MCP server for managing app store preview screenshots with MUSE Preview Maker
Server has 11 tools with complete input schemas and descriptions, but quality is uneven. Naming follows verb_noun convention well (create_, add_, list_, remove_, update_, clear_, open_, generate_, get_, capture_). Descriptions are present but often verbose and lack actionable guidance for LLM selection. Parameters are typed with enums where appropriate (platform, language, deviceId patterns), but many lack detailed constraints or format guidance. Output schemas are not documented, critical gap for agent planning. Error handling is minimal; no recovery guidance visible. Tool composition has issues: create_app_previews conflates multiple concerns (selection, generation, capture, export); clear_all is destructive without sufficient safeguards.
Add a new preview set with screenshot, title, and subtitle. The screenshot will be used to generate an app store preview image.
Capture a screenshot from the running iOS Simulator and add it as a preview set. Claude can use this to automatically capture app screens and generate marketing text.
Clear all saved preview sets.
Complete workflow to create App Store preview images. Steps: 1) User selects platform (iOS/macOS/watchOS), 2) User selects language, 3) Claude generates marketing text, 4) Auto-capture screenshots, 5) Send to MUSE Preview Maker app or save to folder.
Generate all preview images using MUSE Preview Maker. The app will process all saved preview sets and export them.
Get current settings (default device type, palette, output directory).
Output schemas not documented. LLMs cannot plan downstream calls or extract required fields (e.g., what does list_previews return? What fields are in each preview object?). This forces agents to guess structure and wastes tokens on parsing.
create_app_previews conflates multiple responsibilities: platform selection, language selection, text generation, screenshot capture, and export. Should be split into separate tools (e.g., generate_preview_text, capture_app_screenshots, export_to_muse) so agents can compose them independently.
clear_all is destructive but lacks confirmation safeguards. The 'confirm' boolean parameter is weak, agents may pass true without understanding consequences. Should require explicit resource ID confirmation or a dry-run preview.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | D | 56 | 2026-07-28+ | v2 |
List all saved preview sets with their details.
Open MUSE Preview Maker app. If previews are saved, they will be loaded automatically.
Remove a specific preview set by its ID.
Update an existing preview set.
Update default settings.
Descriptions lack actionable guidance for LLM selection. E.g., 'Add a new preview set with screenshot, title, and subtitle' does not explain when to call this vs create_app_previews, or what happens next. Descriptions should answer: What does it do? When to use it? What does it return?
Parameter descriptions lack format constraints. E.g., screenshotPath says 'Absolute path to the screenshot image file' but does not specify: must exist? which formats (PNG, JPG)? max size? deviceId accepts 'iphone_6_7, iphone_6_5, ipad_12_9, mac_retina, etc.' but 'etc.' is vague, should enumerate all valid values or reference a discovery tool.