MCP server for MAScope — World MiniApp reviews and analytics powered by verified humans
MAScope MCP demonstrates solid tool design with consistent naming patterns (all verb_noun), comprehensive parameter schemas using Zod, and clear descriptions. All 6 tools are READ_ONLY and well-scoped. However, output schemas are not explicitly documented, responses are JSON-stringified without field-level documentation. Error handling is basic (generic API errors without recovery guidance). Parameter descriptions are good but could be more prescriptive about constraints. No tool annotations (readOnlyHint, etc.) despite all being read-only.
Compare 2-5 World MiniApps side by side. Returns rank, users, opens, rating, review count, and trend data for each app. Useful for competitive analysis.
Get detailed information about a specific World MiniApp, including rank, user metrics, verified ratings, trend data (7-day sparkline), and developer info.
Get World ID-verified reviews for a MiniApp. Every review is posted by a human verified through proof-of-personhood. Includes ratings, comments, sub-ratings (usability/recommendation/reliability), and a summary with rating distribution.
Get a breakdown of all World MiniApp categories with app count, total weekly users, and top 3 apps per category. Useful for understanding the MiniApp ecosystem landscape.
Get currently trending World MiniApps sorted by growth rate. Shows which apps are gaining users or opens the fastest, with trend direction, change percentage, and sparkline data.
Search World MiniApps by keyword, category, or sort criteria. Returns app name, description, rank, weekly users, rating, and review count. All reviews on MAScope are verified by World ID (proof-of-personhood).
Output schemas not documented. Tools return JSON.stringify(data) without specifying what fields the LLM should expect. This forces LLMs to infer structure and risks misuse of returned data.
Error handling is generic. apiCall() throws raw API errors without recovery guidance. E.g., 'MAScope API error (404)' tells the LLM nothing about what to do next. Should include 'Try search_miniapps() first' or 'Invalid app_id format' hints.
No tool annotations. All 6 tools are read-only but lack readOnlyHint annotation. This prevents clients from optimizing caching, retry logic, or safety checks.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | D | 59 | 2026-07-28+ | v2 |
Parameter descriptions lack explicit constraints. E.g., 'limit' says '1-50, default 20' in text but should also validate and reject out-of-range values with actionable error messages.
No pagination metadata in responses. Tools accept limit/offset but responses don't indicate total count or whether more results exist. LLMs cannot reliably paginate without this.