TaskDoc MCP

TaskDoc MCP

Enables management of Kanban-backed task checkpoint documents in Obsidian, supporting task creation, querying, checkpoint commits, and rich block storage via MCP tools.

Category
访问服务器

README

TaskDoc MCP

TaskDoc MCP is a desktop-only Obsidian plugin for Kanban-backed task checkpoint documents. It keeps long-term task records concise while preserving large tables, Mermaid diagrams, CP checklists, specifications, and test evidence as addressable rich blocks.

This repository is an early test release extracted from the Taskflow design in win-console.

Alpha scope

  • Register one or more Obsidian Kanban Markdown boards.
  • Treat a board as a project, a column as a task type, and a linked card as a task.
  • Create a linked task document and card through MCP.
  • Commit validated checkpoint cores with fixed rendering.
  • Store Markdown tables, Mermaid, domain checklists, configuration, specifications, and evidence as rich blocks up to 24,576 characters by default.
  • Write a mutable handoff capsule and resume a task without loading its full history or rich block bodies.
  • Move a card between type columns and synchronize its checkbox completion state.
  • Run a token-authenticated Streamable HTTP MCP endpoint on 127.0.0.1.
  • Configure boards, MCP, document budgets, and diagnostics from the Obsidian plugin settings page.

PHA synchronization is intentionally unavailable in this alpha. The adapter/outbox boundary and settings placeholder are included, but a real implementation requires the concrete PHA task/comment API, especially comment update semantics.

Requirements

  • Obsidian Desktop 1.11.4 or newer.
  • An Obsidian Kanban board represented as Markdown headings and checkbox cards.
  • An MCP client that supports remote/Streamable HTTP servers.

Do not enable the old win-console Taskflow writer and TaskDoc MCP against the same files at the same time.

Install from a release

  1. Download taskdoc-mcp-0.1.0-alpha.1.zip from the GitHub Release.
  2. Extract it to <vault>/.obsidian/plugins/taskdoc-mcp/.
  3. Confirm that the folder directly contains main.js, manifest.json, and styles.css.
  4. Enable TaskDoc MCP under Obsidian → Settings → Community plugins.
  5. Open Obsidian → Settings → TaskDoc MCP.

The plugin is desktop-only because the MCP server uses the Node.js HTTP API.

Configure a board

  1. In Boards, choose Add.
  2. Enter a vault-relative board file such as Projects/win-console.md.
  3. Enter the task document folder, for example Tasks/win-console.
  4. Choose Scan headings.
  5. Give each discovered column a stable type ID and validation profile.

Task semantics are fixed:

Obsidian object TaskDoc meaning
Board Project
H2 column Task type
Linked checkbox card Task
Checked card Completed task
Task document H2 section Checkpoint/subtask

Connect an MCP client

Use Copy client config on the settings page. The generated configuration contains the current endpoint and Bearer token. A representative OpenCode configuration is:

{
  "mcp": {
    "taskdoc": {
      "type": "remote",
      "url": "http://127.0.0.1:27124/mcp",
      "enabled": true,
      "oauth": false,
      "headers": {
        "Authorization": "Bearer <generated-token>"
      }
    }
  }
}

The token is stored in Obsidian SecretStorage. Regenerating it invalidates all copied client configurations.

MCP tools

Tool Purpose
task_catalog List configured boards and mapped type columns.
task_query Query linked tasks by project, type, or state.
task_resume Return a bounded resume packet without rich block bodies.
task_read Read an outline, checkpoint, or paged rich block.
task_create Create a task document and linked Kanban card.
task_card_update Move a task to another type column or reopen it as active.
task_checkpoint_commit Create or replace a structured checkpoint core.
task_block_put Create or replace a typed rich content block.
task_handoff Replace the mutable cross-session resume capsule.
task_finalize Validate evidence, then complete or archive the task card.

Writes use stable IDs, expected revisions, and request idempotency keys. Free-form append logs, session todo lists, and arbitrary frontmatter mutation are not exposed. The alpha idempotency-result cache lasts for the current plugin runtime; after reloading Obsidian, reread the task before retrying an uncertain write.

Development

npm.cmd install
npm.cmd run check

The production bundle is written to main.js.

Known alpha limitations

  • PHA remote synchronization is not active without its API specification.
  • Direct manual edits inside plugin-managed checkpoint regions are detected as conflicts; import/reconciliation UI is intentionally conservative.
  • The plugin understands the standard Markdown shape of Kanban boards. It does not depend on private runtime APIs from a Kanban fork.
  • Automatic conversion of manually typed, unlinked cards is reserved but not enabled in this alpha; create tasks through MCP.
  • A crash between rich-block asset and manifest writes can leave a detectable mismatch; there is no automatic mutation-journal repair yet.
  • Obsidian must remain running for MCP access.
  • The request-id result cache is not yet persisted across an Obsidian/plugin restart.

See TASKFLOW_VNEXT_DESIGN.md for the complete design and acceptance criteria.

推荐服务器

Baidu Map

Baidu Map

百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。

官方
精选
JavaScript
Playwright MCP Server

Playwright MCP Server

一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。

官方
精选
TypeScript
Magic Component Platform (MCP)

Magic Component Platform (MCP)

一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。

官方
精选
本地
TypeScript
Audiense Insights MCP Server

Audiense Insights MCP Server

通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。

官方
精选
本地
TypeScript
VeyraX

VeyraX

一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。

官方
精选
本地
graphlit-mcp-server

graphlit-mcp-server

模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。

官方
精选
TypeScript
Kagi MCP Server

Kagi MCP Server

一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。

官方
精选
Python
e2b-mcp-server

e2b-mcp-server

使用 MCP 通过 e2b 运行代码。

官方
精选
Neon MCP Server

Neon MCP Server

用于与 Neon 管理 API 和数据库交互的 MCP 服务器

官方
精选
Exa MCP Server

Exa MCP Server

模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。

官方
精选