引導式測驗 · 中等 等級
將 OpenAPI 規格轉換為 LLM 工具測驗
將 OpenAPI 操作對應到工具架構、管理端點、驗證參數、保護憑證並強制執行應用程式端授權。
問題 1 的 8
What does converting an OpenAPI spec into LLM tools primarily involve mapping?
預覽每個問題8 問題
當您參加測驗時,就會出現答案和解釋。
- What does converting an OpenAPI spec into LLM tools primarily involve mapping?
- Each endpoint's path, method and parameters into a tool schema a model can call
- Database table names into tool names
- CSS styles into tool descriptions
- Server uptime logs into tool metadata
- Why would a team expose only 6 of 140 documented endpoints as tools?
- Because OpenAPI only allows 6 endpoints total
- Because including the full spec can overwhelm the model's context and increase the chance it picks the wrong endpoint
- Because tools must be written in alphabetical order
- Because the API only supports 6 HTTP methods
- Why can an application require confirmation before a tool calls a sensitive endpoint?
- OpenAPI fully defines the product’s confirmation policy
- Models cannot return JSON arguments
- The application owns risk policy and user confirmation; an API description alone does not decide these
- HTTP methods never change data
- Which OpenAPI version family aligns its Schema Object with JSON Schema 2020-12, easing conversion to JSON-Schema-based tool parameters?
- YAML front matter
- Protocol Buffers
- XML DTD
- OpenAPI 3.1 and later
- How are API credentials from an OpenAPI security scheme typically handled when converting to tools?
- They are injected server-side at execution time, after the model has chosen the call and arguments
- The model is given the raw API key as a tool parameter to fill in
- They are ignored entirely and requests are sent unauthenticated
- They are hardcoded into the model's weights
- In the weather API example, why did the model inconsistently omit the units parameter?
- Because the API rejected all requests with units included
- Because units was marked optional with no example value, giving the model little guidance
- Because the parameter name was misspelled in the code
- Because OpenAPI does not support optional parameters
- According to the guide, what is a common misconception about feeding an entire large OpenAPI file to a model as tools?
- That it is technically impossible to attempt
- That OpenAPI files cannot contain more than 10 endpoints
- That it works reliably out of the box regardless of tool count or description clarity
- That models cannot process JSON Schema at all
- What part of the conversion process most often still requires human judgment rather than pure automation?
- Copying the HTTP method into the tool name
- Formatting the JSON response
- Choosing the server's port number
- Curating which endpoints to include and rewriting descriptions for model clarity