@alpacachen/dsh-kanban 1.11.3 → 1.12.1

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
package/README.md CHANGED
@@ -1,87 +1,116 @@
1
- <p align="center">
2
- <picture>
3
- <source media="(max-width: 600px)" srcset="docs/readme/cover-en-mobile.png">
4
- <img src="docs/readme/cover-en.png" alt="dsh-kanban — Plan together. See work move. A shared board for you and your agent in DeepSeek Harness." width="100%">
5
- </picture>
6
- </p>
1
+ <h1 align="center">dsh-kanban</h1>
2
+ <p align="center"><strong>A board for people. Workflows for AI.</strong><br>Plan, run and review work inside DeepSeek Harness.</p>
7
3
 
8
4
  <p align="center">
9
- <a href="https://www.npmjs.com/package/@alpacachen/dsh-kanban"><img src="https://img.shields.io/npm/v/@alpacachen/dsh-kanban?style=flat-square&color=3867e8&label=npm" alt="npm version"></a>
10
- <a href="LICENSE"><img src="https://img.shields.io/badge/license-MIT-3867e8?style=flat-square" alt="MIT license"></a>
5
+ <a href="https://www.npmjs.com/package/@alpacachen/dsh-kanban"><img src="https://img.shields.io/npm/v/@alpacachen/dsh-kanban?style=flat-square&color=7651c9" alt="npm version"></a>
6
+ <a href="https://github.com/alpacachen/dsh-kanban/actions/workflows/ci.yml"><img src="https://github.com/alpacachen/dsh-kanban/actions/workflows/ci.yml/badge.svg" alt="CI"></a>
7
+ <a href="LICENSE"><img src="https://img.shields.io/badge/license-MIT-7651c9?style=flat-square" alt="MIT license"></a>
11
8
  <a href="https://awesome-dsh-plugin.com"><img src="https://awesome-dsh-plugin.com/badge.svg" alt="Awesome DSH Plugin"></a>
12
9
  </p>
13
10
 
14
11
  <p align="center">
15
12
  <a href="README.zh.md">简体中文</a> · <strong>English</strong><br>
16
- <a href="#get-started">Get started</a> · <a href="docs/guide.md">User guide</a> · <a href="docs/guide.md#agent-tool-reference">Agent tools</a>
13
+ <a href="#get-started">Get started</a> · <a href="#kanban--work-you-lead">Kanban</a> · <a href="#workflow--work-ai-runs">Workflow</a> · <a href="#documentation">Documentation</a>
17
14
  </p>
18
15
 
19
- <br>
16
+ One plugin, two ways to work. Use **Kanban** to organize tasks and decide what happens next. Use **Workflow** to give AI a repeatable sequence, follow its execution, and step in at review points.
20
17
 
21
- Turn plans from your conversations into tasks you can move forward. You and your agent share one board for planning, context, and progress, all inside DeepSeek Harness.
18
+ Both live under **Kanban · Workflow** in the sidebar. Switch modes beside the workspace name; the plugin remembers your last mode and workspace.
22
19
 
23
- ![Kanban board with search and filters above tasks in Todo, In progress, Review, and Done](docs/readme/board-en.png)
20
+ | | Kanban · Work you lead | Workflow · Work AI runs |
21
+ | --- | --- | --- |
22
+ | Start with | A task or a plan from a conversation | A reusable process designed with your Agent |
23
+ | Move forward | You or your Agent update cards | AI executes stages in order |
24
+ | Keep control | Priorities, deadlines and review columns | Human approval, feedback, retry and stop |
25
+ | Keep the context | Descriptions, screenshots, comments and activity | Stage conversations, reports and task-specific files |
24
26
 
25
- <p align="center"><sub>Current plugin UI · Isolated demo data · A separate board for every workspace</sub></p>
27
+ ## Get started
26
28
 
27
- ## From a conversation to the next step
29
+ Requires **DSH 0.2.2-rc.3**, with desktop and Web support. For Workflow, configure an execution model in DSH first.
28
30
 
29
- ### Describe the goal. Let your agent create the tasks.
31
+ **Desktop** — Open **Plugins**, add `@alpacachen/dsh-kanban`, and enable it.
30
32
 
31
- Ask your agent to break down a requirement, write implementation steps and acceptance criteria, and assign labels and priorities. With 19 built-in tools, the plan goes straight onto the board.
33
+ **Web** — Install the plugin, then restart your Web instance to load it:
32
34
 
33
- > Break phone sign-in into testable tasks. Check existing cards first to avoid duplicates. Write implementation steps and acceptance criteria for each card, and mark launch blockers as P0.
35
+ ```sh
36
+ dsh plugin --profile web add @alpacachen/dsh-kanban
37
+ ```
34
38
 
35
- ### See the progress. Take over at any time.
39
+ Open **Kanban · Workflow** in the sidebar and choose a workspace. Stay in **Kanban** to create your first card, or switch to **Workflow**, choose **Create workflow in chat**, and describe a process.
36
40
 
37
- Drag cards between stages, then use search and filters to find the next task. Markdown descriptions, comments, and activity history keep the work's context in one place.
41
+ > Workflow is included in **1.12.0**. Until that release is published, installing from npm gives you the latest published version. To try this branch, [build and install it locally](docs/testing-images.md#local-workflow-build).
38
42
 
39
- | What matters | How to find it |
40
- | --- | --- |
41
- | Important work | P0 / P1 / P2 priorities and custom labels |
42
- | Upcoming deadlines | Due dates, due-today and overdue filters |
43
- | Existing decisions | Search card IDs, titles, full descriptions, and comments |
43
+ ## Kanban — Work you lead
44
44
 
45
- ### Bring the context back into conversation.
45
+ Turn plans into visible tasks. Move cards yourself, ask your Agent to keep them up to date, and take the next step when you are ready.
46
46
 
47
- Paste or drop screenshots into a task, then choose **Create new conversation** to bring its text and saved images into a new conversation in the selected workspace. Everything starts as an editable draft for you to review and send.
47
+ ![Kanban board with task search, priority and label filters, and four task columns](docs/readme/board-en.png)
48
48
 
49
- [Image and saving rules →](docs/guide.md#usage-details) · [More prompt examples →](docs/guide.md#prompt-examples)
49
+ <sub>Actual plugin UI with isolated example data. Each workspace has its own board.</sub>
50
50
 
51
- ## Get started
51
+ - **Plan together.** Ask your Agent to break down a goal, write acceptance criteria and create cards with 19 built-in board tools.
52
+ - **Find the next task.** Search titles, descriptions, comments and IDs; combine priorities, labels and due-date filters.
53
+ - **Keep evidence close.** Add Markdown, screenshots, comments and activity history to a card.
54
+ - **Continue in chat.** Bring a card and its saved images into a new conversation as an editable draft, ready for you to send.
52
55
 
53
- **Desktop** — Open **Plugins**, add `@alpacachen/dsh-kanban`, and enable it.
56
+ Try asking your Agent:
54
57
 
55
- **Web** — Run the command below, then restart `dsh web` to load the plugin.
58
+ > Break phone sign-in into testable tasks. Check existing cards first to avoid duplicates. Add implementation steps and acceptance criteria, and mark launch blockers as P0.
56
59
 
57
- ```sh
58
- dsh plugin --profile web add @alpacachen/dsh-kanban
59
- ```
60
+ [Kanban guide →](docs/guide.md) · [All 19 board tools →](docs/guide.md#agent-tool-reference)
60
61
 
61
- Open **Board** in the upper-left sidebar, then click the workspace name at the top to switch boards. Start with four default stages:
62
+ ## Workflow — Work AI runs
62
63
 
63
- **Todo → In Progress → Review → Done**
64
+ Describe a process once, then give it new inputs. AI runs its stages using DSH's normal tools and authorization; you can follow the work and decide at human review points.
64
65
 
65
- Requires DSH **0.2.2-rc.3**, on Web and desktop. Upgrade DSH before installing Kanban 1.11.3.
66
+ <picture>
67
+ <source media="(max-width: 600px)" srcset="docs/readme/workflow-en-mobile.svg">
68
+ <img src="docs/readme/workflow-en.svg" alt="Example workflow: read materials, draft a report, wait for human review, then finalize; feedback returns to drafting" width="100%">
69
+ </picture>
66
70
 
67
- <details>
68
- <summary>Check the Web plugin configuration</summary>
71
+ <sub>An example process. Human review pauses execution until you approve, return it with feedback, or stop.</sub>
69
72
 
70
- ```sh
71
- dsh --profile web --dump-config
72
- ```
73
+ ### Design it in conversation
74
+
75
+ Ask your Agent to create or revise a workflow. It saves the stages, instructions, deliverables and review points. New AI stages inherit the model used in that conversation; you can later edit stage instructions and choose a model and reasoning effort in the stage dialog.
76
+
77
+ > Create a reusable research workflow: read the materials, draft a report with sources and open questions, wait for my review, then finalize the report. If I return it, go back to drafting. Save the workflow without running a task.
78
+
79
+ ### Run it with a new input
80
+
81
+ Expand a workflow in the home list and choose **Create task**. Enter what you need and send it. No separate task name or model setup is required.
73
82
 
74
- Look for a `dsh-kanban` layer. Its presence confirms the configuration, not that the running service has loaded it; restart the Web service after installation.
83
+ - **See the work.** The diagram highlights the current stage. Click a stage to inspect its native DSH conversation, available reasoning text, tool calls and replies.
84
+ - **Review and revise.** Approve to continue or return with feedback. Earlier reports and review decisions remain available.
85
+ - **Keep files separate.** Every task has its own output directory, shared by its stages. Preview files in place, open the directory, or delete deliverables with confirmation.
86
+ - **Resume deliberately.** Retry a failed stage after checking the cause. After a crash, reconcile the old execution before retrying; completed upstream stages are retained.
75
87
 
76
- </details>
88
+ ### Improve the process from its history
77
89
 
78
- ## Explore further
90
+ Choose **Optimize from history** to start a conversation with an analysis prompt. The Agent can inspect previous task results, review feedback, stage conversations and recorded timings. Changes apply to future tasks; existing tasks keep their original workflow version.
79
91
 
80
- - [User guide](docs/guide.md) — Editing, search, images, and collaboration.
81
- - [Agent tool reference](docs/guide.md#agent-tool-reference) — All 19 tools, arguments, and completion rules.
82
- - [Data and backups](docs/guide.md#data-and-backups) — Keep both the board file and the DSH attachment store to preserve images.
83
- - [Development and validation](docs/testing-images.md) — Node.js 26, pnpm 11.8.0, builds, and tests.
84
- - [Issues and ideas](https://github.com/alpacachen/dsh-kanban/issues) — Report a problem or share feedback.
92
+ [Workflow guide →](docs/workflow-guide.en.md) · [Agent interface (中文) →](docs/workflow-agent-interface-review.md) · [Validation and recovery boundaries (中文) →](docs/workflow-validation.md)
93
+
94
+ ## What to expect
95
+
96
+ - **Native DSH execution.** Workflow adds orchestration and human review, with no separate tool-permission system or extra model-turn, tool-call or stage-time budgets.
97
+ - **Sequential stages.** Stages in a task run in order. Tool concurrency and model-request retries follow DSH; a failed workflow stage is retried by you.
98
+ - **Small handoffs.** Stages receive upstream conclusions, evidence references and a file index. Full reports are read when needed instead of carrying the entire conversation forward.
99
+ - **Review scope is explicit.** Approval applies to the saved stage reports. It does not snapshot project files or undo changes made by tools.
100
+ - **Two interface languages.** Controls, owned error messages and dates follow DSH's Chinese or English setting. Your inputs, prompts and saved outputs retain their original language.
101
+
102
+ If DSH is forcibly killed, previously launched commands can survive on macOS. Check that old commands and external operations have stopped before confirming recovery. The plugin will not silently rerun an interrupted stage.
103
+
104
+ ## Documentation
105
+
106
+ | You want to… | Read |
107
+ | --- | --- |
108
+ | Use cards, search, images and chat handoff | [Kanban guide](docs/guide.md) |
109
+ | Create, run, review and improve workflows | [Workflow guide](docs/workflow-guide.en.md) |
110
+ | Understand board tools | [Board tool reference](docs/guide.md#agent-tool-reference) |
111
+ | Back up your data | [Kanban data](docs/guide.md#data-and-backups) · [Workflow data](docs/workflow-guide.en.md#data-and-backups) |
112
+ | Build and validate the plugin | [Development checks](docs/testing-images.md) · [Workflow acceptance record](docs/workflow-validation.md) |
113
+ | Report a problem or suggest a change | [GitHub Issues](https://github.com/alpacachen/dsh-kanban/issues) |
85
114
 
86
115
  ---
87
116
 
package/README.zh.md CHANGED
@@ -1,87 +1,116 @@
1
- <p align="center">
2
- <picture>
3
- <source media="(max-width: 600px)" srcset="docs/readme/cover-zh-mobile.png">
4
- <img src="docs/readme/cover-zh.png" alt="dsh-kanban — 一起规划,让进展可见。你与 Agent 共用的 DeepSeek Harness 看板。" width="100%">
5
- </picture>
6
- </p>
1
+ <h1 align="center">dsh-kanban</h1>
2
+ <p align="center"><strong>面向人的看板,面向 AI 的工作流。</strong><br>在 DeepSeek Harness 中规划、执行与审核工作。</p>
7
3
 
8
4
  <p align="center">
9
- <a href="https://www.npmjs.com/package/@alpacachen/dsh-kanban"><img src="https://img.shields.io/npm/v/@alpacachen/dsh-kanban?style=flat-square&color=3867e8&label=npm" alt="npm 版本"></a>
10
- <a href="LICENSE"><img src="https://img.shields.io/badge/license-MIT-3867e8?style=flat-square" alt="MIT 许可"></a>
5
+ <a href="https://www.npmjs.com/package/@alpacachen/dsh-kanban"><img src="https://img.shields.io/npm/v/@alpacachen/dsh-kanban?style=flat-square&color=7651c9" alt="npm 版本"></a>
6
+ <a href="https://github.com/alpacachen/dsh-kanban/actions/workflows/ci.yml"><img src="https://github.com/alpacachen/dsh-kanban/actions/workflows/ci.yml/badge.svg" alt="CI"></a>
7
+ <a href="LICENSE"><img src="https://img.shields.io/badge/license-MIT-7651c9?style=flat-square" alt="MIT 许可"></a>
11
8
  <a href="https://awesome-dsh-plugin.com"><img src="https://awesome-dsh-plugin.com/badge.svg" alt="Awesome DSH Plugin"></a>
12
9
  </p>
13
10
 
14
11
  <p align="center">
15
12
  <strong>简体中文</strong> · <a href="README.md">English</a><br>
16
- <a href="#开始使用">开始使用</a> · <a href="docs/guide.zh.md">使用指南</a> · <a href="docs/guide.zh.md#agent-工具参考">Agent 工具</a>
13
+ <a href="#开始使用">开始使用</a> · <a href="#kanban面向人的看板">Kanban</a> · <a href="#workflow面向-ai-的工作流">Workflow</a> · <a href="#文档导航">文档</a>
17
14
  </p>
18
15
 
19
- <br>
16
+ 一个插件,两种工作方式。用 **Kanban** 安排任务、掌握进度、决定下一步;用 **Workflow** 把可重复的步骤交给 AI,查看执行过程,在需要判断的地方参与审核。
17
+
18
+ 两种功能共用侧边栏的 **「看板·工作流」** 入口,在工作区名称右侧切换;插件会记住上次打开的模式和工作区。
19
+
20
+ | | Kanban · 由你推进 | Workflow · 由 AI 执行 |
21
+ | --- | --- | --- |
22
+ | 从哪里开始 | 一项任务,或对话中拆解出的计划 | 与 Agent 一起设计的可复用流程 |
23
+ | 怎么推进 | 你或 Agent 更新卡片 | AI 按阶段依次执行 |
24
+ | 怎么掌控 | 优先级、截止日期、验收列 | 人工审核、反馈、重试与终止 |
25
+ | 留下什么 | 描述、截图、评论与活动记录 | 阶段会话、报告与任务独立产物 |
26
+
27
+ ## 开始使用
28
+
29
+ 需要 **DSH 0.2.2-rc.3**,支持桌面端与 Web。使用 Workflow 前,先在 DSH 中配置可用模型。
20
30
 
21
- 把对话里的计划变成可推进的任务。你和 Agent 在同一块看板上拆解工作、补充上下文、更新进度,始终留在 DeepSeek Harness 中。
31
+ **桌面端** — 打开「插件」,添加并启用 `@alpacachen/dsh-kanban`。
32
+
33
+ **Web** — 安装插件后,重启 Web 实例以加载:
34
+
35
+ ```sh
36
+ dsh plugin --profile web add @alpacachen/dsh-kanban
37
+ ```
22
38
 
23
- ![Kanban 看板:顶部搜索与筛选,下方展示待办、进行中、待验收和已完成四列任务](docs/readme/board-zh.png)
39
+ 打开侧边栏的 **「看板·工作流」**,选择工作区。在 **Kanban** 中创建第一张卡片,或切到 **Workflow**,点击「对话创建流程」,描述你希望重复执行的工作。
24
40
 
25
- <p align="center"><sub>当前插件界面 · 隔离示例数据 · 每个工作区独立保存</sub></p>
41
+ > Workflow 随 **1.12.0** 提供。在该版本发布前,npm 安装得到的仍是当前已发布版本;尝试本分支请[本地构建安装](docs/testing-images.md#local-workflow-build)。
26
42
 
27
- ## 从一句话,到下一步行动
43
+ ## Kanban:面向人的看板
28
44
 
29
- ### 说清目标,Agent 来建卡
45
+ 把计划变成看得见的任务。自己拖动卡片,也可以让 Agent 整理和更新,下一步由你决定。
30
46
 
31
- 让 Agent 拆解需求,写下实现步骤与验收标准,安排标签和优先级。19 个内置工具,让规划直接落到看板上。
47
+ ![Kanban 看板:任务搜索、优先级与标签筛选,以及四列任务](docs/readme/board-zh.png)
48
+
49
+ <sub>插件真实界面,使用隔离示例数据。每个工作区有独立看板。</sub>
50
+
51
+ - **一起规划。** 让 Agent 拆解目标、写验收标准,用 19 个看板工具直接创建卡片。
52
+ - **找到下一件事。** 搜索标题、描述、评论和 ID,组合优先级、标签与截止日期筛选。
53
+ - **上下文留在任务里。** Markdown、截图、评论和活动记录,放在同一张卡片中。
54
+ - **带着任务继续聊。** 把卡片文字和已保存的图片带入新会话草稿,检查后再发送。
55
+
56
+ 可以这样告诉 Agent:
32
57
 
33
58
  > 把手机号登录拆成可验收的任务,先检查已有卡片,避免重复。每张卡写清实现步骤和验收标准,把上线阻塞项标为 P0。
34
59
 
35
- ### 看清进展,随时接手
60
+ [看板使用指南 →](docs/guide.zh.md) · [全部 19 个看板工具 →](docs/guide.zh.md#agent-工具参考)
36
61
 
37
- 拖动卡片调整阶段,用搜索和筛选找到下一件事。任务描述支持 Markdown,评论和活动记录留下推进过程。
62
+ ## Workflow:面向 AI 的工作流
38
63
 
39
- | 关注什么 | 怎么找到 |
40
- | --- | --- |
41
- | 重要的事 | P0 / P1 / P2 优先级与自定义标签 |
42
- | 快到期的事 | 截止日期、今天到期与逾期筛选 |
43
- | 已有的结论 | 搜索卡片 ID、标题、完整描述和评论 |
64
+ 把方法设计一次,再交给它不同的输入。AI 沿用 DSH 原有工具和授权执行各阶段;你能查看工作过程,也能在人工节点决定是否继续。
44
65
 
45
- ### 带上上下文,继续对话
66
+ <picture>
67
+ <source media="(max-width: 600px)" srcset="docs/readme/workflow-zh-mobile.svg">
68
+ <img src="docs/readme/workflow-zh.svg" alt="工作流示例:读取材料、起草报告、等待人工审核、整理定稿;打回后返回起草阶段" width="100%">
69
+ </picture>
46
70
 
47
- 把截图粘贴或拖进任务,点击「创建新对话」,将任务文字和已保存的图片一起带到所选工作区的新会话。内容先进入可编辑草稿,由你确认发送。
71
+ <sub>示例流程。到达人工审核点会暂停,等待你通过、填写意见打回,或终止任务。</sub>
48
72
 
49
- [查看图片与保存规则 →](docs/guide.zh.md#使用规则) · [更多对话示例 →](docs/guide.zh.md#对话示例)
73
+ ### 在对话中设计方法
50
74
 
51
- ## 开始使用
75
+ 让 Agent 创建或修改流程,保存阶段、执行要求、交付结果和审核点。新增 AI 阶段继承当时聊天的模型配置;之后可以在阶段弹窗中编辑执行文本,选择模型与推理档位。
52
76
 
53
- **桌面端** — 打开「插件」,添加并启用 `@alpacachen/dsh-kanban`。
77
+ > 帮我创建一个可复用的资料研究流程:先读取材料,再起草带来源和待确认问题的报告,等我审核,通过后整理定稿;打回时回到起草阶段。只保存流程,先不要运行任务。
54
78
 
55
- **Web** — 运行以下命令,安装后重启 `dsh web` 以加载插件。
79
+ ### 输入任务,查看执行
56
80
 
57
- ```sh
58
- dsh plugin --profile web add @alpacachen/dsh-kanban
59
- ```
81
+ 在首页展开工作流,点击「创建任务」,输入本次需要完成的内容即可发送,无需另填任务名称或选择模型。
60
82
 
61
- 点击左上角的 **「看板」**,点击顶部工作区名称切换看板,即可从四个默认阶段开始:
83
+ - **看见工作过程。** 流程图高亮当前阶段,点击节点查看 DSH 原生会话中可见的思考文本、工具调用与回复。
84
+ - **审核与返工。** 通过后继续,填写意见后打回;之前的报告和审核决定仍然保留。
85
+ - **产物互不混杂。** 每个任务拥有独立目录,同一任务的各阶段共用;可原地预览、打开目录,或确认后删除产物。
86
+ - **确认后再恢复。** 失败后先核对原因,再重试当前阶段;异常退出后先核对旧执行,上游已完成阶段保留。
62
87
 
63
- **Todo → In Progress → Review → Done**
88
+ ### 用历史任务改进方法
64
89
 
65
- 要求 DSH **0.2.2-rc.3**,支持 Web 与桌面端。安装 Kanban 1.11.3 前请先升级 DSH。
90
+ 点击「根据历史优化」,带着预置分析要求进入新对话。Agent 可以读取历史结果、审核意见、阶段会话与已记录的耗时来改进流程。修改只影响未来的新任务,已有任务继续使用原流程版本。
66
91
 
67
- <details>
68
- <summary>确认 Web 插件配置</summary>
92
+ [工作流使用指南 →](docs/workflow-guide.md) · [Agent 接口说明 →](docs/workflow-agent-interface-review.md) · [验收记录与恢复边界 →](docs/workflow-validation.md)
69
93
 
70
- ```sh
71
- dsh --profile web --dump-config
72
- ```
94
+ ## 运行方式
73
95
 
74
- 输出中应包含 `dsh-kanban` 配置层。配置存在不代表当前服务已加载,安装后仍需重启 Web 服务。
96
+ - **复用 DSH。** Workflow 负责调度与人工审核,不另设工具权限,也不增加模型轮数、工具次数或阶段时长预算。
97
+ - **阶段依次执行。** 同一任务按流程顺序推进。工具并发与模型请求重试沿用 DSH;失败阶段由你选择重试。
98
+ - **简短交接。** 阶段之间传递上游结论、依据位置和产物索引,需要时再读完整报告,不反复携带整段聊天。
99
+ - **审核对象明确。** 通过针对当前保存的阶段报告,不会为项目文件生成快照,也不会撤销工具已经做出的改动。
100
+ - **中英文界面。** 控件、内置错误提示和日期跟随 DSH 语言;任务输入、阶段要求和历史产物保持原文。
75
101
 
76
- </details>
102
+ macOS 上强杀 DSH 后,先前启动的命令可能继续运行。恢复前需要核对旧命令及外部操作已停止;插件不会静默重跑中断阶段。
77
103
 
78
- ## 继续了解
104
+ ## 文档导航
79
105
 
80
- - [使用指南](docs/guide.zh.md) — 编辑、搜索、图片与协作。
81
- - [Agent 工具参考](docs/guide.zh.md#agent-工具参考) — 全部 19 个工具、参数与完成状态规则。
82
- - [数据与备份](docs/guide.zh.md#数据与备份) — 图片备份需同时保留看板文件和 DSH 附件存储。
83
- - [开发与验证](docs/testing-images.md) — Node.js 26、pnpm 11.8.0,构建和测试说明。
84
- - [问题与建议](https://github.com/alpacachen/dsh-kanban/issues) — 提交问题或分享使用反馈。
106
+ | 想了解什么 | 阅读 |
107
+ | --- | --- |
108
+ | 卡片、搜索、图片与对话交接 | [看板指南](docs/guide.zh.md) |
109
+ | 创建、执行、审核和优化流程 | [工作流指南](docs/workflow-guide.md) |
110
+ | Agent 如何管理看板 | [看板工具参考](docs/guide.zh.md#agent-工具参考) |
111
+ | 备份数据 | [看板数据](docs/guide.zh.md#数据与备份) · [工作流数据](docs/workflow-guide.md#数据与备份) |
112
+ | 开发与验证 | [开发检查](docs/testing-images.md) · [工作流验收记录](docs/workflow-validation.md) |
113
+ | 问题与建议 | [GitHub Issues](https://github.com/alpacachen/dsh-kanban/issues) |
85
114
 
86
115
  ---
87
116
 
package/cordis.patch.yml CHANGED
@@ -4,3 +4,5 @@
4
4
  - insert:
5
5
  - id: dsh-kanban
6
6
  name: "@alpacachen/dsh-kanban"
7
+ - id: dsh-kanban-workflow
8
+ name: "@alpacachen/dsh-kanban/workflow"
package/docs/guide.md CHANGED
@@ -4,7 +4,7 @@
4
4
 
5
5
  ## Using the board
6
6
 
7
- Open **Board** in the upper-left sidebar and click the workspace name at the top to switch boards. The board remembers your selection. Switching workspaces does not create or navigate to a conversation; each workspace keeps its own board.
7
+ Open **Kanban · Workflow** in the upper-left sidebar and click the workspace name at the top to switch boards. The board remembers your selection. Switching workspaces does not create or navigate to a conversation; each workspace keeps its own board.
8
8
 
9
9
  | Action | How it works |
10
10
  | --- | --- |
package/docs/guide.zh.md CHANGED
@@ -4,7 +4,7 @@
4
4
 
5
5
  ## 使用看板
6
6
 
7
- 点击左上角的「看板」,点击顶部工作区名称切换看板。看板记住上次的选择,切换工作区不会创建或切换会话。每个工作区的数据仍独立保存。
7
+ 点击左上角的「看板·工作流」,点击顶部工作区名称切换看板。看板记住上次的选择,切换工作区不会创建或切换会话。每个工作区的数据仍独立保存。
8
8
 
9
9
  | 操作 | 使用方式 |
10
10
  | --- | --- |
@@ -0,0 +1,28 @@
1
+ <svg xmlns="http://www.w3.org/2000/svg" width="600" height="900" viewBox="0 0 600 900" role="img" aria-labelledby="title desc">
2
+ <title id="title">A workflow with human review</title>
3
+ <desc id="desc">Read, draft, review, finalize. Feedback returns to drafting.</desc>
4
+ <defs><marker id="arrow" viewBox="0 0 10 10" refX="9" refY="5" markerWidth="7" markerHeight="7" orient="auto-start-reverse"><path d="M0 0L10 5L0 10" fill="none" stroke="#7b7391" stroke-width="1.6"/></marker></defs>
5
+ <rect width="600" height="900" rx="20" fill="#f7f6fa"/>
6
+ <g font-family="-apple-system,BlinkMacSystemFont,Segoe UI,PingFang SC,Microsoft YaHei,sans-serif">
7
+ <rect x="86" y="55" width="428" height="132" rx="12" fill="#fff" stroke="#ddd9e6"/>
8
+ <text x="110" y="87" font-size="16" fill="#766785">1 · AI</text>
9
+ <text x="110" y="127" font-size="28" font-weight="600" fill="#241d34">Read materials</text>
10
+ <text x="110" y="160" font-size="17" fill="#625b70">Collect sources</text>
11
+ <path d="M300.0 195V246" fill="none" stroke="#7b7391" stroke-width="2" marker-end="url(#arrow)"/>
12
+ <rect x="86" y="258" width="428" height="132" rx="12" fill="#fff" stroke="#ddd9e6"/>
13
+ <text x="110" y="290" font-size="16" fill="#766785">2 · AI</text>
14
+ <text x="110" y="330" font-size="28" font-weight="600" fill="#241d34">Draft a report</text>
15
+ <text x="110" y="363" font-size="17" fill="#625b70">Conclusions + evidence</text>
16
+ <path d="M300.0 398V449" fill="none" stroke="#7b7391" stroke-width="2" marker-end="url(#arrow)"/>
17
+ <rect x="86" y="461" width="428" height="132" rx="12" fill="#ebe4fb" stroke="#9b7fd1"/>
18
+ <text x="110" y="493" font-size="16" fill="#766785">3 · Human</text>
19
+ <text x="110" y="533" font-size="28" font-weight="600" fill="#241d34">Human review</text>
20
+ <text x="110" y="566" font-size="17" fill="#625b70">Approve to continue</text>
21
+ <path d="M300.0 601V652" fill="none" stroke="#7b7391" stroke-width="2" marker-end="url(#arrow)"/>
22
+ <rect x="86" y="664" width="428" height="132" rx="12" fill="#fff" stroke="#ddd9e6"/>
23
+ <text x="110" y="696" font-size="16" fill="#766785">4 · AI</text>
24
+ <text x="110" y="736" font-size="28" font-weight="600" fill="#241d34">Finalize</text>
25
+ <text x="110" y="769" font-size="17" fill="#625b70">Save task deliverables</text>
26
+ <path d="M80 527H40V324H75" fill="none" stroke="#8a6ab9" stroke-width="2" stroke-dasharray="6 6" marker-end="url(#arrow)"/>
27
+ <text x="90" y="882" font-size="18" fill="#72539d">Return with feedback</text>
28
+ </g></svg>
@@ -0,0 +1,28 @@
1
+ <svg xmlns="http://www.w3.org/2000/svg" width="1440" height="360" viewBox="0 0 1440 360" role="img" aria-labelledby="title desc">
2
+ <title id="title">A workflow with human review</title>
3
+ <desc id="desc">Read, draft, review, finalize. Feedback returns to drafting.</desc>
4
+ <defs><marker id="arrow" viewBox="0 0 10 10" refX="9" refY="5" markerWidth="7" markerHeight="7" orient="auto-start-reverse"><path d="M0 0L10 5L0 10" fill="none" stroke="#7b7391" stroke-width="1.6"/></marker></defs>
5
+ <rect width="1440" height="360" rx="20" fill="#f7f6fa"/>
6
+ <g font-family="-apple-system,BlinkMacSystemFont,Segoe UI,PingFang SC,Microsoft YaHei,sans-serif">
7
+ <rect x="44" y="70" width="296" height="140" rx="12" fill="#fff" stroke="#ddd9e6"/>
8
+ <text x="68" y="102" font-size="16" fill="#766785">1 · AI</text>
9
+ <text x="68" y="142" font-size="28" font-weight="600" fill="#241d34">Read materials</text>
10
+ <text x="68" y="175" font-size="17" fill="#625b70">Collect sources</text>
11
+ <path d="M348 140.0H384" fill="none" stroke="#7b7391" stroke-width="2" marker-end="url(#arrow)"/>
12
+ <rect x="396" y="70" width="296" height="140" rx="12" fill="#fff" stroke="#ddd9e6"/>
13
+ <text x="420" y="102" font-size="16" fill="#766785">2 · AI</text>
14
+ <text x="420" y="142" font-size="28" font-weight="600" fill="#241d34">Draft a report</text>
15
+ <text x="420" y="175" font-size="17" fill="#625b70">Conclusions + evidence</text>
16
+ <path d="M700 140.0H736" fill="none" stroke="#7b7391" stroke-width="2" marker-end="url(#arrow)"/>
17
+ <rect x="748" y="70" width="296" height="140" rx="12" fill="#ebe4fb" stroke="#9b7fd1"/>
18
+ <text x="772" y="102" font-size="16" fill="#766785">3 · Human</text>
19
+ <text x="772" y="142" font-size="28" font-weight="600" fill="#241d34">Human review</text>
20
+ <text x="772" y="175" font-size="17" fill="#625b70">Approve to continue</text>
21
+ <path d="M1052 140.0H1088" fill="none" stroke="#7b7391" stroke-width="2" marker-end="url(#arrow)"/>
22
+ <rect x="1100" y="70" width="296" height="140" rx="12" fill="#fff" stroke="#ddd9e6"/>
23
+ <text x="1124" y="102" font-size="16" fill="#766785">4 · AI</text>
24
+ <text x="1124" y="142" font-size="28" font-weight="600" fill="#241d34">Finalize</text>
25
+ <text x="1124" y="175" font-size="17" fill="#625b70">Save task deliverables</text>
26
+ <path d="M896 224V282H544V224" fill="none" stroke="#8a6ab9" stroke-width="2" stroke-dasharray="6 6" marker-end="url(#arrow)"/>
27
+ <text x="640" y="318" font-size="18" fill="#72539d">Return with feedback</text>
28
+ </g></svg>
@@ -0,0 +1,28 @@
1
+ <svg xmlns="http://www.w3.org/2000/svg" width="600" height="900" viewBox="0 0 600 900" role="img" aria-labelledby="title desc">
2
+ <title id="title">包含人工审核的工作流</title>
3
+ <desc id="desc">依次读取材料、起草、审核和定稿;反馈返回起草阶段。</desc>
4
+ <defs><marker id="arrow" viewBox="0 0 10 10" refX="9" refY="5" markerWidth="7" markerHeight="7" orient="auto-start-reverse"><path d="M0 0L10 5L0 10" fill="none" stroke="#7b7391" stroke-width="1.6"/></marker></defs>
5
+ <rect width="600" height="900" rx="20" fill="#f7f6fa"/>
6
+ <g font-family="-apple-system,BlinkMacSystemFont,Segoe UI,PingFang SC,Microsoft YaHei,sans-serif">
7
+ <rect x="86" y="55" width="428" height="132" rx="12" fill="#fff" stroke="#ddd9e6"/>
8
+ <text x="110" y="87" font-size="16" fill="#766785">1 · AI 执行</text>
9
+ <text x="110" y="127" font-size="28" font-weight="600" fill="#241d34">读取材料</text>
10
+ <text x="110" y="160" font-size="17" fill="#625b70">整理来源与任务输入</text>
11
+ <path d="M300.0 195V246" fill="none" stroke="#7b7391" stroke-width="2" marker-end="url(#arrow)"/>
12
+ <rect x="86" y="258" width="428" height="132" rx="12" fill="#fff" stroke="#ddd9e6"/>
13
+ <text x="110" y="290" font-size="16" fill="#766785">2 · AI 执行</text>
14
+ <text x="110" y="330" font-size="28" font-weight="600" fill="#241d34">起草报告</text>
15
+ <text x="110" y="363" font-size="17" fill="#625b70">结论、依据与待确认项</text>
16
+ <path d="M300.0 398V449" fill="none" stroke="#7b7391" stroke-width="2" marker-end="url(#arrow)"/>
17
+ <rect x="86" y="461" width="428" height="132" rx="12" fill="#ebe4fb" stroke="#9b7fd1"/>
18
+ <text x="110" y="493" font-size="16" fill="#766785">3 · 由你确认</text>
19
+ <text x="110" y="533" font-size="28" font-weight="600" fill="#241d34">人工审核</text>
20
+ <text x="110" y="566" font-size="17" fill="#625b70">通过后继续,打回后返工</text>
21
+ <path d="M300.0 601V652" fill="none" stroke="#7b7391" stroke-width="2" marker-end="url(#arrow)"/>
22
+ <rect x="86" y="664" width="428" height="132" rx="12" fill="#fff" stroke="#ddd9e6"/>
23
+ <text x="110" y="696" font-size="16" fill="#766785">4 · AI 执行</text>
24
+ <text x="110" y="736" font-size="28" font-weight="600" fill="#241d34">整理定稿</text>
25
+ <text x="110" y="769" font-size="17" fill="#625b70">保存到本次任务目录</text>
26
+ <path d="M80 527H40V324H75" fill="none" stroke="#8a6ab9" stroke-width="2" stroke-dasharray="6 6" marker-end="url(#arrow)"/>
27
+ <text x="90" y="882" font-size="18" fill="#72539d">填写意见,返回起草</text>
28
+ </g></svg>
@@ -0,0 +1,28 @@
1
+ <svg xmlns="http://www.w3.org/2000/svg" width="1440" height="360" viewBox="0 0 1440 360" role="img" aria-labelledby="title desc">
2
+ <title id="title">包含人工审核的工作流</title>
3
+ <desc id="desc">依次读取材料、起草、审核和定稿;反馈返回起草阶段。</desc>
4
+ <defs><marker id="arrow" viewBox="0 0 10 10" refX="9" refY="5" markerWidth="7" markerHeight="7" orient="auto-start-reverse"><path d="M0 0L10 5L0 10" fill="none" stroke="#7b7391" stroke-width="1.6"/></marker></defs>
5
+ <rect width="1440" height="360" rx="20" fill="#f7f6fa"/>
6
+ <g font-family="-apple-system,BlinkMacSystemFont,Segoe UI,PingFang SC,Microsoft YaHei,sans-serif">
7
+ <rect x="44" y="70" width="296" height="140" rx="12" fill="#fff" stroke="#ddd9e6"/>
8
+ <text x="68" y="102" font-size="16" fill="#766785">1 · AI 执行</text>
9
+ <text x="68" y="142" font-size="28" font-weight="600" fill="#241d34">读取材料</text>
10
+ <text x="68" y="175" font-size="17" fill="#625b70">整理来源与任务输入</text>
11
+ <path d="M348 140.0H384" fill="none" stroke="#7b7391" stroke-width="2" marker-end="url(#arrow)"/>
12
+ <rect x="396" y="70" width="296" height="140" rx="12" fill="#fff" stroke="#ddd9e6"/>
13
+ <text x="420" y="102" font-size="16" fill="#766785">2 · AI 执行</text>
14
+ <text x="420" y="142" font-size="28" font-weight="600" fill="#241d34">起草报告</text>
15
+ <text x="420" y="175" font-size="17" fill="#625b70">结论、依据与待确认项</text>
16
+ <path d="M700 140.0H736" fill="none" stroke="#7b7391" stroke-width="2" marker-end="url(#arrow)"/>
17
+ <rect x="748" y="70" width="296" height="140" rx="12" fill="#ebe4fb" stroke="#9b7fd1"/>
18
+ <text x="772" y="102" font-size="16" fill="#766785">3 · 由你确认</text>
19
+ <text x="772" y="142" font-size="28" font-weight="600" fill="#241d34">人工审核</text>
20
+ <text x="772" y="175" font-size="17" fill="#625b70">通过后继续,打回后返工</text>
21
+ <path d="M1052 140.0H1088" fill="none" stroke="#7b7391" stroke-width="2" marker-end="url(#arrow)"/>
22
+ <rect x="1100" y="70" width="296" height="140" rx="12" fill="#fff" stroke="#ddd9e6"/>
23
+ <text x="1124" y="102" font-size="16" fill="#766785">4 · AI 执行</text>
24
+ <text x="1124" y="142" font-size="28" font-weight="600" fill="#241d34">整理定稿</text>
25
+ <text x="1124" y="175" font-size="17" fill="#625b70">保存到本次任务目录</text>
26
+ <path d="M896 224V282H544V224" fill="none" stroke="#8a6ab9" stroke-width="2" stroke-dasharray="6 6" marker-end="url(#arrow)"/>
27
+ <text x="640" y="318" font-size="18" fill="#72539d">填写意见,返回起草</text>
28
+ </g></svg>
@@ -34,3 +34,15 @@ Keep authentication logs, generated reports, and personal-workspace screenshots
34
34
  ## Desktop acceptance
35
35
 
36
36
  Verify the actual installed plugin: search and the three adjacent filters, Settings actions, image selection/paste/drop, native full-size preview, Escape/close/backdrop behavior, and task-draft preservation. Check both current- and new-conversation handoff: saved task images must appear as removable composer attachments without submitting the message. Use disposable test tasks for writes.
37
+
38
+ ## Local Workflow build
39
+
40
+ The 1.12.0 source includes both Kanban and Workflow. Before the npm release is available, build this checkout with the commands above, then install it into your intended profile using DSH's normal local-plugin flow. For example, for a separate Web profile:
41
+
42
+ ```sh
43
+ dsh plugin --profile web add file:/absolute/path/to/dsh-kanban
44
+ ```
45
+
46
+ Reload the plugin or restart that Web instance after installation. Do not restart a DSH instance that is executing work without checking its tasks first. Use a separate disposable profile for fault injection.
47
+
48
+ `node scripts/check-workflow-executor.mjs` exercises an installed DSH kernel with a local model fixture; configure `DSH_INSTALL_DIR` to its DSH package directory. It does not call a paid model. Real-model, human-review and interruption checks are recorded in [Workflow validation](workflow-validation.md).