AgentSkillsCN

todoist-tasks

根据自然语言创建Todoist任务。当用户提及某项任务、待办事项、提醒,或他们需要完成的某件事时使用。可通过“我需要……”“提醒我……”“添加任务”等语句,或任何用户希望追踪的可执行事项来触发任务创建。

SKILL.md
--- frontmatter
name: todoist-tasks
description: Create Todoist tasks from natural language. Use when the user mentions a task, to-do, reminder, or something they need to do. Triggers on phrases like "I need to", "remind me to", "add task", or any actionable item the user wants to track.

Todoist Task Creation

Create tasks in Todoist from natural language input using the MCP Todoist tools.

User's Todoist Structure

All tasks go to Inbox — no separate projects.

Labels for categorization:

  • professional — PhD hunting, job hunting, networking with faculty/professors, LinkedIn, career counseling, industry networking
  • social — social activities, friends, dating, community

Active vs Backlog (determined by due date):

  • Due today = active
  • Due on other days or no due date = backlog

Defaults

FieldDefaultOverride when...
ProjectInboxAlways Inbox
LabelNoneContent matches professional or social domain
Priorityp4 (lowest)User says "high priority", "important", "urgent"
Due dateNone (backlog)User specifies "today", "tomorrow", a date, etc.

Label Auto-Detection

Infer label from task content:

Keywords/ContextLabel
faculty, professor, PhD, grad school, research, academia, department, advisor, LinkedIn, recruiter, job, career, industry, interview, resume, networking (professional)professional
friends, hangout, party, dinner, drinks, social event, birthday, dating, Hinge, communitysocial
Everything elseno label

Workflow

  1. Parse the user's natural language for: task content, label hints, priority, due date
  2. Apply defaults for any unspecified fields
  3. Call mcp__todoist__add-tasks with the parsed values
  4. Confirm creation with task name and key details

MCP Tool Reference

code
mcp__todoist__add-tasks
Parameters:
- tasks: array of task objects
  - content: string (required) — task title
  - description: string — additional details
  - labels: array of strings — e.g., ["professional"] or ["social"]
  - priority: "p1" | "p2" | "p3" | "p4" — p1 highest, p4 lowest
  - dueString: string — natural language date like "tomorrow", "next Friday"
  - duration: string — e.g., "2h", "30m"

Examples

User: "I need to email Professor Smith about research opportunities" → Inbox, Label: professional, Priority: p4, No due date

User: "Update my LinkedIn profile today" → Inbox, Label: professional, Priority: p4, Due: today

User: "Get groceries tomorrow" → Inbox, No label, Priority: p4, Due: tomorrow

User: "High priority: prepare for interview on Friday" → Inbox, Label: professional, Priority: p1, Due: Friday

User: "Grab coffee with Sarah this weekend" → Inbox, Label: social, Priority: p4, Due: this weekend