asana

name	asana
description	Show Asana tasks due in the next 7 days across both workspaces and interactively mark tasks complete.

/asana

Quick view of upcoming Asana tasks with interactive completion. Pure chat skill — no file writes.

Inputs

• $1 (optional): number of days to look ahead. Defaults to 7.

Procedure

1. Fetch upcoming tasks

Issue both calls in a single tool-use block so they run in parallel. Both workspaces use asana_get_my_tasks — this returns only tasks assigned to the user (not collaborator/follower tasks, which search_tasks incorrectly includes).

• Work (mcp__asana_<work_slug>__asana_get_my_tasks, workspace <asana_work_workspace_gid>):
  • completed_since: now (returns incomplete tasks only)
  • opt_fields: name,due_on,due_at,projects.name,permalink_url
  • Client-side filter: keep tasks where due_on is non-null AND due_on <= today + $1 days. Sort ascending by due_on.
• Personal (mcp__asana_<personal_slug>__asana_get_my_tasks, workspace <asana_personal_workspace_gid>):
  • completed_since: now (returns incomplete tasks only)
  • opt_fields: name,due_on,due_at,projects.name,permalink_url
  • Client-side filter: keep tasks where due_on is non-null AND due_on <= today + $1 days. Sort ascending by due_on.

Why get_my_tasks over search_tasks: search_tasks with assignee_any: me returns tasks where the user is a collaborator or follower — not just the assignee. This floods the list with downstream pipeline and team-watched tasks that aren't actually theirs. get_my_tasks returns only truly assigned tasks. Free-tier workspaces also require get_my_tasks because search_tasks is paid-plan-only (402 Payment Required otherwise).

Overdue tasks (due_on < today, not completed) fall naturally inside the due_on <= today + $1 days window and will appear in both result sets without a separate query.

2. Display the list

Format as a single unified table with Personal on the left and Work on the right. Date group headers are rendered as bold-text rows spanning the table (use the Personal column for the label, leave other cells empty). Using one table guarantees consistent column widths throughout.

Numbering: assign numbers by counting through all Personal tasks first (across all date groups), then all Work tasks (across all date groups). This keeps the numbering contiguous within each workspace.

## Asana — next 7 days

| #  | Personal                                | #  | Work                                    |
|----|-----------------------------------------|----|-----------------------------------------|
|    | 🔴 **Overdue**                          |    |                                         |
| 1  | Task name — was due Apr 3               | 8  | Task name — was due Apr 5               |
|    | 📅 **Today (Apr 7)**                    |    |                                         |
| 2  | Task name (Project)                     | 9  | Task name (Project)                     |
| 3  | Task name (Project)                     |    |                                         |
|    | 📅 **Tomorrow (Apr 8)**                 |    |                                         |
| 4  | Task name                               |    |                                         |
|    | 📅 **Fri Apr 10**                       |    |                                         |
| 5  | Task name                               | 10 | Task name (Project)                     |

• Single table for the entire view — this forces all columns to the same width.
• Date groups appear as bold label rows with an emoji prefix: 🔴 for Overdue, 📅 for all other date groups. This adds color to make date headers visually distinct from task rows.
• Numbering order: first pass numbers all Personal tasks sequentially across all date groups (1, 2, 3, …), then second pass numbers all Work tasks continuing from where Personal left off. Within each workspace, numbers follow date order.
• Show project name in parentheses if available.
• When one column has more tasks than the other in a date group, leave the shorter column's cells empty.
• Omit date groups that have no tasks.
• If no tasks at all, say so and stop.

3. Interactive check-off (free-text)

After displaying the numbered list, ask the user to type the numbers of tasks they want to mark complete. Example prompt:

Type the numbers of tasks to check off (e.g. 1, 3, 8, 11), or none to skip.

• Accept comma-separated numbers, ranges (3-5), or none/skip to exit.
• Validate that all numbers correspond to tasks in the displayed list.
• If any numbers are invalid, report which ones and re-prompt once.

4. Mark complete

For all selected tasks:

• Call asana_update_task on the correct workspace's MCP (asana_work or asana_personal) with completed: true.
• Run all update calls in parallel where possible.
• Report success/failure inline: Completed: <task name> or Failed: <task name> — <error>.
• If no tasks were selected, say "No tasks marked complete" and stop.

Notes

• This is a read + interactive update skill. No vault files are created or modified.
• Cap at 25 tasks per workspace to keep the list actionable.
• Tasks without a due date are excluded (this skill is specifically for deadline-driven work).
• Both workspaces use asana_get_my_tasks (not search_tasks) to ensure only truly assigned tasks appear. search_tasks with assignee_any: me returns collaborator/follower tasks too, flooding the list with pipeline items that the user isn't responsible for. Free-tier workspaces also require get_my_tasks because search_tasks is paid-plan-only (402 on search_tasks).
• Never serialize the Work and Personal fetches. They are independent — always issue them in the same tool-use block.