MCP server for searching YouTube videos and generating summaries from video transcripts
Two tools with minimal descriptions (both under 50 chars) and basic schemas. 'search' lacks parameter descriptions; 'createSummary' has only videoId with minimal context. No output schemas documented. No error handling guidance. No parameter constraints (enums, ranges). Naming is verb-based but descriptions fail to explain WHEN to use each tool or WHAT they return. Missing dependency hints (e.g., 'search first to get videoId'). No indication of API rate limits, timeout behavior, or failure modes.
Generate a summary from a Youtube transcript
Search for Youtube videos
Tool descriptions are too short (under 50 chars each). 'Search for Youtube videos' and 'Generate a summary from a Youtube transcript' lack context on WHEN to use, WHAT is returned, and dependencies between tools.
No output schemas documented. LLMs cannot plan downstream calls or extract required fields (e.g., does search return videoId? What fields does createSummary return?). Breaks tool composition.
Parameter 'query' in search tool has no description. Parameter 'videoId' in createSummary has minimal description. LLMs cannot infer expected format, constraints, or examples.
No error handling guidance. No indication of what happens on invalid videoId, network failure, quota exhaustion, or missing transcript. Agents cannot recover from failures.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 40 | 2026-07-28+ | v2 |
No pagination or result limits documented for search. If YouTube returns 1000+ results, context window explodes. Missing limit parameter and total_count in response.