novelws 3.0.0 → 4.0.0

package/templates/commands/relations.md

@@ -2,10 +2,10 @@
 name: relations
 description: 管理和追踪角色关系变化
 argument-hint: [update | show | history | check] [角色] [关系] [目标角色] [--volume vol-XX | --character-focus 角色名]
-allowed-tools: Read(//spec/tracking/relationships.json), Read(//spec/tracking/relationships.json), Write(//spec/tracking/relationships.json), Write(//spec/tracking/relationships.json), Bash(find:*), Bash(*)
+allowed-tools: Read(//tracking/relationships.json), Read(//tracking/relationships.json), Write(//tracking/relationships.json), Write(//tracking/relationships.json), Bash(find:*), Bash(*)
 scripts:
-  sh: .specify/scripts/bash/manage-relations.sh
-  ps: .specify/scripts/powershell/manage-relations.ps1
+  sh: resources/scripts/bash/manage-relations.sh
+  ps: resources/scripts/powershell/manage-relations.ps1
 ---
 
 # 角色关系管理
@@ -54,8 +54,8 @@ const relationships = await mcp.call('novelws-mcp/query_relationships', {
 ### Layer 2: 分片 JSON（次优）
 
 ```bash
-# 当 spec/tracking/volumes/ 存在时
-relationships_data=$(cat spec/tracking/volumes/vol-03/relationships.json)
+# 当 tracking/volumes/ 存在时
+relationships_data=$(cat tracking/volumes/vol-03/relationships.json)
 ```
 
 **适用场景**：
@@ -67,7 +67,7 @@ relationships_data=$(cat spec/tracking/volumes/vol-03/relationships.json)
 
 ```bash
 # 传统模式，加载完整文件
-relationships_data=$(cat spec/tracking/relationships.json)
+relationships_data=$(cat tracking/relationships.json)
 ```
 
 **向下兼容**：小型项目（< 300 章）继续使用单文件模式
@@ -120,7 +120,7 @@ relationships_data=$(cat spec/tracking/relationships.json)
 
 ## 数据存储
 
-关系数据存储在 `spec/tracking/relationships.json`：
+关系数据存储在 `tracking/relationships.json`：
 ```json
 {
   "characters": {
@@ -163,7 +163,7 @@ relationships_data=$(cat spec/tracking/relationships.json)
 
 关系数据更新遵循以下协议：
 
-### 分片模式（spec/tracking/volumes/ 存在）
+### 分片模式（tracking/volumes/ 存在）
 
 1. **确定目标卷**：
    - 如果命令指定 `--volume vol-XX`，写入对应卷
@@ -177,13 +177,13 @@ relationships_data=$(cat spec/tracking/relationships.json)
 2. **写入分片文件**：
    ```bash
    # 写入指定卷的 relationships.json
-   Write(spec/tracking/volumes/${target_volume}/relationships.json)
+   Write(tracking/volumes/${target_volume}/relationships.json)
    ```
 
 3. **更新全局摘要**（如果关系变化影响角色摘要）：
    ```bash
    # 更新 characters-summary.json 的关系统计
-   Write(spec/tracking/summary/characters-summary.json)
+   Write(tracking/summary/characters-summary.json)
    ```
 
 4. **触发 MCP 同步**（如果启用）：
@@ -198,7 +198,7 @@ relationships_data=$(cat spec/tracking/relationships.json)
 
 直接写入完整 `relationships.json`：
 ```bash
-Write(spec/tracking/relationships.json)
+Write(tracking/relationships.json)
 ```
 
 ### 写入示例
@@ -211,8 +211,8 @@ const targetVolume = `vol-${String(volumeNum).padStart(2, '0')}`;
 
 // 读取目标卷的 relationships 分片
 const relPath = isSharded
-  ? `spec/tracking/volumes/${targetVolume}/relationships.json`
-  : 'spec/tracking/relationships.json';
+  ? `tracking/volumes/${targetVolume}/relationships.json`
+  : 'tracking/relationships.json';
 
 const relationships = JSON.parse(fs.readFileSync(relPath));
 

package/templates/commands/revise.md

@@ -1,6 +1,7 @@
 ---
 description: 系统性修改润色：四层修改流程（结构→节奏→一致性→文字），支持分层修改、指定范围、快速模式
 argument-hint: [--layer=structure|pacing|consistency|polish] [--chapters=1-10] [--chapter=5] [--volume vol-XX] [--quick]
+recommended-model: claude-sonnet-4-5-20250929 # 修改任务；重大改写可用 opus
 allowed-tools: Bash(find:*), Bash(wc:*), Bash(grep:*), Read(//**), Write(//stories/**), Bash(*)
 ---
 
@@ -69,14 +70,14 @@ const plotData = await mcp.call('novelws-mcp/query_plot', {
 #### Layer 2: 分片 JSON（次优）
 
 ```bash
-# 当 spec/tracking/volumes/ 存在时
+# 当 tracking/volumes/ 存在时
 if [[ -n "$VOLUME_FILTER" ]]; then
   # 仅加载指定卷的数据
-  cat spec/tracking/volumes/${VOLUME_FILTER}/character-state.json
-  cat spec/tracking/volumes/${VOLUME_FILTER}/relationships.json
+  cat tracking/volumes/${VOLUME_FILTER}/character-state.json
+  cat tracking/volumes/${VOLUME_FILTER}/relationships.json
 else
   # 加载所有卷的数据
-  for vol in spec/tracking/volumes/vol-*/; do
+  for vol in tracking/volumes/vol-*/; do
     cat "$vol/character-state.json"
     cat "$vol/relationships.json"
   done
@@ -92,8 +93,8 @@ fi
 
 ```bash
 # 传统模式，加载完整文件
-cat spec/tracking/character-state.json
-cat spec/tracking/relationships.json
+cat tracking/character-state.json
+cat tracking/relationships.json
 ```
 
 **向下兼容**：小型项目（< 300 章）继续使用单文件模式
@@ -109,24 +110,24 @@ cat spec/tracking/relationships.json
 - 创作计划：`stories/*/creative-plan.md`
 - **角色状态**（三层回退）：
   - Layer 3: `mcp.call('novelws-mcp/query_characters', {volume})`
-  - Layer 2: `spec/tracking/volumes/${volume}/character-state.json`
-  - Layer 1: `spec/tracking/character-state.json`
+  - Layer 2: `tracking/volumes/${volume}/character-state.json`
+  - Layer 1: `tracking/character-state.json`
 - **关系网络**（三层回退）：
   - Layer 3: `mcp.call('novelws-mcp/query_relationships', {volume})`
-  - Layer 2: `spec/tracking/volumes/${volume}/relationships.json`
-  - Layer 1: `spec/tracking/relationships.json`
+  - Layer 2: `tracking/volumes/${volume}/relationships.json`
+  - Layer 1: `tracking/relationships.json`
 
 **如有则加载**：
 - `/analyze` 报告：`stories/*/analysis-report.md`
-- 风格参考：`.specify/memory/style-reference.md`
-- 个人风格：`.specify/memory/personal-voice.md`
+- 风格参考：`resources/memory/style-reference.md`
+- 个人风格：`resources/memory/personal-voice.md`
 - **情节追踪**（三层回退）：
   - Layer 3: `mcp.call('novelws-mcp/query_plot', {volume})`
-  - Layer 2: `spec/tracking/volumes/${volume}/plot-tracker.json`
-  - Layer 1: `spec/tracking/plot-tracker.json`
+  - Layer 2: `tracking/volumes/${volume}/plot-tracker.json`
+  - Layer 1: `tracking/plot-tracker.json`
 
 **加载 Skill**：
-- `.specify/templates/skills/writing-techniques/revision-polish/SKILL.md`（核心修改流程）
+- `.claude/skills/writing-techniques/revision-polish/SKILL.md`（核心修改流程）
 
 ### 4. 初稿评估
 

package/templates/commands/search.md

@@ -2,6 +2,7 @@
 name: search
 description: 全文搜索章节内容和 tracking 数据（支持中文分词、正则表达式、范围过滤）
 argument-hint: <关键词> [--type=content|tracking|all] [--volume vol-XX] [--fuzzy] [--regex]
+recommended-model: claude-haiku-4-5-20251001 # 简单查询，速度优先
 allowed-tools: Read, Grep, Bash
 ---
 
@@ -45,15 +46,15 @@ const results = await mcp.call('novelws-mcp/search_content', {
 ### Layer 2: 分片 Grep（次优）
 
 ```bash
-# 当 spec/tracking/volumes/ 存在时
+# 当 tracking/volumes/ 存在时
 if [[ -n "$VOLUME_FILTER" ]]; then
   # 仅搜索指定卷
   grep -rn "$QUERY" "stories/*/content/" | grep "ch-$(volume_to_range $VOLUME_FILTER)"
-  grep -rn "$QUERY" "spec/tracking/volumes/$VOLUME_FILTER/"
+  grep -rn "$QUERY" "tracking/volumes/$VOLUME_FILTER/"
 else
   # 搜索所有内容
   grep -rn "$QUERY" "stories/*/content/"
-  grep -rn "$QUERY" "spec/tracking/volumes/"
+  grep -rn "$QUERY" "tracking/volumes/"
 fi
 ```
 
@@ -67,7 +68,7 @@ fi
 ```bash
 # 传统模式，搜索所有文件
 grep -rn "$QUERY" "stories/*/content/"
-grep -rn "$QUERY" "spec/tracking/"
+grep -rn "$QUERY" "tracking/"
 ```
 
 **向下兼容**：小型项目（< 300 章）使用标准 Grep
@@ -101,7 +102,7 @@ const isRegex = args.regex || false;   // 正则表达式
 ### Step 2: 检测数据模式
 
 ```javascript
-const isSharded = fs.existsSync('spec/tracking/volumes/');
+const isSharded = fs.existsSync('tracking/volumes/');
 const hasMCP = fs.existsSync('mcp-servers.json');
 
 // 选择搜索策略
@@ -178,11 +179,11 @@ if [[ -n "$VOLUME_FILTER" ]]; then
   done
 
   # 搜索 tracking 数据
-  grep -rn "$QUERY" "spec/tracking/volumes/$VOLUME_FILTER/"
+  grep -rn "$QUERY" "tracking/volumes/$VOLUME_FILTER/"
 else
   # 搜索所有内容
   grep -rn "$QUERY" "stories/*/content/"
-  grep -rn "$QUERY" "spec/tracking/volumes/"
+  grep -rn "$QUERY" "tracking/volumes/"
 fi
 ```
 
@@ -195,7 +196,7 @@ QUERY="$1"
 grep -rn "$QUERY" "stories/*/content/"
 
 # 搜索 tracking 数据
-grep -rn "$QUERY" "spec/tracking/" --exclude-dir=volumes
+grep -rn "$QUERY" "tracking/" --exclude-dir=volumes
 ```
 
 ### Step 4: 处理搜索结果
@@ -261,7 +262,7 @@ const resultsWithContext = sortedResults.map(result => ({
 
 ### 角色数据
 
-**文件**：spec/tracking/volumes/vol-03/character-state.json
+**文件**：tracking/volumes/vol-03/character-state.json
 
 **匹配项**：
 ```json
@@ -276,7 +277,7 @@ const resultsWithContext = sortedResults.map(result => ({
 
 ### 情节数据
 
-**文件**：spec/tracking/volumes/vol-03/plot-tracker.json
+**文件**：tracking/volumes/vol-03/plot-tracker.json
 
 **匹配项**：
 ```json

package/templates/commands/specify.md

@@ -1,10 +1,10 @@
 ---
 description: 定义故事规格，明确要创造什么样的作品
 argument-hint: [故事描述] [--world]
-allowed-tools: Read(//stories/**/specification.md), Read(//stories/**/specification.md), Write(//stories/**/specification.md), Write(//stories/**/specification.md), Read(//.specify/memory/constitution.md), Read(//.specify/memory/constitution.md), Bash(find:*), Bash(*)
+allowed-tools: Read(//stories/**/specification.md), Read(//stories/**/specification.md), Write(//stories/**/specification.md), Write(//stories/**/specification.md), Read(//resources/memory/constitution.md), Read(//resources/memory/constitution.md), Bash(find:*), Bash(*)
 scripts:
-  sh: .specify/scripts/bash/specify-story.sh --json
-  ps: .specify/scripts/powershell/specify-story.ps1 -Json
+  sh: resources/scripts/bash/specify-story.sh --json
+  ps: resources/scripts/powershell/specify-story.ps1 -Json
 ---
 
 用户输入：$ARGUMENTS
@@ -99,7 +99,7 @@ else:
 
 ### 2. 检查宪法合规性
 
-如果存在 `.specify/memory/constitution.md`：
+如果存在 `resources/memory/constitution.md`：
 - 加载宪法原则
 - 确保规格符合宪法价值观
 - 在规格中引用相关原则

package/templates/commands/tasks.md

@@ -2,8 +2,8 @@
 description: 将创作计划分解为可执行的任务清单
 allowed-tools: Read(//stories/**/creative-plan.md), Read(//stories/**/creative-plan.md), Read(//stories/**/specification.md), Read(//stories/**/specification.md), Write(//stories/**/tasks.md), Write(//stories/**/tasks.md), Bash(find:*), Bash(*)
 scripts:
-  sh: .specify/scripts/bash/tasks-story.sh
-  ps: .specify/scripts/powershell/generate-tasks.ps1
+  sh: resources/scripts/bash/tasks-story.sh
+  ps: resources/scripts/powershell/generate-tasks.ps1
 ---
 
 基于创作计划生成具体的、可执行的任务列表。
@@ -63,16 +63,16 @@ const characters = await mcp.call('novelws-mcp/query_characters', {
 ### Layer 2: 分片 JSON（次优）
 
 ```bash
-# 当 spec/tracking/volumes/ 存在时
+# 当 tracking/volumes/ 存在时
 # 读取摘要数据
-plot_summary=$(cat spec/tracking/summary/plot-summary.json)
+plot_summary=$(cat tracking/summary/plot-summary.json)
 
 # 提取未解决伏笔和紧急伏笔
 unresolved_plots=$(echo "$plot_summary" | jq '.unresolved')
 urgent_plots=$(echo "$plot_summary" | jq '.urgent')
 
 # 读取角色摘要
-character_summary=$(cat spec/tracking/summary/characters-summary.json)
+character_summary=$(cat tracking/summary/characters-summary.json)
 absent_characters=$(echo "$character_summary" | jq '.absent')
 ```
 
@@ -85,8 +85,8 @@ absent_characters=$(echo "$character_summary" | jq '.absent')
 
 ```bash
 # 传统模式，加载完整文件并手动过滤
-plot_tracker=$(cat spec/tracking/plot-tracker.json)
-character_state=$(cat spec/tracking/character-state.json)
+plot_tracker=$(cat tracking/plot-tracker.json)
+character_state=$(cat tracking/character-state.json)
 
 # 手动计算紧急度、缺席章节等
 ```
@@ -313,8 +313,8 @@ character_state=$(cat spec/tracking/character-state.json)
 4. 自动设置依赖关系（顺序依赖）
 5. 按三层 Fallback 读取 tracking 数据生成修复/维护任务：
    - **MCP 查询（优先）**：`query_plot --status=overdue` 获取延迟伏笔，`stats_volume` 获取进度
-   - **分片 JSON（次优）**：读取 `spec/tracking/summary/plot-summary.json` 获取未解决伏笔
-   - **单文件 JSON（兜底）**：读取 `spec/tracking/plot-tracker.json`
+   - **分片 JSON（次优）**：读取 `tracking/summary/plot-summary.json` 获取未解决伏笔
+   - **单文件 JSON（兜底）**：读取 `tracking/plot-tracker.json`
 6. 从 `/analyze` 反馈中提取修复任务
 7. 从 `/track --check` 中提取维护任务
 

package/templates/commands/timeline.md

@@ -2,10 +2,10 @@
 name: timeline
 description: 管理和验证故事时间线
 argument-hint: [add | check | show | sync] [--volume vol-XX | --recent N]
-allowed-tools: Read(//spec/tracking/timeline.json), Read(//spec/tracking/timeline.json), Write(//spec/tracking/timeline.json), Write(//spec/tracking/timeline.json), Read(//stories/**/content/**), Read(//stories/**/content/**), Bash(find:*), Bash(*)
+allowed-tools: Read(//tracking/timeline.json), Read(//tracking/timeline.json), Write(//tracking/timeline.json), Write(//tracking/timeline.json), Read(//stories/**/content/**), Read(//stories/**/content/**), Bash(find:*), Bash(*)
 scripts:
-  sh: .specify/scripts/bash/check-timeline.sh
-  ps: .specify/scripts/powershell/check-timeline.ps1
+  sh: resources/scripts/bash/check-timeline.sh
+  ps: resources/scripts/powershell/check-timeline.ps1
 ---
 
 # 时间线管理
@@ -53,8 +53,8 @@ const events = await mcp.call('novelws-mcp/query_timeline', {
 ### Layer 2: 分片 JSON（次优）
 
 ```bash
-# 当 spec/tracking/volumes/ 存在时
-timeline_data=$(cat spec/tracking/volumes/vol-03/timeline.json)
+# 当 tracking/volumes/ 存在时
+timeline_data=$(cat tracking/volumes/vol-03/timeline.json)
 ```
 
 **适用场景**：
@@ -66,7 +66,7 @@ timeline_data=$(cat spec/tracking/volumes/vol-03/timeline.json)
 
 ```bash
 # 传统模式，加载完整文件
-timeline_data=$(cat spec/tracking/timeline.json)
+timeline_data=$(cat tracking/timeline.json)
 ```
 
 **向下兼容**：小型项目（< 300 章）继续使用单文件模式
@@ -109,7 +109,7 @@ timeline_data=$(cat spec/tracking/timeline.json)
 
 ## 时间线数据
 
-时间线信息存储在 `spec/tracking/timeline.json` 中：
+时间线信息存储在 `tracking/timeline.json` 中：
 - 故事内时间（年/月/日）
 - 章节对应关系
 - 重要事件标记
@@ -136,7 +136,7 @@ timeline_data=$(cat spec/tracking/timeline.json)
 
 时间线数据更新遵循以下协议：
 
-### 分片模式（spec/tracking/volumes/ 存在）
+### 分片模式（tracking/volumes/ 存在）
 
 1. **确定目标卷**：
    - 如果命令指定 `--volume vol-XX`，写入对应卷
@@ -150,13 +150,13 @@ timeline_data=$(cat spec/tracking/timeline.json)
 2. **写入分片文件**：
    ```bash
    # 写入指定卷的 timeline.json
-   Write(spec/tracking/volumes/${target_volume}/timeline.json)
+   Write(tracking/volumes/${target_volume}/timeline.json)
    ```
 
 3. **更新全局摘要**（如果事件跨度影响摘要统计）：
    ```bash
    # 更新 timeline-summary.json 的统计信息
-   Write(spec/tracking/summary/timeline-summary.json)
+   Write(tracking/summary/timeline-summary.json)
    ```
 
 4. **触发 MCP 同步**（如果启用）：
@@ -171,7 +171,7 @@ timeline_data=$(cat spec/tracking/timeline.json)
 
 直接写入完整 `timeline.json`：
 ```bash
-Write(spec/tracking/timeline.json)
+Write(tracking/timeline.json)
 ```
 
 ### 写入示例
@@ -184,8 +184,8 @@ const targetVolume = `vol-${String(volumeNum).padStart(2, '0')}`;
 
 // 读取目标卷的 timeline 分片
 const timelinePath = isSharded
-  ? `spec/tracking/volumes/${targetVolume}/timeline.json`
-  : 'spec/tracking/timeline.json';
+  ? `tracking/volumes/${targetVolume}/timeline.json`
+  : 'tracking/timeline.json';
 
 const timeline = JSON.parse(fs.readFileSync(timelinePath));
 

package/templates/commands/track-init.md

@@ -1,10 +1,10 @@
 ---
 name: track-init
 description: 初始化追踪系统，基于故事大纲设置追踪数据
-allowed-tools: Read(//stories/**/specification.md), Read(//stories/**/specification.md), Read(//stories/**/creative-plan.md), Read(//stories/**/creative-plan.md), Write(//spec/tracking/**), Write(//spec/tracking/**), Bash(find:*), Bash(grep:*), Bash(wc:*), Bash(*)
+allowed-tools: Read(//stories/**/specification.md), Read(//stories/**/specification.md), Read(//stories/**/creative-plan.md), Read(//stories/**/creative-plan.md), Write(//tracking/**), Write(//tracking/**), Bash(find:*), Bash(grep:*), Bash(wc:*), Bash(*)
 scripts:
-  sh: .specify/scripts/bash/init-tracking.sh
-  ps: .specify/scripts/powershell/init-tracking.ps1
+  sh: resources/scripts/bash/init-tracking.sh
+  ps: resources/scripts/powershell/init-tracking.ps1
 ---
 
 # 初始化追踪系统
@@ -38,13 +38,13 @@ resource-loading:
 1. **读取基础数据**
    - 读取 `stories/*/specification.md` 获取故事设定
    - 读取 `stories/*/creative-plan.md` 获取章节规划
-   - 读取 `.specify/config.json` 获取写作方法
+   - 读取 `resources/config/config.json` 获取写作方法
 
 2. **初始化追踪文件**
 
    **重要**：优先从 `specification.md` 第五章读取线索管理规格，填充到追踪文件。
 
-   创建或更新 `spec/tracking/plot-tracker.json`：
+   创建或更新 `tracking/plot-tracker.json`：
    - 从 `specification.md 5.1节` 读取所有线索定义
    - 从 `specification.md 5.3节` 读取所有交汇点
    - 从 `specification.md 5.4节` 读取所有伏笔
@@ -111,15 +111,15 @@ resource-loading:
    }
    ```
 
-   创建或更新 `spec/tracking/timeline.json`：
+   创建或更新 `tracking/timeline.json`：
    - 根据章节规划设置时间节点
    - 标记重要时间事件
 
-   创建或更新 `spec/tracking/relationships.json`：
+   创建或更新 `tracking/relationships.json`：
    - 从角色设定提取初始关系
    - 设置派系分组
 
-   创建或更新 `spec/tracking/character-state.json`：
+   创建或更新 `tracking/character-state.json`：
    - 初始化角色状态
    - 设置起始位置