From 8d8ebb87869f8784e3ef213b27cde0235a46ffb2 Mon Sep 17 00:00:00 2001 From: kurihada Date: Wed, 17 Dec 2025 13:47:58 +0800 Subject: [PATCH] =?UTF-8?q?feat(core):=20=E6=9B=B4=E6=96=B0=20web=20?= =?UTF-8?q?=E5=B7=A5=E5=85=B7=E6=8F=8F=E8=BF=B0=E5=B9=B6=E6=B7=BB=E5=8A=A0?= =?UTF-8?q?=E5=9F=9F=E5=90=8D=E8=BF=87=E6=BB=A4=E5=8F=82=E6=95=B0?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit - 更新 web_extract 工具描述为英文版本 - 更新 web_search 工具描述，添加搜索日期提示 - 为 web_search 添加 allowed_domains 和 blocked_domains 参数 - 添加 GET_CURRENT_DATE_FN 模板函数支持动态日期 --- packages/core/src/template/renderer.ts | 5 +++ .../tools/descriptions/web/web_extract.txt | 33 ++++++++-------- .../src/tools/descriptions/web/web_search.txt | 38 ++++++++++--------- packages/core/src/tools/web/web_search.ts | 14 +++++++ 4 files changed, 55 insertions(+), 35 deletions(-) diff --git a/packages/core/src/template/renderer.ts b/packages/core/src/template/renderer.ts index ad65160..5506c9a 100644 --- a/packages/core/src/template/renderer.ts +++ b/packages/core/src/template/renderer.ts @@ -482,6 +482,11 @@ export function createToolDescriptionContext( CUSTOM_TIMEOUT_MS: () => CUSTOM_TIMEOUT_MS, MAX_TIMEOUT_MS: () => MAX_TIMEOUT_MS, MAX_OUTPUT_CHARS: () => MAX_OUTPUT_CHARS, + GET_CURRENT_DATE_FN: () => { + const now = new Date(); + const pad = (n: number) => n.toString().padStart(2, '0'); + return `${now.getFullYear()}-${pad(now.getMonth() + 1)}-${pad(now.getDate())} ${pad(now.getHours())}:${pad(now.getMinutes())}:${pad(now.getSeconds())}`; + }, }; return { diff --git a/packages/core/src/tools/descriptions/web/web_extract.txt b/packages/core/src/tools/descriptions/web/web_extract.txt index e56af1a..a296295 100644 --- a/packages/core/src/tools/descriptions/web/web_extract.txt +++ b/packages/core/src/tools/descriptions/web/web_extract.txt @@ -1,19 +1,16 @@ -从指定的网页 URL 提取内容。使用 Tavily Extract API 智能解析网页，返回结构化的文本内容。 +- Fetches content from a specified URL and processes it using an AI model +- Takes a URL and a prompt as input +- Fetches the URL content, converts HTML to markdown +- Processes the content with the prompt using a small, fast model +- Returns the model's response about the content +- Use this tool when you need to retrieve and analyze web content -适用场景： -- 获取网页文章的完整内容 -- 提取文档页面的详细信息 -- 抓取多个页面进行对比分析 -- 获取网页中的图片列表 -- 深度提取包含表格、嵌入内容的页面 - -参数说明： -- urls: URL 列表（必填，最多 20 个，也可传单个 URL 字符串） -- extract_depth: "basic" 快速提取 / "advanced" 深度提取（含表格等） -- format: "markdown" / "text" 输出格式，默认 markdown -- include_images: 是否包含图片列表，默认 false - -返回内容： -- 每个 URL 的提取内容 -- 图片列表（如果启用） -- 失败的 URL 及错误信息 +Usage notes: + - IMPORTANT: If an MCP-provided web fetch tool is available, prefer using that tool instead of this one, as it may have fewer restrictions. + - The URL must be a fully-formed valid URL + - HTTP URLs will be automatically upgraded to HTTPS + - The prompt should describe what information you want to extract from the page + - This tool is read-only and does not modify any files + - Results may be summarized if the content is very large + - Includes a self-cleaning 15-minute cache for faster responses when repeatedly accessing the same URL + - When a URL redirects to a different host, the tool will inform you and provide the redirect URL in a special format. You should then make a new WebFetch request with the redirect URL to fetch the content. diff --git a/packages/core/src/tools/descriptions/web/web_search.txt b/packages/core/src/tools/descriptions/web/web_search.txt index f47ba8c..9a79502 100644 --- a/packages/core/src/tools/descriptions/web/web_search.txt +++ b/packages/core/src/tools/descriptions/web/web_search.txt @@ -1,21 +1,25 @@ -搜索网络获取最新信息。使用 Tavily API 进行智能搜索，返回相关网页内容和 AI 摘要。 +- Allows Claude to search the web and use the results to inform responses +- Provides up-to-date information for current events and recent data +- Returns search result information formatted as search result blocks, including links as markdown hyperlinks +- Use this tool for accessing information beyond Claude's knowledge cutoff +- Searches are performed automatically within a single API call -这是进行网络搜索的首选工具，不要使用 curl 或 bash 命令来搜索网络。 +CRITICAL REQUIREMENT - You MUST follow this: + - After answering the user's question, you MUST include a "Sources:" section at the end of your response + - In the Sources section, list all relevant URLs from the search results as markdown hyperlinks: [Title](URL) + - This is MANDATORY - never skip including sources in your response + - Example format: -适用场景： -- 查询最新新闻、事件、游戏更新 -- 搜索技术文档、API 参考 -- 获取实时数据（股价、天气等） -- 查找开源项目、库的信息 -- 了解最新的技术趋势 + [Your answer here] -参数说明： -- query: 搜索关键词（必填） -- max_results: 返回结果数量，1-20，默认 5 -- search_depth: "basic" 快速搜索 / "advanced" 深度搜索 -- topic: "general" 通用 / "news" 新闻 / "finance" 财经 -- include_answer: 是否包含 AI 摘要，默认 true + Sources: + - [Source Title 1](https://example.com/1) + - [Source Title 2](https://example.com/2) -返回内容： -- AI 生成的摘要答案 -- 相关网页列表（标题、链接、内容摘要） +Usage notes: + - Domain filtering is supported to include or block specific websites + - Web search is only available in the US + +IMPORTANT - Use the correct year in search queries: + - Today's date is ${GET_CURRENT_DATE_FN()}. You MUST use this year when searching for recent information, documentation, or current events. + - Example: If today is 2025-07-15 and the user asks for "latest React docs", search for "React documentation 2025", NOT "React documentation 2024" diff --git a/packages/core/src/tools/web/web_search.ts b/packages/core/src/tools/web/web_search.ts index 94af535..714ce90 100644 --- a/packages/core/src/tools/web/web_search.ts +++ b/packages/core/src/tools/web/web_search.ts @@ -41,6 +41,16 @@ export const webSearchTool: ToolWithMetadata = { description: '是否包含 AI 生成的摘要答案（默认 true）', required: false, }, + allowed_domains: { + type: 'array', + description: '只包含这些域名的结果', + required: false, + }, + blocked_domains: { + type: 'array', + description: '排除这些域名的结果', + required: false, + }, }, execute: async (params: Record): Promise => { const query = params.query as string; @@ -48,6 +58,8 @@ export const webSearchTool: ToolWithMetadata = { const searchDepth = (params.search_depth as 'basic' | 'advanced') || 'basic'; const topic = (params.topic as 'general' | 'news' | 'finance') || 'general'; const includeAnswer = params.include_answer !== false; + const includeDomains = params.allowed_domains as string[] | undefined; + const excludeDomains = params.blocked_domains as string[] | undefined; // 权限检查 const permissionManager = getPermissionManager(); @@ -94,6 +106,8 @@ export const webSearchTool: ToolWithMetadata = { topic, maxResults, includeAnswer, + includeDomains, + excludeDomains, }); // 格式化输出