Hey, so from the tool output schema provided on the dashboard for the tool - GOOGLEDOCS_GET_DOCUMENT_BY_ID (20260102_00) The doc content should be inside data.body.content.paragraph.elements.text_run.content however, from the tool response, it seems like the content is actually inside data.body.content.paragraph.elements.textRun.content Can you confirm if this is an error with the output schema mentioned on the dashboard?

Hey @Antman, let me check with the team and revert back.

sounds good thanks@

Hey @Antman, team is currently investigating the issue. Also, could you share if this is causing any issues for your use case? Thanks!

So in my use case, I give the llm a flattened list of the output schema, so that the llm can write code and index into the response payload to surface only relevant info into the context window. Since the output schema is incorrect, the code written by my llm was always incorrect, I have now fixed it by manually changing the snake_case to camelCase. However, if some other tool has wrong output schema mentioned in the dashboard then the same problem would happen again. Does this make sense?

Hey Antman I am taking over this from Uday. I think that it is a bug in the published output schema, not in the response. The tool returns Google Docs API's native camelCase JSON (

textRun

,

paragraphStyle

, ...), and the dashboard schema incorrectly shows snake_case for nested fields under

body.content

. We're fixing the schema for

GOOGLEDOCS_GET_DOCUMENT_BY_ID

(and auditing the other Google Workspace tools for the same issue, since you flagged that risk). Keep using the camelCase paths — they're stable and match Google's official API format. We'll update the thread once the corrected schema ships.