todoist-triage

name: todoist-triage description: Triage and clean up Todoist with duplicate detection, stale task cleanup, and ADHD-friendly coaching. Use when the user asks to triage, organize, or clean up their Todoist inbox or tasks.

Todoist Triage

Interactive triage and cleanup for your Todoist task list, designed with Executive Function (EF) and ADHD support in mind. Syncs data to a local SQLite database for efficient analysis, detects issues, suggests improvements, and coaches you toward better task management habits.

Philosophy: EF and ADHD-Friendly Design

This skill acts as an Executive Function coach, not just a task cleaner. It's built around principles that help people who struggle with:

• Task initiation: Identifies "quick wins" (2-5 min tasks) to build momentum
• Task sizing: Breaks down overwhelming tasks into concrete next actions
• Clarity: Ensures tasks are SMART (Specific, Measurable, Attainable, Relevant, Timebound)
• Decision fatigue: Batches similar decisions, offers smart defaults
• Motivation: Celebrates progress, suggests dopamine-friendly task ordering
• Overwhelm: Never shows everything at once; works in focused chunks

Core Principles:

1. One thing at a time - Each question focuses on one decision
2. Momentum first - Start sessions by identifying quick wins
3. Break it down - Large vague tasks become concrete action steps
4. Raise the bar gently - Suggest improvements without judgment
5. Learn patterns - Remember what works for you specifically

Trigger Phrases

• /todoist-triage
• "triage my todoist"
• "clean up my todoist"
• "find duplicate tasks in todoist"
• "organize my todoist inbox"
• "help me get unstuck" (EF-focused mode)
• "what can I do right now" (quick wins mode)

Prerequisites

• Todoist MCP: Must be configured for write operations
• 1Password CLI: API token stored in item "Todoist", field "API Token"
• Python 3: With thefuzz library for fuzzy matching (uv pip install thefuzz)

Overview

This skill operates in 5 phases:

1. SYNC - Fetch data from Todoist Sync API into local SQLite
2. ANALYZE - Detect issues: duplicates, stale tasks, sizing problems, SMART gaps
3. TRIAGE - Interactive decision-making with EF-friendly coaching
4. CONFIRM - Review summary of all planned changes
5. APPLY - Execute changes via MCP tools, update memory

File Locations

skills/todoist-triage/
├── SKILL.md                   # This file
├── assets/
│   └── TRIAGE_MEMORY.md       # Template for user preferences
└── data/                      # Runtime data (gitignored)
    ├── todoist.db             # SQLite database
    └── sync_state.json        # Sync token and timestamps

Phase 1: SYNC

Step 1.1: Initialize Database

Create the SQLite database if it doesn't exist:

DB_PATH="skills/todoist-triage/data/todoist.db"
mkdir -p "skills/todoist-triage/data"

sqlite3 "$DB_PATH" << 'EOF'
CREATE TABLE IF NOT EXISTS sync_state (
    id INTEGER PRIMARY KEY CHECK (id = 1),
    sync_token TEXT NOT NULL DEFAULT '*',
    last_full_sync DATETIME,
    last_incremental_sync DATETIME
);
INSERT OR IGNORE INTO sync_state (id, sync_token) VALUES (1, '*');

CREATE TABLE IF NOT EXISTS projects (
    id TEXT PRIMARY KEY,
    name TEXT NOT NULL,
    parent_id TEXT,
    is_inbox BOOLEAN DEFAULT 0,
    is_archived BOOLEAN DEFAULT 0,
    synced_at DATETIME NOT NULL DEFAULT CURRENT_TIMESTAMP
);

CREATE TABLE IF NOT EXISTS items (
    id TEXT PRIMARY KEY,
    content TEXT NOT NULL,
    description TEXT,
    project_id TEXT,
    section_id TEXT,
    parent_id TEXT,
    labels TEXT,
    priority INTEGER DEFAULT 1,
    due_date TEXT,
    is_completed BOOLEAN DEFAULT 0,
    added_at DATETIME,
    content_normalized TEXT,
    content_hash TEXT,
    synced_at DATETIME NOT NULL DEFAULT CURRENT_TIMESTAMP
);

CREATE TABLE IF NOT EXISTS triage_findings (
    id INTEGER PRIMARY KEY AUTOINCREMENT,
    session_id TEXT NOT NULL,
    finding_type TEXT NOT NULL,
    item_id TEXT,
    related_item_id TEXT,
    similarity_score REAL,
    details TEXT,
    suggested_action TEXT,
    user_decision TEXT,
    decided_at DATETIME,
    applied_at DATETIME
);

CREATE INDEX IF NOT EXISTS idx_items_content_hash ON items(content_hash);
CREATE INDEX IF NOT EXISTS idx_items_added ON items(added_at);
CREATE INDEX IF NOT EXISTS idx_findings_session ON triage_findings(session_id);
EOF

Step 1.2: Fetch Todoist API Token

TODOIST_TOKEN=$(op item get "Todoist" --fields "API Token" --reveal)

Step 1.3: Call Todoist Sync API

SYNC_TOKEN=$(sqlite3 "$DB_PATH" "SELECT sync_token FROM sync_state WHERE id = 1;")

RESPONSE=$(curl -s -X POST "https://api.todoist.com/sync/v9/sync" \
  -H "Authorization: Bearer $TODOIST_TOKEN" \
  -d "sync_token=$SYNC_TOKEN" \
  -d 'resource_types=["all"]')

echo "$RESPONSE" > "skills/todoist-triage/data/sync_response.json"

Phase 2: ANALYZE

Detect Exact Duplicates

SELECT i1.id, i2.id, i1.content
FROM items i1
JOIN items i2 ON i1.content_hash = i2.content_hash AND i1.id < i2.id
WHERE i1.is_completed = 0 AND i2.is_completed = 0;

Detect Fuzzy Duplicates

Use Python with thefuzz for similarity detection (threshold: 80%):

from thefuzz import fuzz

for t1 in tasks:
    for t2 in tasks:
        score = fuzz.ratio(t1['content'].lower(), t2['content'].lower())
        if score >= 80:
            # Found fuzzy duplicate

Detect Stale Tasks

• No due date + age > 60 days
• Overdue > 14 days
• In Inbox > 7 days

SELECT * FROM items
WHERE is_completed = 0
  AND due_date IS NULL
  AND julianday('now') - julianday(added_at) > 60;

Task Sizing Analysis

Identify tasks that need breakdown:

• Vague verbs: plan, organize, figure out, deal with
• Multiple "and"s in task name
• Very long task names (>12 words)

Quick wins:

• Specific verbs: call, email, buy, schedule
• Short task names (<6 words)

Phase 3: TRIAGE (Interactive)

Present Summary

AskUserQuestion(
  questions=[{
    "question": "Analysis complete! Found:\n* Q quick wins\n* X duplicates\n* Z stale tasks\n\nWhat to tackle?",
    "header": "Summary",
    "options": [
      {"label": "Quick wins", "description": "Build momentum"},
      {"label": "Duplicates", "description": "Clean up duplicates"},
      {"label": "Stale tasks", "description": "Review old tasks"},
      {"label": "Full triage", "description": "All categories"}
    ]
  }]
)

Handle Duplicates

AskUserQuestion(
  questions=[{
    "question": "Found X tasks with identical content. How to handle?",
    "header": "Duplicates",
    "options": [
      {"label": "Keep newest", "description": "Delete older duplicates"},
      {"label": "Keep oldest", "description": "Delete newer duplicates"},
      {"label": "Review each", "description": "Manual decision"}
    ]
  }]
)

Quick Wins (Momentum Building)

AskUserQuestion(
  questions=[{
    "question": "Here are 5 tasks you could finish in 15 minutes:\n1. Call dentist\n2. Reply to email\n3. Order supplies\n\nKnock some out now?",
    "header": "Quick Wins",
    "options": [
      {"label": "Show all", "description": "See full list"},
      {"label": "Batch by type", "description": "Group calls, emails"},
      {"label": "Start with #1", "description": "Let's do it"},
      {"label": "Skip", "description": "Move to other triage"}
    ]
  }]
)

Large Task Breakdown

AskUserQuestion(
  questions=[{
    "question": "'Plan vacation to Japan'\n\nThis looks like a project. What's the VERY NEXT physical action?",
    "header": "Break Down",
    "options": [
      {"label": "Research flights", "description": "Search dates/prices"},
      {"label": "Pick dates", "description": "Check calendar"},
      {"label": "Ask travel buddy", "description": "Coordinate"},
      {"label": "Help me think", "description": "Walk through it"}
    ]
  }]
)

Phase 4: CONFIRM

AskUserQuestion(
  questions=[{
    "question": "Ready to apply:\n* Delete: 8 duplicates\n* Complete: 3 stale\n* Update: 6 with due dates\n\nProceed?",
    "header": "Confirm",
    "options": [
      {"label": "Apply all", "description": "Execute changes"},
      {"label": "Dry run", "description": "Show what would happen"},
      {"label": "Review again", "description": "Go back"},
      {"label": "Cancel", "description": "Discard all"}
    ]
  }]
)

Phase 5: APPLY

Execute via MCP

# Complete tasks
mcp__todoist__complete-tasks(ids: ["id1", "id2", ...])

# Update tasks
mcp__todoist__update-tasks(tasks: [
  {id: "x", dueString: "tomorrow"},
  {id: "y", projectId: "new_project"}
])

# Create projects
mcp__todoist__add-projects(projects: [{name: "New Project"}])

Update TRIAGE_MEMORY.md

Record session results and learned patterns:

### Session 20260113_143022

**Changes applied:**
- Deleted 8 duplicates
- Completed 3 stale tasks
- Updated 6 tasks with due dates

**Patterns learned:**
- Tasks mentioning 'Sarah' -> Work project
- Tasks starting with 'Call' -> Add @phone label

Configuration (TRIAGE_MEMORY.md)

Copy assets/TRIAGE_MEMORY.md to data/TRIAGE_MEMORY.md and customize:

stale_thresholds:
  no_due_date_days: 60
  overdue_days: 14
  inbox_days: 7

excluded_projects:
  - "Someday/Maybe"
  - "Reference"

duplicate_default: "keep_newest"
similarity_threshold: 80

Error Handling

Tools Used

• Bash: Database init, API calls
• Python: Fuzzy matching, analysis
• SQLite: Local data storage
• AskUserQuestion: Interactive decisions
• mcp__todoist__*: Task operations