| name | todoist |
| description | Todoist dashboard and router. Shows inbox count, overdue count, project health, and routes to the right sub-skill. Use when the user says 'todoist', 'check my tasks', 'task overview', 'what's on my plate', 'show me my todos', or any general Todoist inquiry that doesn't clearly map to triage/organize/update/extract. |
Todoist Dashboard
Entry point for the todoist-flow suite. Show the user where things stand, then route them to the right tool.
Flow
Step 1: Fetch State
Spawn todoist-fetch to gather dashboard data in a fresh context window:
Agent(subagent_type: "todoist-fetch", prompt: "Fetch dashboard data: 1) get-overview for inbox count and project summary, 2) find-tasks with filter 'overdue' for overdue count — flag any 3+ weeks stale, 3) get-productivity-stats for completion trends. Return structured overview with counts.")
Step 2: Present Dashboard
Format as a concise status board:
## Todoist Status
| Metric | Count |
|--------|-------|
| Inbox | X tasks |
| Overdue | X tasks (Y are 3+ weeks stale) |
| Due today | X tasks |
| Active projects | X |
### Health Signals
- [good/warn/critical] Overdue debt: X tasks
- [good/warn/critical] Inbox backlog: X tasks
- [good/warn/critical] P1 inflation: X tasks at P1
Step 3: Route
Based on the state, suggest the most impactful next action:
| Signal | Recommendation |
|---|
| Overdue > 10 | "Run /triage to process overdue tasks" |
| Inbox > 20 | "Run /triage to clear inbox backlog" |
| P1 count > 5 | "Run /update to review priority inflation" |
| Projects > 15 or empty projects exist | "Run /organize to restructure projects" |
| Many noun-only or URL tasks | "Run /extract to pull out reference material" |
| Everything looks healthy | "Looking good. Run /update for routine hygiene." |
Present the suggestion but let the user decide. Don't auto-invoke sub-skills.
Research Integration
If the user asks for deeper context on their productivity patterns, invoke the research skill:
Skill(skill="todoist-flow:research", args="[specific topic the user asked about]")
This is opt-in — only when the user explicitly asks for research.
