@hupan56/wlkj 2.7.11 → 3.0.0

package/package.json

@@ -1,29 +1,29 @@
-{
-  "name": "@hupan56/wlkj",
-  "version": "2.7.11",
-  "description": "AI Product R&D Workflow - PRD/Prototype/Search/Task/Report",
-  "bin": {
-    "wlkj": "bin/cli.js"
-  },
-  "files": [
-    "bin/",
-    "templates/",
-    "MAC-VERIFY.md"
-  ],
-  "keywords": [
-    "workflow",
-    "ai",
-    "prd",
-    "pipeline",
-    "qoder",
-    "product",
-    "team"
-  ],
-  "license": "MIT",
-  "publishConfig": {
-    "access": "public"
-  },
-  "engines": {
-    "node": ">=16"
-  }
-}
+{
+  "name": "@hupan56/wlkj",
+  "version": "3.0.0",
+  "description": "AI Product R&D Workflow - PRD/Prototype/Search/Task/Report",
+  "bin": {
+    "wlkj": "bin/cli.js"
+  },
+  "files": [
+    "bin/",
+    "templates/",
+    "MAC-VERIFY.md"
+  ],
+  "keywords": [
+    "workflow",
+    "ai",
+    "prd",
+    "pipeline",
+    "qoder",
+    "product",
+    "team"
+  ],
+  "license": "MIT",
+  "publishConfig": {
+    "access": "public"
+  },
+  "engines": {
+    "node": ">=16"
+  }
+}

package/templates/.qoder/.runtime/ctx-cache-5660152f1d6dd819.md

@@ -0,0 +1,23 @@
+# Context Pack: 考勤 (platform: web)
+搜索词扩展: attendance clock time/clock timeclock
+
+## 1. 相关代码文件 (0 个, 取前 12)
+- (无命中 — 换业务词重试, 或该功能为全新)
+
+## 2. 同类页面示例 (风格锚点, 读 1-2 个)
+
+## 3. 相关字段 (真实字段名->中文标题)
+- (无直接相关字段)
+
+## 4. 相关历史 PRD (data/docs/prd/)
+- (无相关 PRD)
+
+## 5. 相关 API (取前 8 / 共 0)
+- (无相关 API)
+
+## 6. 风格真源速查
+- Web Token: data/index/ref-vben-style.json (primary=hsl(214.58 86.34% 59.8%))
+- 看板/大屏: data/index/ref-chart-style.json (大屏深蓝风 monitor / 浅色卡片风 workbench)
+- 图标 (禁 emoji): data/index/ref-icon.json (Web=antd 内联SVG, APP=vant 字体)
+- 模板: .qoder/templates/prototype-web.html / prototype-app.html
+- PRD 模板: .qoder/templates/prd-full-template.md / prd-quick-template.md

package/templates/.qoder/.runtime/ctx-cache-afdce0dac06b25b0.md

@@ -0,0 +1,23 @@
+# Context Pack: 保险 (platform: web)
+搜索词扩展: ins insurance vehlife/insurance
+
+## 1. 相关代码文件 (0 个, 取前 12)
+- (无命中 — 换业务词重试, 或该功能为全新)
+
+## 2. 同类页面示例 (风格锚点, 读 1-2 个)
+
+## 3. 相关字段 (真实字段名->中文标题)
+- (无直接相关字段)
+
+## 4. 相关历史 PRD (data/docs/prd/)
+- (无相关 PRD)
+
+## 5. 相关 API (取前 8 / 共 0)
+- (无相关 API)
+
+## 6. 风格真源速查
+- Web Token: data/index/ref-vben-style.json (primary=hsl(214.58 86.34% 59.8%))
+- 看板/大屏: data/index/ref-chart-style.json (大屏深蓝风 monitor / 浅色卡片风 workbench)
+- 图标 (禁 emoji): data/index/ref-icon.json (Web=antd 内联SVG, APP=vant 字体)
+- 模板: .qoder/templates/prototype-web.html / prototype-app.html
+- PRD 模板: .qoder/templates/prd-full-template.md / prd-quick-template.md

package/templates/.qoder/.runtime/search-cache-eae7644e7b122f35.txt

@@ -0,0 +1 @@
+code-keyword.json not found. Run: python git_sync.py

package/templates/.qoder/learning/eval-history.jsonl

@@ -0,0 +1,28 @@
+{"a1_reality": 0.0, "a2_style": null, "a3_template": 0.0, "passed": false, "prd": "req-good.md", "score_pct": 0.0, "time": "2026-06-23T15:06:47.512881"}
+{"a1_reality": 1.0, "a2_style": null, "a3_template": 0.0, "passed": false, "prd": "req-poor.md", "score_pct": 57.1, "time": "2026-06-23T15:06:48.151944"}
+{"a1_reality": 0.0, "a2_style": null, "a3_template": 0.0, "passed": false, "prd": "req-good.md", "score_pct": 0.0, "time": "2026-06-23T15:07:06.380363"}
+{"a1_reality": 1.0, "a2_style": null, "a3_template": 0.0, "passed": false, "prd": "req-poor.md", "score_pct": 57.1, "time": "2026-06-23T15:07:20.277445"}
+{"a1_reality": 0.5, "a2_style": null, "a3_template": 0.0, "passed": false, "prd": "req-good.md", "score_pct": 28.6, "time": "2026-06-23T15:08:26.677411"}
+{"a1_reality": 1.0, "a2_style": null, "a3_template": 0.0, "passed": false, "prd": "req-poor.md", "score_pct": 57.1, "time": "2026-06-23T15:08:27.209288"}
+{"a1_reality": 0.5, "a2_style": null, "a3_template": 0.0, "passed": false, "prd": "req-good.md", "score_pct": 28.6, "time": "2026-06-23T15:08:55.802944"}
+{"a1_reality": 1.0, "a2_style": null, "a3_template": 0.0, "passed": false, "prd": "req-poor.md", "score_pct": 57.1, "time": "2026-06-23T15:08:56.234057"}
+{"a1_reality": 0.5, "a2_style": null, "a3_template": 0.0, "passed": false, "prd": "req-good.md", "score_pct": 28.6, "time": "2026-06-23T15:11:32.306461"}
+{"a1_reality": 1.0, "a2_style": null, "a3_template": 0.0, "passed": false, "prd": "req-poor.md", "score_pct": 57.1, "time": "2026-06-23T15:11:32.761987"}
+{"a1_reality": 0.5, "a2_style": null, "a3_template": 0.0, "passed": false, "prd": "req-good.md", "score_pct": 28.6, "time": "2026-06-23T15:19:28.644771"}
+{"a1_reality": 1.0, "a2_style": null, "a3_template": 0.0, "passed": false, "prd": "req-poor.md", "score_pct": 57.1, "time": "2026-06-23T15:19:29.240469"}
+{"a1_reality": 0.5, "a2_style": null, "a3_template": 0.0, "passed": false, "prd": "req-good.md", "score_pct": 28.6, "time": "2026-06-23T15:26:51.416116"}
+{"a1_reality": 1.0, "a2_style": null, "a3_template": 0.0, "passed": false, "prd": "req-poor.md", "score_pct": 57.1, "time": "2026-06-23T15:26:52.129295"}
+{"a1_reality": 0.5, "a2_style": null, "a3_template": 0.0, "passed": false, "prd": "req-good.md", "score_pct": 28.6, "time": "2026-06-23T15:32:16.110663"}
+{"a1_reality": 1.0, "a2_style": null, "a3_template": 0.0, "passed": false, "prd": "req-poor.md", "score_pct": 57.1, "time": "2026-06-23T15:32:16.669572"}
+{"a1_reality": 0.5, "a2_style": null, "a3_template": 0.0, "passed": false, "prd": "req-good.md", "score_pct": 28.6, "time": "2026-06-23T15:35:31.618489"}
+{"a1_reality": 1.0, "a2_style": null, "a3_template": 0.0, "passed": false, "prd": "req-poor.md", "score_pct": 57.1, "time": "2026-06-23T15:35:32.085857"}
+{"a1_reality": 0.5, "a2_style": null, "a3_template": 0.0, "passed": false, "prd": "req-good.md", "score_pct": 28.6, "time": "2026-06-23T15:44:50.176279"}
+{"a1_reality": 1.0, "a2_style": null, "a3_template": 0.0, "passed": false, "prd": "req-poor.md", "score_pct": 57.1, "time": "2026-06-23T15:44:50.849511"}
+{"a1_reality": 0.5, "a2_style": null, "a3_template": 0.0, "passed": false, "prd": "req-good.md", "score_pct": 28.6, "time": "2026-06-23T15:49:54.375114"}
+{"a1_reality": 1.0, "a2_style": null, "a3_template": 0.0, "passed": false, "prd": "req-poor.md", "score_pct": 57.1, "time": "2026-06-23T15:49:54.852542"}
+{"a1_reality": 0.5, "a2_style": null, "a3_template": 0.0, "passed": false, "prd": "req-good.md", "score_pct": 28.6, "time": "2026-06-23T16:09:00.572162"}
+{"a1_reality": 1.0, "a2_style": null, "a3_template": 0.0, "passed": false, "prd": "req-poor.md", "score_pct": 57.1, "time": "2026-06-23T16:09:01.184157"}
+{"a1_reality": 0.5, "a2_style": null, "a3_template": 0.0, "passed": false, "prd": "req-good.md", "score_pct": 28.6, "time": "2026-06-23T16:28:54.661703"}
+{"a1_reality": 1.0, "a2_style": null, "a3_template": 0.0, "passed": false, "prd": "req-poor.md", "score_pct": 57.1, "time": "2026-06-23T16:28:55.227433"}
+{"a1_reality": 0.5, "a2_style": null, "a3_template": 0.0, "passed": false, "prd": "req-good.md", "score_pct": 28.6, "time": "2026-06-23T16:46:57.600191"}
+{"a1_reality": 1.0, "a2_style": null, "a3_template": 0.0, "passed": false, "prd": "req-poor.md", "score_pct": 57.1, "time": "2026-06-23T16:46:58.135094"}

package/templates/data/index/wiki-index.json

@@ -0,0 +1,8 @@
+{
+  "index": {},
+  "stats": {
+    "projects": 0,
+    "pages": 0,
+    "keywords": 0
+  }
+}

package/templates/qoder/agents/insight-planning.md

@@ -23,7 +23,7 @@ You are the insight-planning sub-agent. Do NOT spawn another insight-planning ag
 
 ## Phase 1: 取现状（输入，不是本 agent 的活）
 - 从 dispatch prompt 或 context_pack 结果读现状摘要
-- 若现状信息不足 → 回报 caller"现状锚定不全，建议先跑 context_pack 或先做 discover"
+- 若现状信息不足 → 回报 caller"现状锚定不全，建议先跑 context_pack 或先做 clarify 对焦"
 
 ## Phase 2: 理想态（如果有 research 报告）
 1. 读最近一份 `workspace/members/{dev}/drafts/insight-research-*.md`

package/templates/qoder/agents/insight-research.md

@@ -1,12 +1,13 @@
-# Insight Research Agent（外部调研 · 多视角脑暴）
+# Insight Research Agent（外部调研 · 多视角 + 方法论框架）
 
-> 本 agent 服务于 `/wl-insight research` 类型。完整规则见 `.qoder/commands/optional/wl-insight.md`。
-> 来源：整体从 prd-research.md 搬迁而来（脑暴 + web 搜索能力归 insight）。
+> 本 agent 服务于 `/wl-insight` 的 research 段。完整规则见 `.qoder/commands/optional/wl-insight.md`。
+> ⚠️ **research 段只在用户明确同意"要调研"后才触发**(见 skill 第 2 段)。
+>    绝不擅自调研——那是上一版的毛病。
 
 ## Required: Load Context First (Pull-Based)
 This platform does NOT auto-inject context. Before doing anything else:
-1. Read the dispatch prompt to get: assigned perspective, topic,
-   current state summary (from context_pack), developer name
+1. Read the dispatch prompt to get: assigned perspective, **assigned methodology**,
+   topic, current state summary (from context_pack), developer name
 2. If perspective is missing from dispatch prompt, STOP and report back:
    "perspective not specified - caller must assign one of: 竞品对标/行业趋势/技术方案/用户场景"
 
@@ -18,16 +19,25 @@ Do NOT spawn another insight-research agent.
 **全部输出必须简体中文。**（技术专有名词可保留英文）
 
 ## Your Role
-你是 N 个调研 agent 中的一个，负责**一个特定视角**的外部调研。
-你的视角在 dispatch prompt 里指定。你的产出是**中间产物**——
-caller（/wl-insight research）会把所有 agent 的报告合并成一份调研报告，
-然后删除中间产物。
+你是 N 个调研 agent 中的一个，负责**一个特定视角**的外部调研，
+并用**指定的方法论框架**组织分析（让结论结构化、可对比，不是散点罗列）。
+你的视角和方法论在 dispatch prompt 里指定。你的产出是**中间产物**——
+caller 会把所有 agent 的报告合并成一份调研报告，然后删除中间产物。
+
+## 方法论框架（按视角，dispatch 会指定用哪个）
+
+| 视角 | 方法论 | 输出形态 |
+|------|--------|---------|
+| 竞品对标 | **竞品矩阵** | 功能对比表(我们 vs 竞品A vs B)+ 差异化机会 |
+| 技术方案 | **技术选型对比** | 方案A vs B(成本/风险/成熟度/社区)矩阵 + 推荐 |
+| 用户场景 | **JTBD + Kano** | "用户在完成什么 job" + 需求分类(基础/期望/兴奋) |
+| 行业趋势 | **趋势时间线** | 过去→现在→未来节点 + 机会窗口 |
 
 ## Workflow
-1. 从 dispatch prompt 读你的视角：竞品对标 / 行业趋势 / 技术方案 / 用户场景 之一
+1. 从 dispatch prompt 读你的**视角 + 方法论**
 2. 读 dispatch prompt 里的**本项目现状摘要**（让调研有针对性，不泛泛而谈）
 3. 用 WebSearch 搜 3-5 个权威源（官方文档 / 行业报告 / 技术博客 / 公开案例）
-4. 从你的视角分析，重点找：**外部做得好而我们没做的**（Gap 信号）
+4. **用指定的方法论框架**组织分析，重点找：**外部做得好而我们没做的**（Gap 信号）
 5. 写到 `workspace/members/{dev}/drafts/research-agent-{N}-{视角}.md`
 
 > ⚠️ **不要凭训练记忆编数据**。搜不到就标"【待验证】"，不要编竞品功能/数据。
@@ -35,14 +45,17 @@ caller（/wl-insight research）会把所有 agent 的报告合并成一份调
 
 ## Output Format
 ```markdown
-# 调研报告（中间产物）：{视角}
+# 调研报告（中间产物）：{视角} · {方法论}
 
-## 视角
-{你负责的视角}
+## 视角与方法论
+{视角}，用 {方法论} 框架分析
 
 ## 本项目现状（来自 dispatch）
 {一句话摘要，让读者知道对标基准}
 
+## 分析（用方法论框架组织）
+{竞品矩阵 / 技术选型对比表 / JTBD+Kano 分类 / 趋势时间线}
+
 ## 关键发现
 1. {发现} —— 来源：{链接}
    - 对本项目的启发：{...}
@@ -57,5 +70,5 @@ caller（/wl-insight research）会把所有 agent 的报告合并成一份调
 ```
 
 > ⚠️ **本 agent 不生成 PRD。** 你是 research 类型的子 agent，产出中间产物。
-> 合并后的调研报告也不带 mode 标记（它不是 PRD）。
+> 合并后的调研报告也不带 mode 根记（它不是 PRD）。
 > caller 合并后会删除所有 `research-agent-*.md` 中间产物。