- 截断/畸形 HTML 不再漏残标签与半截 script 正文(正则 [\s\S]*? 遇缺闭标签整体失配的必然缺陷) - class/id 页壳启发式限定块级容器,不再连坐删正文 - 相对 href/src 按重定向终址绝对化,抓回的链接可直接作为下一次 fetch 参数 - 标题走 DOM 规则取(修复先删 head 后取 title 会取空的问题) - 新增解析硬帽 10MB:超限与解析抛错均降级轻量剥标签,不空手而归 - 截断提示文案区分「下载少了」与「解析坏了」
20 lines
483 B
JSON
20 lines
483 B
JSON
{
|
|
"name": "agent-app",
|
|
"version": "2.3.0",
|
|
"description": "自用通用 Agent 系统:多模型调度 + 任务目录 + 工具调用 + H5 Web UI",
|
|
"main": "server/server.js",
|
|
"scripts": {
|
|
"start": "node server/server.js",
|
|
"test": "node test/smoke.js"
|
|
},
|
|
"dependencies": {
|
|
"better-sqlite3": "11.3.0",
|
|
"cron-parser": "5.10.1",
|
|
"express": "4.21.2",
|
|
"multer": "1.4.5-lts.1",
|
|
"sharp": "0.35.4",
|
|
"turndown": "7.2.0",
|
|
"yaml": "2.6.1"
|
|
}
|
|
}
|