A Model Context Protocol (MCP) server that connects AI models to GROWI wiki content. Enables LLMs to search and retrieve information from your organization's knowledge base for accurate, context-aware responses.
The GROWI MCP server has 13 tools with significant quality gaps. Tool naming is generally verb-first and clear (suggest_path, get_comments, add_comment, etc.), following convention well. However, schema and parameter documentation are largely absent from the visible source. The sample code shows tool registration files exist but their actual schema definitions, parameter types, and descriptions are not visible in the provided source excerpt. This forces conservative scoring: without seeing explicit Zod schemas or JSON Schema definitions, input validation rules, parameter types, and output field documentation, we cannot credit schema quality. The server loads tools dynamically via `loadTools(server)` but the actual tool definitions in `/src/tools/*/register.ts` files are not included in the source provided. Based on the visible code structure alone, the framework (fastmcp with Zod for validation) is sound, but the actual tool specifications are opaque.
Tool definitions not visible in source. Register.ts files are imported but their actual content (Zod schemas, parameter descriptions, output types) is not provided. Cannot verify schema quality, parameter types, or descriptions for any tool.
No visible parameter documentation. Tool names inferred from file structure and imports. Without explicit schema definitions, cannot verify that parameters have descriptions, types, enums, or validation rules.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-22 | F | 30 | 2026-07-28+ | v2 |
| 2026-04-07 | F | 35 | - | v1 |
No visible output schema documentation. Cannot verify that tool responses document their return fields, types, pagination, or field chaining (e.g., whether create_page returns a page_id that edit_page can accept).
Destructive operations (delete_pages, remove_comment) have no visible confirmation or dry-run mechanism. High risk of accidental data loss.
No visible error handling strategy. Cannot verify that errors include actionable recovery guidance, retry classification, or clarification of constraints.
Prompts feature enabled but not evaluated. Single prompt 'summarizePagePrompt' referenced but definition not visible in source.