TaskFlow MCP πβ
@pinkpixel-dev
About TaskFlow MCP πβ
A task management Model Context Protocol (MCP) server that helps AI assistants break down user requests into manageable tasks with subtasks, dependencies, and notes. Enforces a structured workflow with user approval steps.
Config
Add this server to your MCP-compatible client using the configuration below.
{
"mcpServers": {
"taskflow": {
"command": "npx",
"args": [
"-y",
"@pinkpixel/taskflow-mcp"
],
"env": {
"TASK_MANAGER_FILE_PATH": "/path/to/tasks.json"
}
}
}
}Tools
24Register a new user request and plan its associated tasks. You must provide 'originalRequest' and 'tasks', and optionally 'splitDetails'. Tasks can now include subtasks, which are smaller units of work that make up a task. All subtasks must be completed before a task can be marked as done. You can also include: - 'dependencies': List of project or task-specific dependencies (libraries, tools, etc.) - 'notes': General notes about the project (preferences, guidelines, etc.) - 'outputPath': Path to save a Markdown file with the task plan for reference. It's recommended to use absolute paths (e.g., 'C:/Users/username/Documents/task-plan.md') rather than relative paths for more reliable file creation. This tool initiates a new workflow for handling a user's request. The workflow is as follows: 1. Use 'plan_task' to register a request and its tasks (with optional subtasks, dependencies, and notes). 2. After adding tasks, you MUST use 'get_next_task' to retrieve the first task. A progress table will be displayed. 3. Use 'get_next_task' to retrieve the next uncompleted task. 4. If the task has subtasks, complete each subtask using 'mark_subtask_done' before marking the task as done. 5. **IMPORTANT:** After marking a task as done, a progress table will be displayed showing the updated status of all tasks. The assistant MUST NOT proceed to another task without the user's approval. Ask the user for approval before proceeding. 6. Once the user approves the completed task, you can proceed to 'get_next_task' again to fetch the next pending task. 7. Repeat this cycle until all tasks are done. 8. After all tasks are completed, 'get_next_task' will indicate that all tasks are done. At this point, ask the user for confirmation that the entire request has been completed satisfactorily. 9. If the user wants more tasks, you can use 'add_tasks_to_request' or 'plan_task' to add new tasks and continue the cycle. The critical point is to always wait for user approval after completing each task and after all tasks are done. Do not proceed automatically, UNLESS the user has explicitly told you to continue with all tasks and that you do not need approval.
Given a 'requestId', return the next pending task (not done yet). If all tasks are completed, it will indicate that no more tasks are left and that you must ask the user what to do next. A progress table showing the current status of all tasks will be displayed with each response. If the same task is returned again or if no new task is provided after a task was marked as done, you MUST NOT proceed. In such a scenario, you must prompt the user for approval before calling 'get_next_task' again. Do not skip the user's approval step. In other words: - After calling 'mark_task_done', do not call 'get_next_task' again until the user has given approval for the completed task. - If 'get_next_task' returns 'all_tasks_done', it means all tasks have been completed. At this point, confirm with the user that all tasks have been completed, and optionally add more tasks via 'add_tasks_to_request' or 'plan_task'.
Mark a given task as done after you've completed it. Provide 'requestId' and 'taskId', and optionally 'completedDetails'. After marking a task as done, a progress table will be displayed showing the updated status of all tasks. After this, DO NOT proceed to 'get_next_task' again until the user has explicitly approved the completed task. Ask the user for approval before continuing.
Get details of a specific task by 'taskId'. This is for inspecting task information at any point.
List all requests with their basic information and summary of tasks. This provides a quick overview of all requests in the system.
Add new tasks to an existing request. This allows extending a request with additional tasks. Tasks can include subtasks and dependencies. A progress table will be displayed showing all tasks including the newly added ones.
Update an existing task's title and/or description. Only uncompleted tasks can be updated. A progress table will be displayed showing the updated task information.
Delete a specific task from a request. Only uncompleted tasks can be deleted. A progress table will be displayed showing the remaining tasks after deletion.
Add subtasks to an existing task. Provide 'requestId', 'taskId', and 'subtasks' array. Subtasks are smaller units of work that make up a task. All subtasks must be completed before a task can be marked as done. A progress table will be displayed showing the updated task with its subtasks.
Mark a subtask as done. Provide 'requestId', 'taskId', and 'subtaskId'. A progress table will be displayed showing the updated status of all tasks and subtasks. All subtasks must be completed before a task can be marked as done.
Update a subtask's title or description. Provide 'requestId', 'taskId', 'subtaskId', and optionally 'title' and/or 'description'. Only uncompleted subtasks can be updated. A progress table will be displayed showing the updated task with its subtasks.
Delete a subtask from a task. Provide 'requestId', 'taskId', and 'subtaskId'. Only uncompleted subtasks can be deleted. A progress table will be displayed showing the updated task with its remaining subtasks.
Export the current status of all tasks in a request to a file. This tool saves the current state of tasks, subtasks, dependencies, and notes to a file for reference. You can specify: - 'format': 'markdown', 'json', or 'html' - 'outputPath': Full path to save the file, or just a directory path - 'filename': Optional custom filename (auto-generated if not provided) Path handling: - If outputPath is a directory, filename will be auto-generated as '{project-name}_tasks.{ext}' - If outputPath includes filename, it will be used as-is - Relative paths are resolved from current working directory - If no path specified, saves to current working directory
Add a note to a request. Notes can contain important information about the project, such as user preferences or guidelines. Notes are displayed in the task progress table and can be referenced when working on tasks.
Update an existing note's title or content. Provide the 'requestId' and 'noteId', and optionally 'title' and/or 'content' to update.
Delete a note from a request. Provide the 'requestId' and 'noteId' of the note to delete.
Add a dependency to a request or task. Dependencies can be libraries, tools, or other requirements needed for the project or specific tasks. If 'taskId' is provided, the dependency will be added to that specific task. Otherwise, it will be added to the request.
Get the current prompts configuration including instructions, taskPrefix, and taskSuffix settings. This tool helps you view the current global prompts settings that are applied to all tasks.
Set the global prompts configuration with instructions, taskPrefix, and/or taskSuffix. This replaces any existing prompts settings with the new values provided. - 'instructions': General instructions or context shown at the top of each task - 'taskPrefix': Text to prepend before each task description - 'taskSuffix': Text to append after each task description
Update specific parts of the prompts configuration without replacing the entire object. Use this to modify individual fields (instructions, taskPrefix, or taskSuffix) while keeping other settings unchanged.
Remove the entire prompts configuration or specific fields from it. If 'fields' is provided, only those specific fields will be removed. If 'fields' is not provided, the entire prompts configuration will be removed.
Archive completed requests to a separate file to keep the active tasks file clean. If 'requestIds' is provided, only those specific completed requests will be archived. If 'requestIds' is not provided, all completed requests will be archived. This addresses the need to keep active tasks file uncluttered by moving completed work to archive storage.
List archived requests with optional search and filtering capabilities. Provides an overview of all archived requests with their metadata and completion information.
Restore a specific archived request back to the active tasks file. This moves the request from the archive back to active status, allowing you to continue working on it.
Overview
What is TaskFlow MCP?
TaskFlow MCP is a task management Model Context Protocol (MCP) server that helps AI assistants break down user requests into manageable tasks and track their completion. It enforces a structured workflow with user approval steps to ensure tasks are properly tracked and users maintain control over the process.
How to use TaskFlow MCP?
Install it globally with npm install -g @pinkpixel/taskflow-mcp and start with taskflow-mcp, or run locally with npx @pinkpixel/taskflow-mcp. Configure the server by setting the TASK_MANAGER_FILE_PATH environment variable and add it to your MCP clientβs mcp_config.json. AI assistants interact with TaskFlow MCP through tools like plan_task, get_next_task, and mark_task_done.
Key features of TaskFlow MCP?
- Break down complex requests into tasks and subtasks
- Enforce user approval steps for quality control
- Persist tasks to disk in JSON or YAML format
- Export task status reports in Markdown, JSON, or HTML
- Track project and task-level dependencies with version information
- Archive completed requests and restore full history when needed
Use cases of TaskFlow MCP?
- AI assistants managing multi-step project plans with user confirmation
- Breaking a website creation request into design, implementation, and testing tasks
- Tracking progress of complex requests with subtasks and dependencies
- Archiving finished work to keep active task lists clean while preserving history
- Exporting task status reports for sharing with team members
FAQ from TaskFlow MCP
What file formats does TaskFlow MCP support for persistence?
TaskFlow MCP supports both JSON and YAML formats, automatically detected by the file extension (.json, .yaml, .yml). YAML is recommended for better handling of multiline content and human readability.
How do I set the tasks file path?
Use the TASK_MANAGER_FILE_PATH environment variable. You can specify absolute paths (e.g., /home/user/tasks.yaml) or relative paths (e.g., ./project-tasks.yaml). A custom base directory can be set with TASK_MANAGER_BASE_DIR.
Does TaskFlow MCP require any runtime dependencies?
It runs on Node.js and is installed via npm. No additional runtime services or databases are requiredβall data is persisted to a local file.
How does the archive system work?
Archives can be manual (using the archive_completed_requests tool) or automatic (set ARCHIVE_MODE=auto-on-complete). Archived requests can be browsed, searched, and restored using dedicated tools like list_archived_requests and restore_archived_request.
What transport does TaskFlow MCP use?
TaskFlow MCP is a standard MCP server that communicates over stdio, as configured in the MCP clientβs mcp_config.json via the command and args fields.
Frequently asked questions
What file formats does TaskFlow MCP support for persistence?
TaskFlow MCP supports both JSON and YAML formats, automatically detected by the file extension (`.json`, `.yaml`, `.yml`). YAML is recommended for better handling of multiline content and human readability.
How do I set the tasks file path?
Use the `TASK_MANAGER_FILE_PATH` environment variable. You can specify absolute paths (e.g., `/home/user/tasks.yaml`) or relative paths (e.g., `./project-tasks.yaml`). A custom base directory can be set with `TASK_MANAGER_BASE_DIR`.
Does TaskFlow MCP require any runtime dependencies?
It runs on Node.js and is installed via npm. No additional runtime services or databases are requiredβall data is persisted to a local file.
How does the archive system work?
Archives can be manual (using the `archive_completed_requests` tool) or automatic (set `ARCHIVE_MODE=auto-on-complete`). Archived requests can be browsed, searched, and restored using dedicated tools like `list_archived_requests` and `restore_archived_request`.
What transport does TaskFlow MCP use?
TaskFlow MCP is a standard MCP server that communicates over stdio, as configured in the MCP clientβs `mcp_config.json` via the `command` and `args` fields.
Basic information
More Other MCP servers
Mcp
browsermcpBrowser MCP is a Model Context Provider (MCP) server that allows AI applications to control your browser

EverArt
modelcontextprotocolModel Context Protocol Servers
FastMCP v2 π
jlowinπ The fast, Pythonic way to build MCP servers and clients.
MaxKB
1Panel-devπ₯ MaxKB is an open-source platform for building enterprise-grade agents. εΌΊε€§ζη¨ηεΌζΊδΌδΈηΊ§ζΊθ½δ½εΉ³ε°γ

Glasswarp
GlasswarpSee and control a real Windows PC you own β from any MCP client, locally or remotely. Observe (UIA + screenshots), click/type/drag/scroll, launch apps, owner Live View. BYOH: your machine, your key.
Comments