Loading ToolBench
Rakuten Ichiba MCP with unit-price parsing, shipping flags, and compare_ichiba_value. Fork of mrslbt/rakuten-mcp (Ichiba-only).
Strong foundation with complete schemas, bilingual descriptions, and proper error handling. All 5 tools have input schemas with type definitions and parameter descriptions. Tool names follow verb_noun pattern (search, fetch, browse). Descriptions are detailed (100-200 chars) and include Japanese translations. However, output schemas are not explicitly documented in the source, and some parameter descriptions could be more prescriptive about constraints and use cases. Error handling is present but generic. No critical gaps, but lacks the polish of A-grade servers.
Browse Rakuten Ichiba genre hierarchy. Returns top-level genres or nested children/siblings for a given genre ID.
Fetch real-time Rakuten Ichiba item rankings by genre and period. Returns top-ranked items with sales velocity and trend data.
Search products on Rakuten Ichiba marketplace with filtering, sorting, and pagination. Returns items with unit-price parsing, quantity extraction, and shipping verification flags.
Search Rakuten Ichiba products via Item Price Navi (Product Search API). Returns product-level aggregates across multiple shops.
Search Rakuten Ichiba tags by tag ID. Returns tag metadata and associated item counts.
Output schemas not documented. Tool descriptions state what is returned (e.g., 'Returns items with unit-price parsing') but the actual response structure is not formally specified. LLMs cannot plan downstream operations without knowing field names and types.
Parameter descriptions lack prescriptive constraints. E.g., 'sort' enum values are listed but no guidance on when to use '+affiliateRate' vs '-reviewCount'. 'hits' has min/max but no guidance on performance implications of large values.
Error handling in server.ts is generic. RakutenError.toToolError() is called but the actual error messages are not visible in source. Tool descriptions do not guide LLMs on recovery (e.g., 'If rate-limited, retry after 60 seconds').
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | C | 60 | 2026-07-28+ | v2 |
No pagination guidance in descriptions. ichiba_item_search and ichiba_product_search accept 'page' and 'hits' but descriptions do not explain total result count, whether results are exhaustive, or when to stop paginating.
Tool composition: ichiba_item_search and ichiba_product_search appear to overlap (both search products). Descriptions do not clarify when to use one vs. the other, forcing LLMs to guess or try both.