DevTask Mcp
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| DEVTASK_API_KEY | Yes | API key for devtask API | |
| DEVTASK_API_BASE | Yes | Base URL for devtask API |
Instructions
Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.
This server publishes no instructions, or was last inspected before Glama recorded them.
Capabilities
Features and capabilities supported by this server
Protocol revision2025-11-25
| Capability | Details |
|---|---|
| tools | {
"listChanged": false
} |
| prompts | {
"listChanged": false
} |
| resources | {
"subscribe": false,
"listChanged": false
} |
| experimental | {} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| list_dev_tasksA | List dev-tasks from your kanocifer-chat board. Args: status: Filter by lifecycle status. One of: '待评估', '待排期', '进行中', '已搁置', '已完成'. priority: Filter by urgency. One of: 'P0 紧急', 'P1 高', 'P2 中', 'P3 低'. task_type: Filter by kind. One of: '问题', '功能需求', '优化', '技术债'. for_agent: When True, return only tasks ready for agent execution. When False, return only tasks for human. None = no filter. include_deleted: When True, soft-deleted tasks are also returned. page: Page number (1-based). per_page: Items per page, capped at 20. |
| get_dev_taskA | Fetch a single dev-task by its ObjectID. The response includes spec fields (acceptance_criteria, constraints, context_pointers), dependency info (for_agent, blocked_by), and slug. Read them before starting work. context_pointers short-circuits file discovery by telling you exactly which paths are relevant. Use the slug for all human-facing references (e.g. "work on task-42"). Args: task_id: The task's ObjectId hex string (24 hex chars). Use get_dev_task_by_slug if you have a slug like "task-42". |
| get_dev_task_by_slugA | Fetch a single dev-task by its slug (task-1, task-2...). Slug 是任务的人类可读短标识,比 ObjectID 更适合口头引用、看板 UI 和 MCP 对话。其余字段同 get_dev_task。 Args: slug: The task slug, e.g. "task-42". |
| create_dev_taskA | Create a new dev-task. New tasks start at status '待评估'. 返回体里包含 slug (如 "task-1"),后续对话 / 看板 UI / MCP tool 引用 该任务时直接用 slug,比 ObjectID 可读得多。 Args: title: Task title (required). task_type: One of '问题', '功能需求', '优化', '技术债'. priority: One of 'P0 紧急', 'P1 高', 'P2 中', 'P3 低'. scope: Task scope in "<层>-<技术>" format (required). Examples: '前端-React', '后端-Go', 'AI-LangChain', 'Docs-用户手册', '通用'. Free-form string — not a closed enum. description: Short description (optional). detail: Long-form detail (optional). due_date: ISO-8601 datetime string, e.g. '2026-09-01T00:00:00'. Pass whatever your JSON client gives; the backend parses RFC-3339. acceptance_criteria: Conditions that must be met for this task to be considered done. Agent uses these as a self-check before resolve. constraints: Hard boundaries — files not to touch, tech stack requirements, etc. Agent treats these as non-negotiable. context_pointers: Paths to relevant code / docs / ADRs, e.g. 'internal/auth/, docs/adr/0003'. Saves agent a file-discovery pass. for_agent: When True, mark this task as ready for agent execution. Default False (human task). blocked_by: List of task IDs that must be done before this one. Pass [] when no dependencies. |
| update_dev_taskA | Partially update a dev-task. Omitted fields are left unchanged. Args: task_id: The task's ObjectId hex string or slug (task-N). title: New title (optional). description: New description (optional). detail: New detail (optional). task_type: One of '问题', '功能需求', '优化', '技术债'. priority: One of 'P0 紧急', 'P1 高', 'P2 中', 'P3 低'. scope: Free-form "<层>-<技术>" string (see create_dev_task docs). status: One of '待评估', '待排期', '进行中', '已搁置', '已完成'. sort_order: Integer sort key (optional). due_date: ISO-8601 datetime string. acceptance_criteria: Conditions for considering this task done. constraints: Hard boundaries (files / tech stack / etc). context_pointers: Paths to relevant code / docs / ADRs. for_agent: Toggle agent-claimable flag. blocked_by: Replace the dependency list entirely. Pass [] to clear. |
| get_frontier_tasksA | Return tasks the agent can claim next — Pocock's frontier. Frontier = tasks that are:
Ordered by sort_order ASC, then created_at DESC. Use this as your first call when the user says "do the next task" — it tells you exactly what's ready to be worked. Args: limit: Max tasks to return (default 10). |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
No prompts | |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
No resources | |
TDQS
Scored across 6 tools
All six tools have clearly distinct purposes: create, get by ObjectID, get by slug, get frontier tasks, list with filters, and update. No functional overlap.
All tool names follow a consistent verb_noun snake_case pattern (e.g., create_dev_task, get_dev_task_by_slug, list_dev_tasks), making it easy to predict behavior.
Six tools strike a good balance, covering core CRUD plus a specialized frontier query without being too many or too few for a task management MCP server.
The tools cover creation, reading (multiple variants), listing, and updating. The only notable gap is the absence of a dedicated delete tool, though soft-delete is supported via list filtering.