MCP server for searching and retrieving information from Bilibili (Chinese video platform)
Three tools with clear verb-noun naming (search_videos, get_video_info, get_user_info) and basic descriptions. All tools have input schemas with types and parameter descriptions. However, output schemas are not documented, LLMs cannot predict response structure for planning. Parameter descriptions lack format constraints (e.g., BVID format, UID range). No error handling guidance. Descriptions are adequate (50-100 chars) but lack dependency hints and recovery paths. Missing tool annotations (readOnlyHint present in metadata but not in schema). No pagination support despite search_videos returning lists.
Get a Bilibili creator's basic profile, follower count, and bio by UID.
Get a video's title, description, uploader, and play count by BVID.
Search Bilibili videos by keyword and return title, link, and play count.
Output schemas not documented. LLMs cannot predict response structure (fields, types, nested objects) for downstream planning or field extraction.
search_videos returns a list but lacks pagination (limit parameter exists but no offset/page, no total_count, no next_cursor). Large result sets will blow context window.
Parameter descriptions lack format constraints. 'bvid' and 'uid' have no documented format (e.g., 'BVID format: BV + 10 alphanumeric chars', 'UID: positive integer'). LLMs cannot validate inputs.
No error handling guidance. BilibiliAPIError exceptions are raised but tool descriptions do not explain recovery paths (e.g., 'If BVID not found, try search_videos() first').
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | D | 57 | 2026-07-28+ | v2 |
Tool annotations (readOnlyHint) present in metadata but not exposed in input schema. LLMs cannot see that these are safe, idempotent read-only operations.