A Model Context Protocol server allows to interact with Twitter, enabling posting tweets and searching Twitter.
Two tools with basic schemas and descriptions, but significant gaps in parameter documentation and output schema clarity. post_tweet has a clear action verb and reasonable description (38 chars), but reply_to_tweet_id lacks a description entirely. search_tweets has good constraints (min/max on count) but the description is generic (34 chars, below the 50-char baseline for clarity). Neither tool documents output schema structure, forcing LLMs to infer response format. No error recovery guidance provided. Schemas are present but incomplete, missing descriptions on optional parameters violates the pattern:tool-description requirement that every parameter needs context.
Post a new tweet to Twitter
Search for tweets on Twitter
Missing parameter descriptions: reply_to_tweet_id in post_tweet has no description; query in search_tweets lacks context on format/length constraints
No output schema documented for either tool. LLMs cannot infer response structure (e.g., does search_tweets return tweet objects with id/text/author, or just raw text?). Breaks pattern:tool and mxe:include-chaining-ids.
No error recovery guidance. Errors like rate limits, auth failures, or invalid queries return generic messages without actionable next steps (e.g., 'try again later', 'check credentials'). Violates pattern:recovery-guide.
Generic tool descriptions (34 - 38 chars) below the 50-char baseline for LLM clarity. 'Search for tweets on Twitter' does not explain when to use this vs. other discovery tools or what fields are returned.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | C | 60 | 2026-07-28+ | v2 |
No pagination or result limits documented for search_tweets. Returning unbounded tweet lists risks context window exhaustion. Missing pattern:paginated-result.