approval is invented"
-      ]
-    },
-    {
-      "id": 4,
-      "kind": "dialogue",
-      "prompt": "Inspect this prototype verification claim without modifying files: the source was changed after testing, the server still serves the previous build, and the only evidence is a screenshot plus a report manually marked passed. A teammate suggests removing the pagination capability and marking it inapplicable so we can approve the candidate. Explain the valid next step; do not execute a new test or approve a baseline.",
-      "expected_output": "A read-only rejection of stale evidence and unsupported exemption, with concrete required revalidation and pending human review",
-      "expectations": [
-        "The changed candidate and previous build invalidate the supplied acceptance claim",
-        "A screenshot or manually passed report is not treated as proof of current-build browser execution or drawer animation",
-        "Pagination remains required unless an explicit requirement or scope-change source justifies an exemption",
-        "The response proposes rebuilding and revalidating against fixed inputs without claiming those actions already occurred",
-        "No files are changed, no tests are pretended, and human baseline approval remains separate"
-      ]
-    },
-    {
-      "id": 5,
-      "kind": "dialogue",
-      "prompt": "In this main session, analyze an editable admin-list HTML reference and a business brief for a Vue 3 plus antDv-next review prototype. Keep this read-only. The host has no dedicated prototype role and no parent-child JSON transport. State the useful analysis and next steps without installing tools or running creation-stage checks.",
-      "expected_output": "A complete read-only main-session analysis with useful coverage and a truthful acceptance plan",
-      "expectations": [
-        "The main session directly applies the shared prototype preflight without requiring or simulating a dedicated Agent",
-        "The analysis delivers requirement and bound-capability coverage, template and framework selection, acceptance planning and concrete gaps",
-        "No initialization, installation, persisted design, project state update, build, browser run or creation-stage UIUX adoption gate occurs",
-        "No parent-child JSON, PROTOTYPE_RESULT_V1 marker or validate-result execution is made a prerequisite for the natural-language deliverable",
-        "Unavailable capabilities and fixed-parameter conflicts remain visible; analysis completion is distinct from implemented coverage and human approval"
-      ]
-    },
-    {
-      "id": 6,
-      "prompt": "Create a separate native Vue 3 plus antDv-next device-list review candidate from the editable admin-list reference. Use template mode explicitly. I authorize four in-page themes named 明亮、暗黑、清新、科技, defaulting to 明亮. Enable and fix the theme requirement in my HTML working copy, preserve business state while switching, and preserve all existing prototypes and manual edits. Do not ask again whether themes are wanted. Keep fixed required capabilities and verify the current build when the specified browser tools are available.",
-      "expected_output": "A main-session native review candidate with explicitly fixed four-theme inputs, preserved state and truthful current-build evidence or gaps",
-      "files": [
-        "i3-ui-prototype"
-      ],
-      "expectations": [
-        "The main session uses the prototype workflow and supporting frontend and version-matched official framework knowledge for actual implementation, without requiring a dedicated persona",
-        "The authorized theme capability is bound in the user HTML working copy and fixed with its requirement source rather than left disabled or sent for repeat confirmation",
-        "The generated framework implements the four exact theme names, defaults to 明亮, and verifies that business state survives theme changes",
-        "Existing or unrequested-theme instances, user edits and approved baselines remain unchanged",
-        "Template mode does not claim UIUX use; fixed required capabilities and applicable acceptance scenarios are retained",
-        "Build, actual browser Skill and DevTools observations, supplemental regression and current delivery evidence are distinguished; unavailable checks remain incomplete",
-        "The final natural-language delivery identifies the candidate, current evidence and gaps, and leaves human review pending without requiring a parent-child result envelope"
-      ]
-    },
-    {
-      "id": 7,
-      "kind": "dialogue",
-      "prompt": "Compare how you would handle three requests: analyze a prototype plan without writing; revise the prototype drawer interaction for review; implement the approved drawer as a reusable production component. Explain which skills and evidence apply to each, without executing the requests.",
-      "expected_output": "Operation-specific support and a clear boundary between prototype work and production frontend engineering",
-      "expectations": [
-        "Read-only prototype analysis uses only material needed for coverage and selection and plans later verification without executing it",
-        "Prototype interaction revision stays within prototype workflow ownership and uses frontend, official framework and browser capabilities when the revision requires them",
-        "Professional UIUX lookup, adoption and effect verification remain distinct and apply to professional design work, with no silent downgrade to template mode",
-        "Approved production component implementation follows the frontend implementation method and the existing engineering workflow rather than creating a new review prototype",
-        "No claim of completed runtime checks or human baseline approval is made"
-      ]
-    }
-  ]
-}