Agent 工具
这是一个Claude Code插件，使用TypeSafe AI的Jev在agent执行工具调用前验证调用是否符合会话计划，然后允许、询问或阻止调用。
Every tool call your agent makes, checked before it runs. A Claude Code plugin that uses TypeSafe AI's Jev to verify each pending tool call against the session plan, then allows it, asks you, or blocks it. Proof of concept
* Wire format (verified against api.typesafe.ai, 2026-09-18):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。
outOfScope是/否Does current_tool_call read or modify a file, resource, host, or system outside what the plan describes?
contradictsPrevious是/否Does current_tool_call contradict, reverse, or undo the most recent entry in recent_tool_calls?
shouldFlag是/否Independent of the other answers: is this destructive, irreversible, or otherwise something a human should review before it runs?
import { experimental_evaluate as evaluate } from 'ai';
const { answers } = await evaluate({
model: 'typesafe-ai/jev',
state,
questions: {
outOfScope: { type: 'boolean', instructions: 'Does current_tool_call read or modify a file, resource, host, or system outside what the plan describes?' },
contradictsPrevious: { type: 'boolean', instructions: 'Does current_tool_call contradict, reverse, or undo the most recent entry in recent_tool_calls?' },
shouldFlag: { type: 'boolean', instructions: 'Independent of the other answers: is this destructive, irreversible, or otherwise something a human should review before it runs?' },
},
});