Three tools with basic descriptions and minimal schemas. play-flappy-bird and get-high-score have empty input schemas ({}), which is acceptable for no-parameter tools. submit-score has a typed schema with minimum constraint. Descriptions are present but generic (10-100 chars), lacking LLM-optimized guidance on WHEN to use each tool or what happens next. No parameter descriptions for submit-score's 'score' field. No output schemas documented. No error handling guidance. Tool names follow verb_noun pattern (good), but lack clarity on state mutations and side effects.
Get the current Flappy Bird high score
Launch a game of Flappy Bird! A classic side-scrolling game where you tap to keep the bird flying through gaps in pipes.
Submit a score from the Flappy Bird game
submit-score parameter 'score' lacks description. LLMs cannot infer whether this is the player's current score, a delta, or a high-score submission target.
No output schemas documented for any tool. LLMs cannot plan downstream calls or extract structured data. play-flappy-bird returns JSON with 'highScore'; submit-score returns 'highScore' and 'submitted'; get-high-score returns 'highScore'. These should be formally declared.
Descriptions are under 100 characters and lack LLM-optimized guidance. 'Submit a score from the Flappy Bird game' does not explain WHEN to call this vs get-high-score, or what happens if the score is lower than the current high score.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | D | 54 | 2026-07-28+ | v2 |
No error handling guidance. If submit-score fails (e.g., invalid score), the LLM has no recovery path. No indication of retryability or user-fixable errors.
play-flappy-bird description does not clarify that it launches a UI game and returns the current high score. LLMs may not understand the side effect of rendering interactive content.