dart-query
dart-query
用于 Dart AI 任务管理的 MCP 服务器,针对批量操作和最小化上下文使用进行了优化。
dart-query 不再逐个循环处理任务(这会用中间 JSON 填满你的上下文窗口),而是使用 DartQL 选择器和服务器端批量操作,在单次调用中更新数百个任务。通常会消耗约 30K Token 的 50 个任务更新,现在仅需约 200 Token,且不会产生上下文损耗。
快速开始
1. 获取你的 Dart AI Token
访问 https://app.dartai.com/?settings=account 并复制你的 Token(以 dsa_ 开头)。
2. 配置 MCP
npx (推荐)
{
"mcpServers": {
"dart-query": {
"command": "npx",
"args": ["-y", "@standardbeagle/dart-query"],
"env": {
"DART_TOKEN": "dsa_your_token_here"
}
}
}
}SLOP-MCP (v0.10.0+)
slop register dart-query \
--command npx \
--args "-y" "@standardbeagle/dart-query" \
--env DART_TOKEN=dsa_your_token_here \
--scope user3. 验证
info({ level: "overview" })4. 示例:批量更新
// Preview first
batch_update_tasks({
selector: "dartboard = 'Engineering' AND priority = 'high'",
updates: { status: "Doing" },
dry_run: true
})
// Execute
batch_update_tasks({
selector: "dartboard = 'Engineering' AND priority = 'high'",
updates: { status: "Doing" },
dry_run: false
})Related MCP server: MCP Tasks
工具
组 | 工具 | 用途 |
发现 |
| 探索功能、工作区配置 |
任务 CRUD |
| 单个任务操作 |
查询 |
| 通过过滤器或全文搜索查找任务 |
批量 |
| 使用 DartQL 选择器进行批量操作 |
导入 |
| 从 CSV 批量创建并进行验证 |
文档 |
| 文档管理 |
有关完整的参数参考、DartQL 语法和 CSV 导入格式,请参阅 TOOLS.md。
DartQL 选择器
用于在批量操作中定位任务的 SQL-92 WHERE 子句语法:
dartboard = 'Engineering' AND priority = 'high' AND tags CONTAINS 'bug'
due_at < '2026-01-18' AND status <> 'Done'
title LIKE 'Task%' -- starts with
title LIKE '%auth%' -- contains substring运算符: =, !=, <>, >, >=, <, <=, LIKE, IN, NOT IN, BETWEEN, IS NULL, IS NOT NULL, CONTAINS
别名: INCLUDES/HAS → CONTAINS · <> → !=
LIKE 通配符: % = 任意字符,_ = 单个字符(不区分大小写)
安全性
所有 Dart AI 操作均为生产环境(无沙盒)。dart-query 提供:
所有批量操作均支持 试运行模式 (Dry-run mode) — 执行前预览
CSV 导入的 验证阶段 — 在创建任何内容之前捕获错误
批量删除需要 确认标志 (
confirm: true)可恢复删除 — 任务移至回收站,而非永久删除
许可证
MIT
Available Tools
25 toolsadd_task_commentA
Add a comment to a task for status updates, AI rationale, or notes. Supports markdown formatting.
Inputs: dart_id (required), text (required).
add_time_trackingB
Add a time tracking entry to a task. Supports started_at/finished_at or duration_minutes.
Inputs: dart_id (required), started_at (required), finished_at, duration_minutes, +1 more.
attach_urlA
Attach a file from URL to a task. URL must be publicly accessible.
Inputs: dart_id (required), url (required), filename.
batch_delete_tasksA
[Deprecated: use execute_dartql] Batch delete multiple tasks matching a DartQL selector. MOST DANGEROUS OPERATION! CRITICAL: dry_run defaults to true, confirm=true REQUIRED when dry_run=false. Tasks move to trash (recoverable).
Inputs: selector (required), dry_run, confirm, concurrency.
batch_update_tasksA
[Deprecated: use execute_dartql] Batch update multiple tasks matching a DartQL selector. Supports all task fields including relationships. CRITICAL: Always use dry_run=true first to preview changes!
Inputs: selector (required), updates (required), dry_run, concurrency.
create_docB
Create a new document with title, text (markdown supported), and optional folder
Inputs: title (required), text (required), folder.
create_taskA
Create a new task with title, description, status, priority, size, dates, dartboard, assignees, tags, and task relationships (subtasks, blockers, related tasks)
Inputs: title (required), dartboard (required), description, status, +13 more.
delete_docA
Delete a document (moves to trash - recoverable via Dart web UI)
Inputs: doc_id (required).
delete_taskA
Delete a task (moves to trash - recoverable via Dart web UI)
Inputs: dart_id (required).
execute_dartqlA
Execute DartQL UPDATE/DELETE statements. Supports template vars {field}, array literals, comments, and multi-statement (;). Dry-run by default.
Inputs: query (required), dry_run, concurrency.
get_batch_statusA
Retrieve status of a batch operation (update, delete, import, or dartql) by batch_operation_id. Operations are kept in memory for 1 hour.
Inputs: batch_operation_id (required).
get_configA
Get workspace configuration: assignees, dartboards, statuses, tags, priorities, sizes
Inputs: cache_bust, include.
get_dartboardB
Get details about a dartboard including task count. Token-efficient lookup.
Inputs: dartboard_id (required).
get_docA
Get a specific document by doc_id with full text content
Inputs: doc_id (required).
get_folderA
Get details about a folder including doc count. Token-efficient lookup.
Inputs: folder_id (required).
get_taskA
Get a specific task by dart_id with optional comments and relationship details. Returns task relationships (subtasks, blockers, blocking, duplicates, related) with counts and optional expanded titles.
Inputs: dart_id (required), include_comments, include_relationships, expand_relationships.
import_tasks_csvA
Import tasks from CSV file with validation and parallel creation. CRITICAL: ALWAYS use validate_only=true first! Production safety: validate → fix errors → import.
Inputs: dartboard (required), csv_data, csv_file_path, column_mapping, +3 more.
infoB
Progressive discovery of dart-query capabilities - start here
Inputs: level, target.
list_commentsA
List comments on a task with pagination. Token-efficient: returns minimal comment data.
Inputs: task_id (required), limit, offset.
list_docsA
List documents with optional filtering by folder, title_contains, text_contains. Supports pagination.
Inputs: folder, title_contains, text_contains, limit, +1 more.
list_tasksB
Query tasks with filters (assignee, status, dartboard, priority, tags, dates, has_parent), pagination, and detail levels. Parent filter uses client-side filtering.
Inputs: assignee, status, dartboard, priority, +7 more.
move_taskA
Move/reposition a task within a dartboard or to a different dartboard. Supports ordering by index or relative to another task.
Inputs: dart_id (required), dartboard, order, after_id, +1 more.
search_tasksA
Full-text search across tasks with relevance ranking. Alternative to list_tasks for text-based discovery. Supports quoted phrases, exclusions (-term), and inline filters (dartboard:Name, status:Done, assignee:Name).
Inputs: query (required), dartboard, include_completed, limit.
update_docA
Update an existing document (title, text, or folder). Validates references and only sends changed fields.
Inputs: doc_id (required), updates (required).
update_taskA
Update a task. Pass dart_id and any fields to change at the top level. Supports add_to/remove_from for incremental relationship changes and optional comment. Relationship arrays use full replacement unless using add_to/remove_from.
Inputs: dart_id (required), title, description, dartboard, +16 more.
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
TDQS
Scored across 25 tools
Most tools target distinct resources (tasks, docs, folders, config), but there is overlap between list_tasks and search_tasks (both find tasks, though one is filter-based and the other text-based). Also, batch_delete_tasks and batch_update_tasks are deprecated alternatives to execute_dartql, creating potential confusion about which to use. The descriptions do help clarify, but the presence of deprecated tools adds ambiguity.
Tool names follow a mostly consistent verb_noun pattern (get_task, update_task, create_task, delete_task, list_docs, etc.). Some deviations exist like 'execute_dartql' (non-noun object) and 'import_tasks_csv' (instead of import_tasks_csv_file or similar), but these are minor and the pattern is still readable.
With 25 tools, this is on the heavier side for a task management server. Many tools are focused on task lifecycle (create, get, update, delete, move, comment, time tracking), but some like batch operations and import seem redundant with execute_dartql. A few tools could be merged (e.g., batch_update and batch_delete into execute_dartql), but the count is not excessive given the domain complexity.
The server covers core CRUD for tasks (create, read, update, delete) and documents (create, read, update, delete), plus additional features like comments, time tracking, and URL attachments. However, there is no direct way to list or delete comments, and while tasks support relationships, there is no explicit tool to manage dartboard or folder creation/deletion (only get). The presence of deprecated tools suggests some operations are being consolidated, but the surface is fairly complete for task management with some gaps.
Maintenance
Related MCP Connectors
Headless task management for bots, apps and AI agents. Remote MCP server, OAuth or access token.
Nifty's MCP server — exposes tasks, projects, messages, and files as tools for AI agents.
Task management your AI agents can actually run. One line becomes a context-ready task over MCP.
MCP server for your apps' tools and custom tools, plus hosted AI agents and approval-gated workflows
Related MCP Servers
- AlicenseBqualityDmaintenanceA Dart-based MCP server implementation that enables AI-assisted task management, document handling, and workspace organization through standardized tools and seamless Dart integration.105MIT
- AlicenseAqualityDmaintenanceA comprehensive and efficient Model Context Protocol server for task management that works with Claude, Cursor, and other MCP clients, providing powerful search, filtering, and organization capabilities across multiple file formats.5228 npm47MIT
- FlicenseBqualityDmaintenanceA comprehensive task management MCP server built with FastMCP, featuring full CRUD operations, intelligent filtering, and productivity-focused prompts.71-
- FlicenseNot gradedqualityDmaintenanceA minimal MCP server for file-based task management and multi-agent team orchestration, enabling creation, completion, and reassignment of tasks across teams.-