Descriptions are present but minimal (10-50 chars), below the 50-200 char baseline for LLM-optimized descriptions. No parameter descriptions exist (all tools have zero parameters). No output schemas documented. No error handling guidance. Tools lack idempotency hints, permission declarations, or recovery paths. This server fails to meet production-grade definition standards.
Get all available categories for Chuck Norris jokes
Get a random Chuck Norris joke
Get a random dad joke
Get a random Yo Mama joke
All tools have empty input schemas (parameters: {}).
Tool descriptions are 10-50 characters, below the 50-200 char baseline. Descriptions lack WHAT/WHEN/WHY context needed for LLM selection.
No output schemas documented. LLMs cannot predict response structure (e.g., is it {content: [{type, text}]} or {joke: string}?). Breaks downstream tool chaining.
No error handling guidance. If external API fails (chucknorris.io, icanhazdadjoke.com down), tools return raw fetch errors with no recovery hints. LLM cannot self-correct.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 39 | 2026-07-28+ | v2 |
No tool annotations (readOnlyHint, idempotentHint). All tools are read-only and idempotent, but this is not declared in the schema. Agents cannot optimize caching or retry logic.