Task Query Skill

Efficiently answers questions about tasks in the Obsidian Vault via Dataview TABLE queries with file.tasks.

When to Use This Skill

Activate on questions like:

• •"What tasks are open?"
• •"What did I complete today/this week?"
• •"Show me overdue tasks"
• •"What are my tasks about [topic]?"
• •"Which tasks have tag X?"
• •"What was the last completed task about Y?"
• •"Which tasks are due this week?"

Technical Background

Why TABLE instead of TASK Query?

The Obsidian REST API supports only TABLE queries, not TASK queries:

-- Does NOT work via REST API:
TASK WHERE !completed

-- Works via REST API:
TABLE file.tasks as "Tasks" WHERE length(file.tasks) > 0

Task Data Structure

Each task in file.tasks contains:

Comments and Context

Indented bullets under a task appear in children/subtasks:

- [x] Main task completed ✅ 2025-12-09
  - Comment or note about task
  - Additional context

→ The comment is available in children[].text.

Query Patterns

Basic Query (Tool Call)

mcp__obsidian__search_vault
  query: "TABLE file.tasks as \"Tasks\" FROM \"\" WHERE ..."
  queryType: "dataview"

Open Tasks (vault-wide)

TABLE file.tasks as "Tasks"
WHERE any(file.tasks, (t) => !t.completed AND t.text != "")
LIMIT 50

Open Tasks in Specific Folder

TABLE file.tasks as "Tasks"
FROM "Daily"
WHERE any(file.tasks, (t) => !t.completed)
LIMIT 30

Tasks by Topic/Keyword

TABLE file.tasks as "Tasks"
WHERE any(file.tasks, (t) => contains(lower(t.text), "topic"))

Case-insensitive search: contains(lower(t.text), "keyword")

Tasks by Tag

TABLE file.tasks as "Tasks"
WHERE any(file.tasks, (t) => contains(t.tags, "#topic"))

Completed Tasks (recently finished)

TABLE file.tasks as "Tasks"
WHERE any(file.tasks, (t) => t.completed AND t.completion)
SORT file.mtime DESC
LIMIT 20

Overdue Tasks

TABLE file.tasks as "Tasks"
WHERE any(file.tasks, (t) => !t.completed AND t.due AND t.due < date(today))

Tasks Due This Week

TABLE file.tasks as "Tasks"
WHERE any(file.tasks, (t) => !t.completed AND t.due AND t.due >= date(today) AND t.due <= date(today) + dur(7 days))

Combined Search (Keyword + Status)

TABLE file.tasks as "Tasks"
WHERE any(file.tasks, (t) => t.completed AND contains(lower(t.text), "topic"))

Workflow

Step 1: Execute Query

mcp__obsidian__search_vault(
  query: "<dataview-query>",
  queryType: "dataview"
)

Step 2: Filter Result (client-side)

The query returns all tasks of the file. Filter in response:

# Pseudo-code for client filtering
for file in results:
    relevant_tasks = [
        t for t in file.tasks
        if t.completed and "topic" in t.text.lower()
    ]

Step 3: Load Context if Needed

If the user needs more context:

1. •Use path and line from the task
2. •Load the file with mcp__obsidian__get_vault_file
3. •Show surrounding lines or the entire section

Step 4: Format Response

Short answer (task list):

Detailed answer (with context):

## Task: [Description]
- **Status:** Completed on 2025-12-09
- **Due:** 2025-12-04
- **File:** [[Areas/path/to/file.md]]
- **Comments:**
  - Configuration completed
  - Requires further setup

Example Requests and Queries

Extracting Results and Outcomes from Tasks

Convention: Results as Indented Bullets

Tasks with documented outcome follow this pattern:

- [x] Task description ✅ 2025-12-09
  - Result: Configuration completed
  - Next step: Review with team

Workflow: "What results did I achieve on completed tasks?"

Step 1: Query for completed tasks in time range

TABLE file.tasks as "Tasks"
WHERE any(file.tasks, (t) => t.completed AND t.completion >= date(today) - dur(7 days))

Step 2: Client-side filtering and extraction

For each task in result:

1. •Check completed == true and completion in time range
2. •Extract text as task description
3. •Extract children[].text as results/comments
4. •Ignore tasks without children (no documented result)

Step 3: Format response

## Completed Tasks (last 7 days) with Results

### 2025-12-09
- **API integration task**
  - Result: Configuration completed
  - Note: Requires further setup

### 2025-12-08
- **API review**
  - Result: 3 issues found, tickets created

Typical Vault Areas for Tasks

Edge Cases

Ignore Empty Tasks

Some files have empty checkboxes (- [ ] without text). Filter with:

t.text != ""

Tasks vs. Checkboxes in Lists

Dataview only recognizes real tasks (- [ ]), not normal bullets. This is correct.

Large Result Sets

For vault-wide queries, results can be large:

• •Use LIMIT in the query
• •Filter further client-side
• •Ask user for constraints (time range, folder, tag)

Limitations

• •No sorting by task fields: Dataview sorts by file metadata, not by task.completion
• •Client-side filtering needed: The WHERE clause filters files, not tasks - all tasks of the file come along
• •No Tasks plugin features: Recurring tasks, priorities (⏫) are recognized as text, not as fields