Quick Desktop Agent Library

Loading agent details…

✨

Create your own agent

Answer a few questions — we'll build the .md file for you.

Name your agent

Give it a short memorable identity. Lowercase letters, numbers, and dashes only for the ID.

Agent ID *

This is how Quick Desktop will identify your agent. Lowercase letters, numbers, and dashes only.

Display name *

Shown in the library grid and at the top of the file. Name followed by a short tagline works well.

Emoji icon optional

Category optional

Examples: Daily Briefing, Email, Calendar, Meeting Prep, People, Travel, Infrastructure.

Short description *

Shown on the agent card. One to two sentences. Max 280 characters.

Describe what it should do

Write this like you're telling a smart assistant what to do on each run. Include triggers, procedures, and outputs. Plain prose is fine — Quick Desktop reads the whole thing as the system prompt.

Long description (for the library) *

2 to 4 sentences describing what the agent accomplishes and for whom. This is what other people see in the library.

Procedure / instructions *

Describe the full workflow: triggers, steps, data sources, output format, edge cases, rules. The more specific, the better. Quick Desktop will use this as the agent's system prompt.

When should it run?

Examples: Interval for agents that react to new data continuously (e.g. every 15 min). Once daily for morning briefings or nightly reports. Specific days + time for weekly jobs (e.g. Monday 5 PM for a standup draft).

Schedule type

Interval (minutes)

Typical values: 10, 15, 30, 60.

What tools does it need?

Check any tool groups this agent needs. Pick the minimum it actually requires — Quick Desktop denies everything else by default. For the Tuesday-standup-prep example, you'd enable Outlook (calendar + sent email), Slack (read recent messages), Core (activity feed), and run_python (to process the data).

Outlook
Email, calendar, contacts, To-Do lists Slack
Messages, channels, profile lookups Core (feed, memory, state)
Activity feed, memory, agent introspection run_python
In-process Python for data processing File read / write
Local files (state, logs, notes) Web search
External web / URL fetch Phone Tool
Employee directory lookups Zoom meeting creation
Create Zoom meetings on demand

Model

Review and download

Your agent file is ready. Edit it directly below if you want to tweak anything before downloading or saving.

Generated file — edit freely

📤

Upload an existing agent

Drop an .md file — we'll scrub personal info and prep it for the library.

Drop your .md file

Your file is scrubbed entirely in your browser. Nothing is uploaded until you hit "Save to library" on the last step.

📄 Click or drag a .md file here

Max 500 KB · .md / .markdown / .txt

Review what we scrubbed

Everything that looked like personal info (emails, Slack IDs, addresses, phone numbers, local paths, dashboard URLs, Zoom bridges) has been replaced with a placeholder. Scan the preview for any remaining names — we don't auto-scrub human names because that's error-prone. Edit directly to fix anything we missed.

Scrubbed file — edit freely

Tell the library what this is

We extracted what we could from the file. Fill in anything missing.

Agent ID *

Lowercase letters, numbers, and dashes. We'll prefix it with custom- when saving to the library.

Display name *

Emoji icon optional

Category optional

Short description *

Long description *

Save or download

Download the scrubbed .md for your own use, or save it to the library so other people can deploy it too.

Quick Desktop Agent Library

Request access

Set a permanent password

Quick Desktop Agent Library

Pick an agent to deploy

Quick Desktop Agent Library

Sign in

Request access

Set a permanent password

Pick an agent to deploy