Agent 工具
一个小型可扩展的决策到执行框架,适用于 TypeSafe Jev。
A small, extensible decision-to-action harness for TypeSafe Jev
return this.transport('https://api.typesafe.ai/v1/systemone', {链接指向我们抓取当天的那个 commit,行号是准的。
下面是从这个项目源码里原样取出来的 question 组合。
q1是/否is it true?
import { experimental_evaluate as evaluate } from 'ai';
const { answers } = await evaluate({
model: 'typesafe-ai/jev',
state,
questions: {
q1: { type: 'boolean', instructions: 'is it true?' },
},
});ok是/否Is this a greeting?
import { experimental_evaluate as evaluate } from 'ai';
const { answers } = await evaluate({
model: 'typesafe-ai/jev',
state,
questions: {
ok: { type: 'boolean', instructions: 'Is this a greeting?' },
},
});