Agent 工具
Claude Code 插件，使用 TypeSafe 的 Jev 对评审结果、调试假设和设计方案进行评分，提供校准概率而非单一意见。
Claude Code plugin that scores review findings, debug hypotheses and design options with TypeSafe's Jev — calibrated probabilities instead of one more opinion.
import { TypeSafeClient } from '@typesafe-ai/sdk';src/infrastructure/typesafe-jev.ts:1
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。
取自 tests/typesafe-jev.test.ts:37
is_urgent是/否Does this convey urgency?
frustration打分How frustrated is the customer?
0 — Calm1 — Frustrated2 — Very angryimport { experimental_evaluate as evaluate } from 'ai';
const { answers } = await evaluate({
model: 'typesafe-ai/jev',
state,
questions: {
is_urgent: { type: 'boolean', instructions: 'Does this convey urgency?' },
frustration: { type: 'score', instructions: 'How frustrated is the customer?',
criteria: ['Calm', 'Frustrated', 'Very angry'] },
},
});取自 tests/typesafe-jev.test.ts:135
plain是/否Is it fine?
import { experimental_evaluate as evaluate } from 'ai';
const { answers } = await evaluate({
model: 'typesafe-ai/jev',
state,
questions: {
plain: { type: 'boolean', instructions: 'Is it fine?' },
},
});q是/否Is it fine?
import { experimental_evaluate as evaluate } from 'ai';
const { answers } = await evaluate({
model: 'typesafe-ai/jev',
state,
questions: {
q: { type: 'boolean', instructions: 'Is it fine?' },
},
});