OptionalCategoryOptionalDefaultDefault timeout in ms for this specific tool (overrides agent default)
Human-readable description — this is what the LLM reads to decide when to use it
JSON Schema describing the input parameters
Unique identifier for this tool
OptionalOutputJSON Schema describing what the tool returns (optional, for LLM context)
OptionalTrustOptional output trust classification; when 'Untrusted' or 'SemiTrusted', tool output is wrapped in <untrusted_external_data>
Category for grouping in prompts (e.g., 'navigation', 'display', 'data')