sleek-design-mobile-apps_skill

Overview

This skill helps you design and iterate mobile apps using the Sleek AI design platform. It covers creating and managing projects, sending natural-language design instructions, editing specific screens, and rendering screenshots of produced UI. Use it to rapidly prototype UIs, generate screen variants, and fetch visuals for review or handoff.

How this skill works

The skill talks to Sleek's REST API at https://sleek.design/api/v1 using a bearer API key. You can create projects, list components, send chat messages that describe what to build, poll run status, and request screenshots for any component IDs. After any run that creates or updates screens, the skill captures and returns the required screenshots so you always see visuals for the change.

When to use it

• Design a new mobile app from a plain-language brief
• Add or update a specific screen inside an existing project
• List, inspect, or delete Sleek projects and components
• Generate screenshots for review, demos, or documentation
• Run idempotent design requests or poll long-running AI runs

Best practices

• Store the full API key once and use environment variable SLEEK_API_KEY; create keys with only needed scopes
• Prefer ?wait=true for short tasks, otherwise use async + polling with backoff (start 2s, back off to 5s)
• Always follow the screenshot delivery rule: return screenshots for created/updated screens and a combined project image when screens are newly created
• Use idempotency-key for safe retries to avoid duplicate runs
• Keep imageUrls limited to non-sensitive, publicly accessible HTTPS resources

Example use cases

• "Design a fitness tracking app with onboarding, dashboard, and activity screens" — creates a project and generates screens
• "Edit the Settings screen to add a privacy toggle" — target a specific screenId and return the updated screenshot
• "List my projects" — returns project metadata and component lists for inspection
• "Create a new project called 'Onboarding Test'" — POST to projects and begin a chat run to generate initial screens
• "Take screenshots of all screens in project proj_abc" — POST to /api/screenshots with componentIds and transparent background