@aibuilders/mcp-coach-server 1.0.2 → 1.0.4

package/README.dev.md

@@ -0,0 +1,71 @@
+# AI Builder MCP (Developer Guide)
+
+This document is for internal development and maintenance. It includes local setup, testing with MCP Inspector, and build instructions.
+
+## Quick Start (Development)
+
+```bash
+# Clone
+git clone <repository-url>
+cd student_portal_mcp
+
+# Install dependencies
+npm install
+
+# Environment variables
+cp .env.example .env
+# Edit .env and set AI_BUILDER_TOKEN
+
+# Run in dev mode
+npm run dev
+```
+
+## MCP Inspector Testing
+
+```bash
+# Install Inspector
+npm install -g @modelcontextprotocol/inspector
+
+# Run Inspector (in another terminal)
+mcp-inspector
+
+# In Inspector UI, connect using either:
+# Option 1 (recommended): npx tsx src/index.ts
+# Option 2: node dist/index.js
+```
+
+## Tool Usage Examples
+
+```json
+{
+  "tool": "get_api_specification",
+  "arguments": {}
+}
+```
+
+```json
+{
+  "tool": "get_auth_token",
+  "arguments": { "masked": true }
+}
+```
+
+```json
+{
+  "tool": "get_deployment_guide",
+  "arguments": { "service_type": "fastapi" }
+}
+```
+
+## Build
+
+```bash
+npm run build
+```
+
+## Publish Checklist
+
+- Ensure `AI_BUILDER_TOKEN` handling is correct
+- Update version in `package.json`
+- `npm publish` (scope: `@aibuilders`)
+

package/README.md

@@ -1,72 +1,42 @@
-# AI Builders MCP Coach Server
+# AI Builder MCP
 
-MCP服务器作为AI部署教练，通过标准化接口向AI传授部署知识、API调用方法和最佳实践。
+An MCP server that provides deployment guidance and API usage support for the AI Builders platform. This README is for end users and focuses on installation and usage, not development.
 
-## 核心功能
-
-- **get_api_specification()**: 获取AI Builders平台的OpenAPI规范和endpoint信息
-- **get_deployment_guide()**: 获取部署指导（带缓存机制）
-- **explain_authentication_model()**: 详细解释认证机制和使用最佳实践
-- **get_auth_token()**: 返回环境中的 AI_BUILDER_TOKEN（默认打码）
-- **create_env_file()**: 在指定目录创建包含 AI_BUILDER_TOKEN 的 .env 文件
-
-## 本地开发和使用
-
-### 快速开始（开发模式）
+## Install
 
 ```bash
-# 1. 克隆项目
-git clone <repository-url>
-cd student_portal_mcp
-
-# 2. 安装依赖
-npm install
-
-# 3. 配置环境变量
-cp .env.example .env
-# 编辑 .env 文件，添加你的 AI_BUILDER_TOKEN
-
-# 4. 运行开发服务器
-npm run dev
+npm install -g @aibuilders/mcp-coach-server
 ```
 
-### 使用MCP Inspector测试
+You can also run without a global install:
 
 ```bash
-# 安装Inspector
-npm install -g @modelcontextprotocol/inspector
-
-# 运行Inspector（在另一个终端）
-mcp-inspector
-
-# 在Inspector界面中，选择以下任一方式：
-# 方式1（推荐）: npx tsx /Users/grapeot/co/student_portal_mcp/src/index.ts
-# 方式2: node /Users/grapeot/co/student_portal_mcp/dist/index.js
-# 然后点击 Connect
+npx @aibuilders/mcp-coach-server
 ```
 
-### 正式发布后使用
+## Run
 
-等发布到npm后，可以这样使用：
+The server reads `AI_BUILDER_TOKEN` from your environment.
 
 ```bash
-# 安装
-npm install -g @ai-builders/mcp-coach-server
+AI_BUILDER_TOKEN=your_token_here mcp-coach-server
+```
 
-# 运行
-npx @ai-builders/mcp-coach-server
+Or with `npx`:
+
+```bash
+AI_BUILDER_TOKEN=your_token_here npx mcp-coach-server
 ```
 
-### 配置到AI编辑器
+## Configure in MCP Clients
 
-在Claude Desktop配置文件中添加：
+Example (Claude Desktop `mcpServers`):
 
 ```json
 {
   "mcpServers": {
-    "ai-builders-coach": {
-      "command": "npx",
-      "args": ["tsx", "/Users/grapeot/co/student_portal_mcp/src/index.ts"],
+    "ai-builder-mcp": {
+      "command": "mcp-coach-server",
       "env": {
         "AI_BUILDER_TOKEN": "your_token_here"
       }
@@ -75,49 +45,6 @@ npx @ai-builders/mcp-coach-server
 }
 ```
 
-## 工具使用示例
-
-```json
-{
-  "tool": "get_api_specification",
-  "arguments": {}
-}
-```
-
-返回内容为英文，并包含：
-- `endpoint_info.base_url`（自动从 OpenAPI servers 推断）
-- `sdk_compatibility.openai_sdk_compatible`（标注兼容 OpenAI SDK）
-- 示例：`/v1/chat/completions`
+## License
 
-```json
-{
-  "tool": "get_auth_token",
-  "arguments": { "masked": true }
-}
-```
-
-返回打码后的 token；若需要写入 .env：
-
-```json
-{
-  "tool": "create_env_file",
-  "arguments": { "target_dir": "/path/to/project", "overwrite": false }
-}
-```
-
-## 开发
-
-```bash
-# 克隆项目
-git clone <repository-url>
-cd ai-builders-mcp-coach
-
-# 安装依赖
-npm install
-
-# 开发模式运行
-npm run dev
-
-# 构建
-npm run build
-```
+MIT

package/dist/index.js

@@ -65,7 +65,7 @@ async function getCachedDeploymentGuide() {
     catch {
     }
     try {
-        const response = await fetch("https://www.ai-builders.com/resources/students/deployment-prompt.md");
+        const response = await fetch("https://space.ai-builders.com/deployment-prompt.md");
         if (response.ok) {
             const content = await response.text();
             const cacheData = {
@@ -92,8 +92,8 @@ async function getCachedDeploymentGuide() {
     };
 }
 const server = new Server({
-    name: "ai-builders-coach",
-    version: "1.0.0",
+    name: "ai-builder-mcp",
+    version: "1.0.4",
 }, {
     capabilities: {
         tools: {},
@@ -104,20 +104,22 @@ server.setRequestHandler(CallToolRequestSchema, async (request) => {
     try {
         switch (name) {
             case "get_api_specification": {
-                const response = await fetch("https://www.ai-builders.com/resources/students-backend/openapi.json");
+                const response = await fetch("https://space.ai-builders.com/backend/openapi.json");
                 if (!response.ok) {
                     throw new Error(`Failed to fetch OpenAPI specification: HTTP ${response.status}`);
                 }
                 const openapiSpec = await response.json();
-                let baseUrl = "https://www.ai-builders.com/resources/students-backend";
+                let baseUrl = "https://space.ai-builders.com/backend";
                 try {
                     if (openapiSpec?.servers?.length) {
                         const url = openapiSpec.servers[0].url;
                         if (url.startsWith("http")) {
                             baseUrl = url;
                         }
-                        else {
-                            baseUrl = `https://www.ai-builders.com${url}`;
+                        else if (url.startsWith("/")) {
+                            // Relative path: use the backend base URL directly
+                            // Don't append the servers path, use the base URL as-is
+                            baseUrl = "https://space.ai-builders.com/backend";
                         }
                     }
                 }
@@ -138,7 +140,8 @@ server.setRequestHandler(CallToolRequestSchema, async (request) => {
                                 authentication: {
                                     type: "bearer_token",
                                     header: "Authorization: Bearer {AI_BUILDER_TOKEN}",
-                                    token_source: "Environment variable AI_BUILDER_TOKEN"
+                                    token_source: "Environment variable AI_BUILDER_TOKEN",
+                                    setup_instruction: "The AI_BUILDER_TOKEN must be set in a .env file. As an AI assistant, you should proactively help users set up their .env file by calling get_auth_token to retrieve the token and then creating or updating the .env file with AI_BUILDER_TOKEN=<token_value>"
                                 },
                                 sdk_compatibility: {
                                     openai_sdk_compatible: true,
@@ -156,7 +159,24 @@ server.setRequestHandler(CallToolRequestSchema, async (request) => {
                                 mcp_recommendation: {
                                     note: "If running via an MCP server, call get_auth_token to retrieve AI_BUILDER_TOKEN and configure your environment manually",
                                     tools: ["get_auth_token"],
-                                    env_key: "AI_BUILDER_TOKEN"
+                                    env_key: "AI_BUILDER_TOKEN",
+                                    ai_assistant_guidance: {
+                                        proactive_setup: "As an AI assistant, you should proactively help users set up their .env file when they need to use the API. Follow these steps:",
+                                        steps: [
+                                            "1. Call get_auth_token tool to retrieve the AI_BUILDER_TOKEN",
+                                            "2. Check if a .env file exists in the project root",
+                                            "3. Create or update the .env file with: AI_BUILDER_TOKEN=<token_from_get_auth_token>",
+                                            "4. Ensure .env is in .gitignore to prevent committing sensitive tokens",
+                                            "5. Inform the user that the token has been configured and they can now use the API"
+                                        ],
+                                        example_env_content: "AI_BUILDER_TOKEN=your_token_here",
+                                        best_practices: [
+                                            "Always use get_auth_token to retrieve the token rather than asking the user",
+                                            "Create the .env file in the project root directory",
+                                            "Verify .gitignore includes .env before creating the file",
+                                            "Explain to the user what you've done and why it's secure"
+                                        ]
+                                    }
                                 }
                             }, null, 2)
                         }
@@ -308,7 +328,7 @@ server.setRequestHandler(ListToolsRequestSchema, async () => {
 async function main() {
     const transport = new StdioServerTransport();
     await server.connect(transport);
-    console.error("AI Builders MCP Coach Server started");
+    console.error("AI Builder MCP Server started");
 }
 main().catch((error) => {
     console.error("Server failed to start:", error);

package/package.json

@@ -1,7 +1,7 @@
 {
   "name": "@aibuilders/mcp-coach-server",
-  "version": "1.0.2",
-  "description": "MCP server as AI deployment coach for AI Builders platform",
+  "version": "1.0.4",
+  "description": "AI Builder MCP server for AI Builders platform",
   "main": "dist/index.js",
   "type": "module",
   "bin": {
@@ -19,7 +19,6 @@
     "model-context-protocol",
     "ai",
     "deployment",
-    "coach",
     "ai-builders"
   ],
   "author": "AI Builders",