module function · N/llm
createToolResult()
llm.createToolResult(options)
Creates a tool result to send back to the LLM.
- Implementation
- Not started
- Design
- Open for proposals suggest how it should read
A
suitescript api
1llm.createToolResult({2 call: /* llm.ToolCall, required */,3 outputs: /* Object[], required */4});
| Parameter | Type | Required | Description |
|---|---|---|---|
| options.call | llm.ToolCall | required | The original tool call request from the LLM. This parameter links the result to a specific tool call request. |
| options.outputs | Object[] | required | An array of output objects representing the results of running the tool. Each output object typically includes a result property (or another relevant property) that contains the value to send back to the LLM. |
B
mynt syntax
No Mynt syntax yet.
Nothing in the language compiles to llm.createToolResult(options) today. How should it read?
C
community proposals
Competing syntax for createToolResult, from GitHub Discussions. Select two or more to compare them side by side.
Loading proposals from GitHub…
D
suggest an alternative
Write the Mynt you wish you could write.
A good proposal shows the syntax, the SuiteScript it should become, and argues for it: why it is clearer, how it fits the rest of the language, what it makes harder.
- Clarity. Does it say what happens, without ceremony?
- Consistency. Does it read like
load,saveandRef[field]? - Expressiveness. Does it cover the cases people actually write?
- Honesty. Does it hide a NetSuite cost, such as governance, that a reader should see?