ydc-agent 1.0.1 → 1.0.7

package/README.md

@@ -54,8 +54,8 @@ Key modes: `round-robin` (default), `sequential`, `random`
 
 | Tool | Description |
 |------|-------------|
-| `you_search` | Web and news search |
-| `you_express` | Fast AI answers |
+| `you_cached_info` | Web and news search (cached data, limited to 2024) |
+| `you_express` | Fast AI answers (cached data, limited to 2024) |
 | `you_advanced` | Complex reasoning with compute/research tools |
 | `you_agent` | Custom AI agent with intelligent prompt enhancement |
 | `you_chat` | OpenAI-compatible chat with conversation history |
@@ -71,8 +71,8 @@ Tool descriptions include guidance for AI callers:
 
 | Tool | Built-in Feature |
 |------|------------------|
-| `you_search` | Guides AI to use `call_count=5+` for diverse results |
-| `you_express` | Guides AI to use `call_count=5+` for diverse opinions |
+| `you_cached_info` | Guides AI to use `call_count=5+` for diverse results (cached data ≤2024) |
+| `you_express` | Guides AI to use `call_count=5+` for diverse opinions (cached data ≤2024) |
 | `you_advanced` | **AI Capability Rule**: If caller is GPT-3.5+, results are reference material only - use `call_count=5+` for multiple perspectives |
 | `you_agent` | **Failure Tracking**: Auto-tracks failures, suggests switching to `you_advanced` after threshold (default: 3 failures) |
 
@@ -92,7 +92,7 @@ If you are GPT-3.5 or above (GPT-4, Claude 3.5, etc.):
 | Diverse viewpoints | `you_advanced` | `call_count=5-10, summary=true, show_results=true` |
 | Quick opinions | `you_express` | `call_count=10-20, show_results=true` |
 | Deep research | `you_advanced` | `agent_type="advanced-4.5-research", call_count=1` |
-| Web search | `you_search` | `call_count=5-20, show_results=true` |
+| Cached web search | `you_cached_info` | `call_count=5-20, show_results=true` |
 
 ### ⚠️ Common Mistakes
 

package/README_JA.md

@@ -52,8 +52,8 @@ Key モード：`round-robin`（デフォルト）、`sequential`、`random`
 
 | ツール | 説明 |
 |--------|------|
-| `you_search` | ウェブ・ニュース検索 |
-| `you_express` | 高速 AI 回答 |
+| `you_cached_info` | ウェブ・ニュース検索（キャッシュデータ、2024年まで） |
+| `you_express` | 高速 AI 回答（キャッシュデータ、2024年まで） |
 | `you_advanced` | 複雑な推論（計算/リサーチツール付き） |
 | `you_agent` | カスタム AI エージェント（インテリジェントプロンプト強化） |
 | `you_chat` | OpenAI 互換チャット（会話履歴付き） |
@@ -69,8 +69,8 @@ Key モード：`round-robin`（デフォルト）、`sequential`、`random`
 
 | ツール | 組み込み機能 |
 |--------|--------------|
-| `you_search` | AI に `call_count=5+` で多様な検索結果を取得するよう誘導 |
-| `you_express` | AI に `call_count=5+` で多様な意見を取得するよう誘導 |
+| `you_cached_info` | AI に `call_count=5+` で多様な検索結果を取得するよう誘導（キャッシュデータ ≤2024） |
+| `you_express` | AI に `call_count=5+` で多様な意見を取得するよう誘導（キャッシュデータ ≤2024） |
 | `you_advanced` | **AI 能力ルール**：呼び出し元が GPT-3.5+ の場合、結果は参考資料のみ - `call_count=5+` で多様な視点を収集すべき |
 | `you_agent` | **失敗追跡**：失敗回数を自動追跡、閾値（デフォルト 3 回）超過後に `you_advanced` への切り替えを提案 |
 

package/README_ZH_CN.md

@@ -52,8 +52,8 @@ Key 模式：`round-robin`（默认）、`sequential`、`random`
 
 | 工具 | 说明 |
 |------|------|
-| `you_search` | 网页与新闻搜索 |
-| `you_express` | 快速 AI 回答 |
+| `you_cached_info` | 网页与新闻搜索（缓存数据，限于 2024 年） |
+| `you_express` | 快速 AI 回答（缓存数据，限于 2024 年） |
 | `you_advanced` | 复杂推理（含计算/研究工具） |
 | `you_agent` | 自定义 AI 代理（智能提示增强） |
 | `you_chat` | OpenAI 兼容聊天（含对话历史） |
@@ -69,8 +69,8 @@ Key 模式：`round-robin`（默认）、`sequential`、`random`
 
 | 工具 | 内置功能 |
 |------|----------|
-| `you_search` | 引导 AI 使用 `call_count=5+` 获取多元搜索结果 |
-| `you_express` | 引导 AI 使用 `call_count=5+` 获取多元意见 |
+| `you_cached_info` | 引导 AI 使用 `call_count=5+` 获取多元搜索结果（缓存数据 ≤2024） |
+| `you_express` | 引导 AI 使用 `call_count=5+` 获取多元意见（缓存数据 ≤2024） |
 | `you_advanced` | **AI 能力规则**：如果调用方是 GPT-3.5+，结果仅作为参考资料 - 应使用 `call_count=5+` 收集多元观点 |
 | `you_agent` | **累错追踪**：自动追踪失败次数，超过阈值（默认 3 次）后建议切换到 `you_advanced` |
 

package/README_ZH_TW.md

@@ -52,8 +52,8 @@ Key 模式：`round-robin`（預設）、`sequential`、`random`
 
 | 工具 | 說明 |
 |------|------|
-| `you_search` | 網頁與新聞搜尋 |
-| `you_express` | 快速 AI 回答 |
+| `you_cached_info` | 網頁與新聞搜尋（快取資料，限於 2024 年） |
+| `you_express` | 快速 AI 回答（快取資料，限於 2024 年） |
 | `you_advanced` | 複雜推理（含計算/研究工具） |
 | `you_agent` | 自訂 AI 代理（智慧提示增強） |
 | `you_chat` | OpenAI 相容聊天（含對話歷史） |
@@ -69,8 +69,8 @@ Key 模式：`round-robin`（預設）、`sequential`、`random`
 
 | 工具 | 內建功能 |
 |------|----------|
-| `you_search` | 引導 AI 使用 `call_count=5+` 獲取多元搜尋結果 |
-| `you_express` | 引導 AI 使用 `call_count=5+` 獲取多元意見 |
+| `you_cached_info` | 引導 AI 使用 `call_count=5+` 獲取多元搜尋結果（快取資料 ≤2024） |
+| `you_express` | 引導 AI 使用 `call_count=5+` 獲取多元意見（快取資料 ≤2024） |
 | `you_advanced` | **AI 能力規則**：如果調用方是 GPT-3.5+，結果僅作為參考資料 - 應使用 `call_count=5+` 收集多元觀點 |
 | `you_agent` | **累錯追蹤**：自動追蹤失敗次數，超過閾值（預設 3 次）後建議切換到 `you_advanced` |
 

package/index.js

@@ -121,6 +121,9 @@ const OPENAI_SERVER_ACCESS_TOKENS = process.env.YDC_OPENAI_ACCESS_TOKENS || '';
 const PREFER_SUMMARY = process.env.YDC_PREFER_SUMMARY === 'true';
 const DEFAULT_SUMMARY_LANGUAGE = process.env.YDC_SUMMARY_LANGUAGE || 'en';
 
+// ============ ENABLE USAGE HINTS ============
+const SHOW_USAGE_HINTS = process.env.YDC_ENABLE_USAGE_HINTS !== 'false';
+
 // ============ CUSTOM AGENT CONFIGURATION ============
 const CUSTOM_AGENTS_RAW = process.env.YDC_CUSTOM_AGENTS || '';
 const CUSTOM_AGENTS = CUSTOM_AGENTS_RAW.split(',').map(a => a.trim()).filter(a => a);
@@ -271,16 +274,8 @@ class YouAgentsServer {
     this.server.setRequestHandler(ListToolsRequestSchema, async () => ({
       tools: [
         {
-          name: 'you_search',
-          description: `Web and news search via You.com. Supports multiple calls (1-20) with array results. [Performance: ~2s avg, comparable to GPT-3]
-
-⚠️ AI CAPABILITY RULE: If you are GPT-3.5 level or above:
-- Use call_count=5-20 for diverse search results as reference
-- Integrate and verify results with YOUR own knowledge
-
-[RECOMMENDED USAGE]
-- Diverse search results → call_count=5-20, show_results=true
-- Comparison/summary → call_count=3-5, summary=true${PREFER_SUMMARY ? '\n⚡ PREFERRED: Use call_count=3-5 with summary=true for consolidated insights (default enabled)' : ''}`,
+          name: 'you_cached_info',
+          description: `Web search via You.com's cached/indexed data. ⚠️ CACHED DATA LIMITED TO 2024 - if user asks about post-2024 info, respond like: "This tool's data is limited to 2024, but my knowledge is more current (up to [your cutoff]). I can tell you [answer]. Need me to search anyway, or use my knowledge?" Best for: URLs, documentation, pre-2024 topics. call_count: 1=single, 3-5=summary, 5-20=diverse.`,
           inputSchema: {
             type: 'object',
             properties: { 
@@ -295,7 +290,7 @@ class YouAgentsServer {
         },
         {
           name: 'you_express',
-          description: `Fast AI answers with web search. Supports multiple calls (1-20) for diverse perspectives. [Performance: ~2s avg, comparable to GPT-3. Best for quick answers and multiple parallel calls]
+          description: `Fast AI answers with web search. ⚠️ CACHED DATA LIMITED TO 2024. Supports multiple calls (1-20) for diverse perspectives. [Performance: ~2s avg, comparable to GPT-3. Best for quick answers and multiple parallel calls]${SHOW_USAGE_HINTS ? `
 
 ⚠️ AI CAPABILITY RULE: If you are GPT-3.5 level or above:
 - Use call_count=5-20 for diverse reference opinions, then integrate with YOUR own judgment
@@ -303,7 +298,7 @@ class YouAgentsServer {
 
 [RECOMMENDED USAGE]
 - Quick diverse opinions → call_count=10-20, show_results=true
-- Comparison/summary → call_count=5-10, summary=true${PREFER_SUMMARY ? '\n⚡ PREFERRED: Use call_count=5 with summary=true for consolidated insights (default enabled)' : ''}`,
+- Comparison/summary → call_count=5-10, summary=true` : ''}${PREFER_SUMMARY ? '\n⚡ PREFERRED: Use call_count=5 with summary=true for consolidated insights (default enabled)' : ''}`,
           inputSchema: {
             type: 'object',
             properties: { 
@@ -318,7 +313,7 @@ class YouAgentsServer {
         },
         {
           name: 'you_advanced',
-          description: `Advanced AI agent for complex reasoning and research. Supports multiple calls (1-10) for comprehensive analysis. [Performance Guide: basic=~10s (GPT-3.5), +compute=~120s (GPT-4o), +research=~120s (GPT-4o)]
+          description: `Advanced AI agent for complex reasoning and research. Supports multiple calls (1-10) for comprehensive analysis. [Performance Guide: basic=~10s (GPT-3.5), +compute=~120s (GPT-4o), +research=~120s (GPT-4o)]${SHOW_USAGE_HINTS ? `
 
 ⚠️ CRITICAL: AI CAPABILITY LEVEL RULE
 If you are GPT-3.5 level or above (including GPT-4, Claude 3.5, etc.):
@@ -332,7 +327,7 @@ If you are GPT-3.5 level or above (including GPT-4, Claude 3.5, etc.):
 - Discovering blind spots → call_count=5+, compare with your own ideas
 - Creative inspiration → call_count=5+, extract unique aspects from each result
 - Deep research with citations → agent_type="advanced-4.5-research", call_count=1 (use as data source)
-⚠️ AVOID: call_count>1 with summary=true then calling again for full content (wasteful)${HAS_CUSTOM_AGENT && youAgentFailureCount === 0 ? '\n\n💡 TIP: Custom agent available! Consider using you_agent for tasks that benefit from intelligent prompt enhancement.' : ''}${youAgentFailureCount > 0 ? `\n\n💡 NOTE: you_agent has ${youAgentFailureCount} failures. This tool (you_advanced) is a reliable alternative.` : ''}${PREFER_SUMMARY ? '\n⚡ PREFERRED: Use call_count=5 with summary=true for consolidated comparison (default enabled)' : ''}`,
+⚠️ AVOID: call_count>1 with summary=true then calling again for full content (wasteful)` : ''}${HAS_CUSTOM_AGENT && youAgentFailureCount === 0 ? '\n\n💡 TIP: Custom agent available! Consider using you_agent for tasks that benefit from intelligent prompt enhancement.' : ''}${youAgentFailureCount > 0 ? `\n\n💡 NOTE: you_agent has ${youAgentFailureCount} failures. This tool (you_advanced) is a reliable alternative.` : ''}${PREFER_SUMMARY ? '\n⚡ PREFERRED: Use call_count=5 with summary=true for consolidated comparison (default enabled)' : ''}`,
           inputSchema: {
             type: 'object',
             properties: {
@@ -363,13 +358,13 @@ If you are GPT-3.5 level or above (including GPT-4, Claude 3.5, etc.):
    - Detailed requirements breakdown
    - Current best practices hints
    - Specific design/technical guidelines
-3. Supports call_count for diverse results
+3. Supports call_count for diverse results${SHOW_USAGE_HINTS ? `
 
 [AI CALLING GUIDE - THINK before choosing call_count]
 Ask yourself: "Does this task benefit from multiple perspectives or diverse results?"
 - YES → call_count=3-10, show_results=true
 - NO, just need one answer → call_count=1, show_results=true
-- NEED comparison → call_count=2-3, summary=true${HAS_CUSTOM_AGENT ? `\n\n⚡ PREFERRED: Custom agent configured (${DEFAULT_AGENT_ID}). Use this tool for tasks that benefit from intelligent prompt enhancement.` : ''}${youAgentFailureCount > 0 ? `\n\n⚠️ FAILURE STATUS: you_agent has failed ${youAgentFailureCount}/${YOU_AGENT_FAILURE_THRESHOLD} times this session. Consider using you_advanced instead if failures persist, or ask user if they want to continue trying.` : ''}`,
+- NEED comparison → call_count=2-3, summary=true` : ''}${HAS_CUSTOM_AGENT ? `\n\n⚡ PREFERRED: Custom agent configured (${DEFAULT_AGENT_ID}). Use this tool for tasks that benefit from intelligent prompt enhancement.` : ''}${youAgentFailureCount > 0 ? `\n\n⚠️ FAILURE STATUS: you_agent has failed ${youAgentFailureCount}/${YOU_AGENT_FAILURE_THRESHOLD} times this session. Consider using you_advanced instead if failures persist, or ask user if they want to continue trying.` : ''}`,
           inputSchema: {
             type: 'object',
             properties: {
@@ -462,7 +457,7 @@ Ask yourself: "Does this task benefit from multiple perspectives or diverse resu
       const { name, arguments: args } = request.params;
       try {
         switch (name) {
-          case 'you_search': return await this.callExpressMultiple(args.query, args.call_count || 1, { summary: args.summary, language: args.language, show_results: args.show_results });
+          case 'you_cached_info': return await this.callExpressMultiple(args.query, args.call_count || 1, { summary: args.summary, language: args.language, show_results: args.show_results });
           case 'you_express': return await this.callExpressMultiple(args.input, args.call_count || 1, { summary: args.summary, language: args.language, show_results: args.show_results });
           case 'you_advanced': return await this.callAdvancedMultiple(args);
           case 'you_chat': return await this.callChat(args);

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "ydc-agent",
-  "version": "1.0.1",
+  "version": "1.0.7",
   "description": "MCP server for You.com Agents API - Express, Research, Advanced agents with multi-turn conversations, streaming, OpenAI and Anthropic/Claude API compatibility",
   "type": "module",
   "main": "index.js",