MCP Server for Personal Book Library. Ingest ebooks (EPUB, MOBI, AZW3, PDF, TXT), convert to text via Calibre, chunk and embed with Gemini, auto-categorize, and provide semantic search across your entire library.
Server has 6 tools with basic descriptions and some parameter schemas, but significant gaps prevent higher scoring. Tool descriptions are present but generic (avg ~60 chars, below 194-char baseline). Most parameters lack descriptions entirely, only 'query' and 'book_id' have any annotation. No output schemas documented. No error handling guidance. No tool annotations (readOnlyHint/destructiveHint). Naming is reasonable (verb_noun pattern) but parameter descriptions are almost entirely absent, forcing LLMs to guess intent. This is typical of community servers landing in the D/C range.
Find relevant passages from a specific book
First chunks of a book (intro/preface)
Library statistics
Scan inbox folder for new books, convert, chunk, embed, categorize
List all books with metadata
Semantic search across all books
Missing parameter descriptions on 4 of 6 tools. 'ingest_books' and 'list_books' have empty input schemas with no guidance. 'get_library_status' similarly lacks any parameter documentation. LLMs cannot infer what parameters control without explicit descriptions.
No output schemas documented for any tool. Callers do not know what fields to expect (e.g., does search_books return book_id, title, relevance_score, chunk_text?). This forces LLMs to guess and wastes tokens on trial-and-error.
No tool annotations (readOnlyHint, destructiveHint, idempotentHint). 'ingest_books' is marked WRITE but the tool definition lacks destructiveHint annotation. Agents cannot determine which tools are safe to retry without explicit hints.
Inferred effective spec: <=2025-11-25.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 44 | <=2025-11-25 | v2 |
Error handling is absent. No recovery guidance, no categorization of retryable vs fatal errors, no actionable error messages. If a book ingestion fails or a search returns no results, the LLM has no guidance on what to do next.
No pagination support on list_books or search_books. If a library has hundreds of books, returning all results will blow the context window. Missing limit/offset parameters and total_count in response.