@isnow890/naver-search-mcp 1.0.46 → 1.0.48

package/README.md

@@ -3,79 +3,31 @@
 [![한국어](https://img.shields.io/badge/한국어-README-yellow)](README-ko.md)
 
 [![Trust Score](https://archestra.ai/mcp-catalog/api/badge/quality/isnow890/naver-search-mcp)](https://archestra.ai/mcp-catalog/isnow890__naver-search-mcp)
-[![smithery badge](https://smithery.ai/badge/@isnow890/naver-search-mcp)](https://smithery.ai/server/@isnow890/naver-search-mcp)
 [![MCP.so](https://img.shields.io/badge/MCP.so-Naver%20Search%20MCP-blue)](https://mcp.so/server/naver-search-mcp/isnow890)
 
-#### Version History
+MCP server for Naver Search API and DataLab API integration, enabling comprehensive search across various Naver services and data trend analysis.
 
-###### 1.0.45 (2025-09-28)
+## Available on ClawHub for OpenClaw
 
-- Smithery compatibility issues resolved - Now available through Smithery with latest features
-- Excel compatibility issues fixed in category search - Replaced with JSON functionality
-- Web Korean search (`search_webkr`) functionality restored
-- Full compatibility with Smithery platform installations
+You can install this MCP server as an OpenClaw skill from ClawHub:
 
-###### 1.0.44 (2025-08-31)
-
-- `get_current_korean_time` tool added - Essential time context tool for Korean timezone
-- Enhanced all existing tool descriptions to reference time tool for temporal queries
-- Improved temporal context handling for "today", "now", "current" searches
-- Comprehensive Korean time formatting with multiple output formats
-
-###### 1.0.4 (2025-08-21)
-
-- `find_category` tool added - with fuzzy matching and ranking system support
-- Enhanced parameter validation with Zod schema
-- Improved category search workflow
-
-###### 1.0.30 (2025-08-04)
-
-- MCP SDK upgraded to 1.17.1
-- Fixed compatibility issues with Smithery specification changes
-- Added comprehensive DataLab shopping category code documentation
-
-###### 1.0.2 (2025-04-26)
-
-- README updated: cafe article search tool and version history section improved
-
-###### 1.0.1 (2025-04-26)
-
-- Cafe article search feature added
-- Shopping category info added to zod
-- Source code refactored
-
-###### 1.0.0 (2025-04-08)
-
-- Initial release
-
-#### Prerequisites
+```bash
+openclaw skills install naver-search-mcp
+```
 
-- Naver Developers API Key (Client ID and Secret)
-- Node.js 18 or higher
-- NPM 8 or higher
-- Docker (optional, for container deployment)
+The ClawHub skill uses the published npm package internally and requires `NAVER_CLIENT_ID` and `NAVER_CLIENT_SECRET` in your OpenClaw environment.
 
-#### Getting API Keys
+## Quick Start: Use Without API Key
 
-1. Visit [Naver Developers](https://developers.naver.com/apps/#/register)
-2. Click "Register Application"
-3. Enter application name and select ALL of the following APIs:
-   - Search (for blog, news, book search, etc.)
-   - DataLab (Search Trends)
-   - DataLab (Shopping Insight)
-4. Set the obtained Client ID and Client Secret as environment variables
+You can use this server immediately without API keys through [Kakao PlayMCP](https://playmcp.kakao.com/mcp/154). Simply visit the link and start using it right away!
 
 ## Tool Details
 
 ### Available tools:
 
-#### 🕐 Time & Context Tools
-
-- **get_current_korean_time**: Get current Korean time (KST) with comprehensive date/time information. Essential for understanding "today", "now", or "current" context in Korean timezone. Always use this tool when temporal context is needed for searches or analysis.
-
 #### 🆕 Category Search
 
-- **find_category**: Category search tool - No more need to manually check category numbers via URL for trend and shopping insight searches. The LLM will find it out as you say.
+- **find_category**: Category search tool so you no longer need to manually check category numbers in URLs for trend and shopping insight searches. Just describe the category in natural language.
 
 #### Search Tools
 
@@ -103,106 +55,26 @@
 - **datalab_shopping_keyword_by_gender**: Analyze shopping keyword trends by gender
 - **datalab_shopping_keyword_by_age**: Analyze shopping keyword trends by age group
 
-#### Complete Category List:
-
-For a complete list of category codes, you can download from Naver Shopping Partner Center or extract them by browsing Naver Shopping categories.
-
-### 🎯 Business Use Cases & Scenarios
-
-#### 🛍️ E-commerce Market Research
-
-```javascript
-// Fashion trend discovery
-find_category("fashion") → Check top fashion categories and codes
-datalab_shopping_category → Analyze seasonal fashion trends
-datalab_shopping_age → Identify fashion target demographics
-datalab_shopping_keywords → Compare "dress" vs "jacket" vs "coat"
-```
-
-#### 📱 Digital Marketing Strategy
-
-```javascript
-// Beauty industry analysis
-find_category("cosmetics") → Find beauty categories
-datalab_shopping_gender → 95% female vs 5% male shoppers
-datalab_shopping_device → Mobile dominance in beauty shopping
-datalab_shopping_keywords → "tint" vs "lipstick" keyword performance
-```
-
-#### 🏢 Business Intelligence & Competitive Analysis
-
-```javascript
-// Tech product insights
-find_category("smartphone") → Check electronics categories
-datalab_shopping_category → Track iPhone vs Galaxy trends
-datalab_shopping_age → 20-30s as main smartphone buyers
-datalab_shopping_device → PC vs mobile shopping behavior
-```
-
-#### 📊 Seasonal Business Planning
-
-```javascript
-// Holiday shopping analysis
-find_category("gift") → Gift categories
-datalab_shopping_category → Black Friday, Christmas trends
-datalab_shopping_keywords → "Mother's Day gift" vs "birthday gift"
-datalab_shopping_age → Age-based gift purchasing patterns
-```
-
-#### 🎯 Customer Persona Development
-
-```javascript
-// Fitness market analysis
-find_category("exercise") → Sports/fitness categories
-datalab_shopping_gender → Male vs female fitness spending
-datalab_shopping_age → Primary fitness demographics (20-40s)
-datalab_shopping_keywords → "home workout" vs "gym" trend analysis
-```
-
-### 📈 Advanced Analysis Scenarios
-
-#### Market Entry Strategy
-
-1. **Category Discovery**: Use `find_category` to explore market segments
-2. **Trend Analysis**: Identify growing vs declining categories
-3. **Demographic Targeting**: Age/gender analysis for customer targeting
-4. **Competitive Intelligence**: Keyword performance comparison
-5. **Device Strategy**: Mobile vs PC shopping optimization
-
-#### Product Launch Planning
-
-1. **Market Validation**: Category growth trends and seasonality
-2. **Target Customers**: Demographic analysis for product positioning
-3. **Marketing Channels**: Device preferences for advertising strategy
-4. **Competitive Landscape**: Keyword competition and opportunities
-5. **Pricing Strategy**: Category performance and price correlation
-
-#### Performance Monitoring
-
-1. **Category Health**: Monitor product category trends
-2. **Keyword Tracking**: Track brand and product keyword performance
-3. **Demographic Shifts**: Monitor changing customer demographics
-4. **Seasonal Patterns**: Plan inventory and marketing campaigns
-5. **Competitive Benchmarking**: Compare performance against category averages
-
-### Quick Reference: Popular Category Codes
-
-| Category            | Code     | Korean        |
-| ------------------- | -------- | ------------- |
-| Fashion/Clothing    | 50000000 | 패션의류      |
-| Cosmetics/Beauty    | 50000002 | 화장품/미용   |
-| Digital/Electronics | 50000003 | 디지털/가전   |
-| Sports/Leisure      | 50000004 | 스포츠/레저   |
-| Food/Beverages      | 50000008 | 식품/음료     |
-| Health/Medical      | 50000009 | 건강/의료용품 |
-
-💡 **Tip**: Use `find_category` with fuzzy searches like "beauty", "fashion", "electronics" to easily find categories.
+## Getting API Keys
+
+1. Visit [Naver Developers](https://developers.naver.com/apps/#/register) and log in with your Naver account
+2. Click the "Application Registration" (애플리케이션 등록) button
+3. Fill in the application information:
+   - **Application Name**: Enter any name (e.g., "Naver Search MCP")
+   - **Usage**: Select "Search" (검색)
+4. In the API Settings section, check ALL of the following APIs:
+   - **Search** (검색) - Required for blog, news, book, cafe article, web, image, kin, encyclopedia, academic, and local search
+   - **DataLab - Search Trends** (데이터랩 - 검색어 트렌드) - Required for search term trend analysis
+   - **DataLab - Shopping Insight** (데이터랩 - 쇼핑인사이트) - Required for shopping trend analysis
+5. Click "Register" to complete registration
+6. After registration, you'll see your **Client ID** and **Client Secret** on the application detail page
+7. Use these credentials in the configuration below
 
 ## Installation
 
 ### Method 1: NPX Installation (Recommended)
 
-The most reliable way to use this MCP server is through direct NPX installation. For detailed package information, see the [NPM package page](https://www.npmjs.com/package/@isnow890/naver-search-mcp).
+The most reliable way to use this MCP server is through NPX. For detailed package information, see the [NPM package page](https://www.npmjs.com/package/@isnow890/naver-search-mcp).
 
 #### Claude Desktop Configuration
 
@@ -223,9 +95,9 @@ Add to Claude Desktop config file (`%APPDATA%\Claude\claude_desktop_config.json`
 }
 ```
 
-#### Cursor AI Configuration
+#### Claude Code Configuration
 
-Add to `mcp.json`:
+Add to your Claude Code settings:
 
 ```json
 {
@@ -242,38 +114,21 @@ Add to `mcp.json`:
 }
 ```
 
-### Method 2: Smithery Installation (Alternative - Known Issues)
-
-⚠️ **Important Notice**: Smithery installation may experience connection timeouts and hanging issues due to WebSocket relay infrastructure problems. This is a known Smithery platform issue, not a problem with this MCP server code. **NPX installation (Method 1) is strongly recommended for reliable operation.**
-
-#### Known Issues with Smithery:
-- Server initialization hangs or times out
-- `Error -32001: Request timed out`
-- WebSocket connection drops after handshake
-- Server shuts down unexpectedly before processing requests
-
-#### If you still want to try Smithery:
+### Method 2: ClawHub Installation for OpenClaw
 
-##### For Claude Desktop:
+This MCP server can also be used from OpenClaw through the ClawHub skill wrapper. Install it with:
 
 ```bash
-npx -y @smithery/cli@latest install @isnow890/naver-search-mcp --client claude
+openclaw skills install naver-search-mcp
 ```
 
-##### For other AI clients:
+The ClawHub skill uses the same published npm package internally:
 
 ```bash
-# Cursor
-npx -y @smithery/cli@latest install @isnow890/naver-search-mcp --client cursor
-
-# Windsurf
-npx -y @smithery/cli@latest install @isnow890/naver-search-mcp --client windsurf
-
-# Cline
-npx -y @smithery/cli@latest install @isnow890/naver-search-mcp --client cline
+npx -y @isnow890/naver-search-mcp
 ```
 
-**If you experience timeout issues with Smithery, please switch to Method 1 (NPX) for stable operation.**
+Make sure `NAVER_CLIENT_ID` and `NAVER_CLIENT_SECRET` are configured in your OpenClaw environment before using the skill.
 
 ### Method 3: Local Installation
 
@@ -292,7 +147,7 @@ npm run build
 
 ##### Or Download ZIP File
 
-1. Download the latest version from [GitHub Releases](https://github.com/isnow890/naver-search-mcp/releases)
+1. Download the latest version from [GitHub Releases](https://github.com/isnow890/naver-search-mcp)
 2. Extract the ZIP file to your desired location
 3. Navigate to the extracted folder in terminal:
 
@@ -382,51 +237,71 @@ pwd
 
 After completing the configuration, completely close and restart Claude Desktop to activate the Naver Search MCP server.
 
+## Prerequisites
+
+- Naver Developers API Key (Client ID and Secret)
+- Node.js 18 or higher
+- NPM 8 or higher
+
+## License
+
+MIT License
+
 ---
 
-## Alternative Installation Methods
+## Version History
 
-### Method 4: Docker Installation
+### 1.0.48 (2026-05-12)
 
-For containerized deployment:
+- Fixed `find_category` data loading when the MCP server is launched by `npx` from a different working directory
+- Category data is now resolved relative to the installed package before falling back to local development paths
 
-```bash
-docker run -i --rm \
-  -e NAVER_CLIENT_ID=your_client_id \
-  -e NAVER_CLIENT_SECRET=your_client_secret \
-  mcp/naver-search
-```
+### 1.0.47 (2025-01-03)
 
-Docker configuration for Claude Desktop:
+- **Added "today" keyword support** for all DataLab date parameters - no need to call separate time tool
+- **Fixed server termination issue** - MCP server now properly exits when client disconnects
+- **Added graceful shutdown handlers** for SIGINT, SIGTERM, and transport close events
+- **Removed get_current_korean_time tool** - redundant with new "today" keyword feature
+- **Removed memory monitoring module** - resolved setInterval blocking process exit
+- **Special thanks to @gloomyrobot** for reporting the server termination issue
 
-```json
-{
-  "mcpServers": {
-    "naver-search": {
-      "command": "docker",
-      "args": [
-        "run",
-        "-i",
-        "--rm",
-        "-e",
-        "NAVER_CLIENT_ID=your_client_id",
-        "-e",
-        "NAVER_CLIENT_SECRET=your_client_secret",
-        "mcp/naver-search"
-      ]
-    }
-  }
-}
-```
+### 1.0.45 (2025-09-28)
 
-## Build
+- Resolved platform compatibility issues for hosted MCP installations
+- Replaced the Excel export in category search with JSON for better compatibility
+- Restored the `search_webkr` tool for Korean web search
+- Improved hosted platform installation compatibility
 
-Docker build:
+### 1.0.44 (2025-08-31)
 
-```bash
-docker build -t mcp/naver-search .
-```
+- Added the `get_current_korean_time` tool for essential Korea Standard Time context
+- Referenced the time tool across existing tool descriptions for temporal queries
+- Improved handling of "today", "now", and "current" searches with temporal context
+- Expanded Korean date and time formatting outputs with multiple formats
 
-## License
+### 1.0.40 (2025-08-21)
 
-MIT License
+- Added the `find_category` tool with fuzzy matching so you no longer need to check category numbers manually in URLs
+- Enhanced parameter validation with Zod schema
+- Improved the category search workflow
+- Implemented a level-based category ranking system that prioritizes top-level categories
+
+### 1.0.30 (2025-08-04)
+
+- MCP SDK upgraded to 1.17.1
+- Fixed compatibility issues with hosted MCP platform specification changes
+- Added comprehensive DataLab shopping category code documentation
+
+### 1.0.2 (2025-04-26)
+
+- README updated: cafe article search tool and version history section improved
+
+### 1.0.1 (2025-04-26)
+
+- Cafe article search feature added
+- Shopping category info added to zod
+- Source code refactored
+
+### 1.0.0 (2025-04-08)
+
+- Initial release

package/dist/src/handlers/category.handlers.js

@@ -1,8 +1,10 @@
 // Import JSON data - bundle-safe approach
 import { existsSync, readFileSync } from 'fs';
-import { join } from 'path';
+import { dirname, join } from 'path';
+import { fileURLToPath } from 'url';
 // 메모리 누수 방지를 위한 지연 로딩 캐시
 let categoriesCache = null;
+const moduleDir = dirname(fileURLToPath(import.meta.url));
 // Load categories data with fallback paths for different environments
 function getCategoriesData() {
     // 캐시된 데이터가 있으면 반환 (메모리 효율성)
@@ -10,19 +12,21 @@ function getCategoriesData() {
         return categoriesCache;
     }
     try {
-        // Try bundled data path first (dist/data)
-        const bundledPath = join(process.cwd(), 'dist', 'data', 'categories.json');
-        if (existsSync(bundledPath)) {
-            categoriesCache = JSON.parse(readFileSync(bundledPath, 'utf8'));
-            return categoriesCache;
-        }
-        // Fallback to source data path
-        const sourcePath = join(process.cwd(), 'data', 'categories.json');
-        if (existsSync(sourcePath)) {
-            categoriesCache = JSON.parse(readFileSync(sourcePath, 'utf8'));
-            return categoriesCache;
+        const candidatePaths = [
+            // Package-relative path. Works when launched by npx from any cwd.
+            join(moduleDir, '..', '..', 'data', 'categories.json'),
+            // Local development fallbacks.
+            join(process.cwd(), 'dist', 'data', 'categories.json'),
+            join(process.cwd(), 'data', 'categories.json'),
+        ];
+        for (const categoryPath of candidatePaths) {
+            if (existsSync(categoryPath)) {
+                const categoryJson = readFileSync(categoryPath, 'utf8').toString();
+                categoriesCache = JSON.parse(categoryJson);
+                return categoriesCache;
+            }
         }
-        throw new Error('카테고리 데이터 파일을 찾을 수 없습니다');
+        throw new Error(`카테고리 데이터 파일을 찾을 수 없습니다. Tried: ${candidatePaths.join(', ')}`);
     }
     catch (error) {
         console.error('카테고리 데이터 로딩 실패:', error);

package/dist/src/handlers/datalab.handlers.js

@@ -1,4 +1,5 @@
 import { NaverSearchClient } from "../clients/naver-search.client.js";
+import { resolveDateRange } from "../utils/date.utils.js";
 // 클라이언트 인스턴스 (싱글톤)
 const client = NaverSearchClient.getInstance();
 /**
@@ -50,7 +51,8 @@ export const datalabToolHandlers = {
  * @param params DatalabSearch
  */
 export async function handleSearchTrend(params) {
-    return client.searchTrend(params);
+    const { startDate, endDate } = resolveDateRange(params.startDate, params.endDate);
+    return client.searchTrend({ ...params, startDate, endDate });
 }
 /**
  * 쇼핑 카테고리 트렌드 핸들러
@@ -58,7 +60,8 @@ export async function handleSearchTrend(params) {
  * @param params DatalabShopping
  */
 export async function handleShoppingCategoryTrend(params) {
-    return client.datalabShoppingCategory(params);
+    const { startDate, endDate } = resolveDateRange(params.startDate, params.endDate);
+    return client.datalabShoppingCategory({ ...params, startDate, endDate });
 }
 /**
  * 쇼핑 기기별 트렌드 핸들러
@@ -66,9 +69,10 @@ export async function handleShoppingCategoryTrend(params) {
  * @param params DatalabShoppingDevice
  */
 export async function handleShoppingByDeviceTrend(params) {
+    const { startDate, endDate } = resolveDateRange(params.startDate, params.endDate);
     return client.datalabShoppingByDevice({
-        startDate: params.startDate,
-        endDate: params.endDate,
+        startDate,
+        endDate,
         timeUnit: params.timeUnit,
         category: params.category,
         device: params.device,
@@ -80,9 +84,10 @@ export async function handleShoppingByDeviceTrend(params) {
  * @param params DatalabShoppingGender
  */
 export async function handleShoppingByGenderTrend(params) {
+    const { startDate, endDate } = resolveDateRange(params.startDate, params.endDate);
     return client.datalabShoppingByGender({
-        startDate: params.startDate,
-        endDate: params.endDate,
+        startDate,
+        endDate,
         timeUnit: params.timeUnit,
         category: params.category,
         gender: params.gender,
@@ -94,9 +99,10 @@ export async function handleShoppingByGenderTrend(params) {
  * @param params DatalabShoppingAge
  */
 export async function handleShoppingByAgeTrend(params) {
+    const { startDate, endDate } = resolveDateRange(params.startDate, params.endDate);
     return client.datalabShoppingByAge({
-        startDate: params.startDate,
-        endDate: params.endDate,
+        startDate,
+        endDate,
         timeUnit: params.timeUnit,
         category: params.category,
         ages: params.ages,
@@ -108,10 +114,10 @@ export async function handleShoppingByAgeTrend(params) {
  * @param params DatalabShoppingKeywords
  */
 export async function handleShoppingKeywordsTrend(params) {
-    // 키워드 배열을 네이버 API에 맞는 형식으로 변환
+    const { startDate, endDate } = resolveDateRange(params.startDate, params.endDate);
     return client.datalabShoppingKeywords({
-        startDate: params.startDate,
-        endDate: params.endDate,
+        startDate,
+        endDate,
         timeUnit: params.timeUnit,
         category: params.category,
         keyword: params.keyword,
@@ -123,9 +129,10 @@ export async function handleShoppingKeywordsTrend(params) {
  * @param params DatalabShoppingKeywordDevice
  */
 export async function handleShoppingKeywordByDeviceTrend(params) {
+    const { startDate, endDate } = resolveDateRange(params.startDate, params.endDate);
     return client.datalabShoppingKeywordByDevice({
-        startDate: params.startDate,
-        endDate: params.endDate,
+        startDate,
+        endDate,
         timeUnit: params.timeUnit,
         category: params.category,
         keyword: params.keyword,
@@ -138,9 +145,10 @@ export async function handleShoppingKeywordByDeviceTrend(params) {
  * @param params DatalabShoppingKeywordGender
  */
 export async function handleShoppingKeywordByGenderTrend(params) {
+    const { startDate, endDate } = resolveDateRange(params.startDate, params.endDate);
     return client.datalabShoppingKeywordByGender({
-        startDate: params.startDate,
-        endDate: params.endDate,
+        startDate,
+        endDate,
         timeUnit: params.timeUnit,
         category: params.category,
         keyword: params.keyword,
@@ -153,9 +161,10 @@ export async function handleShoppingKeywordByGenderTrend(params) {
  * @param params DatalabShoppingKeywordAge
  */
 export async function handleShoppingKeywordByAgeTrend(params) {
+    const { startDate, endDate } = resolveDateRange(params.startDate, params.endDate);
     return client.datalabShoppingKeywordByAge({
-        startDate: params.startDate,
-        endDate: params.endDate,
+        startDate,
+        endDate,
         timeUnit: params.timeUnit,
         category: params.category,
         keyword: params.keyword,

package/dist/src/index.d.ts

@@ -17,7 +17,7 @@ export declare const configSchema: z.ZodObject<{
 export declare function resetServerInstance(): void;
 export declare function createNaverSearchServer({ config, }: {
     config: z.infer<typeof configSchema>;
-}): import("@modelcontextprotocol/sdk/server/index.js").Server<{
+}): McpServer | import("@modelcontextprotocol/sdk/server/index.js").Server<{
     method: string;
     params?: {
         [x: string]: unknown;
@@ -39,5 +39,5 @@ export declare function createNaverSearchServer({ config, }: {
     _meta?: {
         [x: string]: unknown;
     } | undefined;
-}> | McpServer;
+}>;
 export default createNaverSearchServer;