@opencangjie/skills 0.0.37 → 0.0.39

package/CHANGELOG.md

@@ -1,4 +1,4 @@
-# Changelog
+# Changelog
 
 All notable changes to the `@opencangjie/skills` SDK will be documented in this file.
 
@@ -7,7 +7,7 @@ All notable changes to the `@opencangjie/skills` SDK will be documented in this
 ### Fixed
 - **Windows `npx skills stop` command**: Now uses `taskkill /F /PID` on Windows to properly terminate runtime process
 - Previous `process.kill(pid, 'SIGTERM')` didn't work on Windows for detached processes
-- **Runtime v0.0.13**: Temporary directory now uses sibling `temp` directory of `SKILL_INSTALL_PATH`
+- **Runtime v0.0.16**: Temporary directory now uses sibling `temp` directory of `SKILL_INSTALL_PATH`
 - If `SKILL_INSTALL_PATH=./skills`, temp directory will be `./temp/git-clone`
 - This ensures proper permissions and accessibility for the temp directory
 

package/README.md

@@ -32,10 +32,48 @@ npx @opencangjie/skills add ./my-skill
 npx skills install-runtime
 
 # Or specify a version
-npx skills install-runtime --runtime-version 0.0.13
+npx skills install-runtime --runtime-version 0.0.16
 ```
 
-### 2. Start the Runtime
+### 2. Configure the Runtime
+
+Before starting the runtime, you need to configure the AI model API key. The runtime requires an AI model to process skill execution and natural language understanding.
+
+Edit the `.env` file in the runtime directory:
+- **Windows**: `%USERPROFILE%\.agentskills-runtime\<platform>-<arch>\release\.env`
+- **macOS/Linux**: `~/.agentskills-runtime/<platform>-<arch>/release/.env`
+
+Where `<platform>-<arch>` is your system architecture, e.g., `win-x64`, `darwin-x64`, or `linux-x64`
+
+Add your AI model configuration (choose one provider):
+
+```ini
+# Option 1: StepFun (阶跃星辰)
+MODEL_PROVIDER=stepfun
+MODEL_NAME=step-1-8k
+STEPFUN_API_KEY=your_stepfun_api_key_here
+STEPFUN_BASE_URL=https://api.stepfun.com/v1
+
+# Option 2: DeepSeek
+MODEL_PROVIDER=deepseek
+MODEL_NAME=deepseek-chat
+DEEPSEEK_API_KEY=your_deepseek_api_key_here
+
+# Option 3: 华为云 MaaS
+MODEL_PROVIDER=maas
+MAAS_API_KEY=your_maas_api_key_here
+MAAS_BASE_URL=https://api.modelarts-maas.com/v2
+MAAS_MODEL_NAME=qwen3-coder-480b-a35b-instruct
+
+# Option 4: Sophnet
+MODEL_PROVIDER=sophnet
+SOPHNET_API_KEY=your_sophnet_api_key_here
+SOPHNET_BASE_URL=https://www.sophnet.com/api/open-apis/v1
+```
+
+> **Note**: Without proper AI model configuration, the runtime will fail to start with an error like "Get env variable XXX_API_KEY error."
+
+### 3. Start the Runtime
 
 ```bash
 # Start in background (default)
@@ -48,7 +86,7 @@ npx skills start --foreground
 npx skills start --port 3000 --host 0.0.0.0
 ```
 
-### 3. Manage Skills
+### 4. Manage Skills
 
 ```bash
 # Find and install skills
@@ -72,7 +110,7 @@ Download and install the AgentSkills runtime binary.
 
 ```bash
 npx skills install-runtime
-npx skills install-runtime --runtime-version 0.0.13
+npx skills install-runtime --runtime-version 0.0.16
 ```
 
 #### `npx skills start`
@@ -136,11 +174,23 @@ npx skills find skill --source atomgit --limit 5
 Install a skill from GitHub or local path.
 
 ```bash
+# Install from local directory
 npx skills add ./my-skill
+
+# Install from GitHub repository
 npx skills add github.com/user/skill-repo
 npx skills add github.com/user/skill-repo --branch develop
+
+# Install from multi-skill repository (specify subdirectory)
+npx skills add https://github.com/user/skills-repo/tree/main/skills/my-skill
+npx skills add https://atomgit.com/user/skills-repo/tree/main/skills/skill-creator
+
+# Install with options
+npx skills add github.com/user/skill-repo -y  # Skip confirmation
 ```
 
+> **Tip**: For repositories containing multiple skills, use the `/tree/<branch>/<skill-path>` format to specify the exact subdirectory. This avoids the interactive selection prompt.
+
 #### `npx skills list`
 
 List installed skills.
@@ -653,18 +703,26 @@ onMounted(loadSkills);
 // apps/backend/package.json
 {
   "dependencies": {
-    "@opencangjie/skills": "^0.0.13"
+    "@opencangjie/skills": "^0.0.16"
   }
 }
 
 // apps/uctoo-app-client-pc/package.json
 {
   "dependencies": {
-    "@opencangjie/skills": "^0.0.13"
+    "@opencangjie/skills": "^0.0.16"
   }
 }
 ```
 
+## Links
+
+- [NPM Package](https://www.npmjs.com/package/@opencangjie/skills)
+- [Documentation](https://atomgit.com/uctoo/agentskills-runtime#readme)
+- [Repository](https://atomgit.com/uctoo/agentskills-runtime)
+- [Issue Tracker](https://atomgit.com/uctoo/agentskills-runtime/issues)
+- [Python SDK](../python)
+
 ## License
 
 MIT

package/README_cn.md

@@ -32,10 +32,48 @@ npx @opencangjie/skills add ./my-skill
 npx skills install-runtime
 
 # 或指定版本
-npx skills install-runtime --version 0.0.1
+npx skills install-runtime --runtime-version 0.0.16
 ```
 
-### 2. 启动运行时
+### 2. 配置运行时
+
+在启动运行时之前，您需要配置 AI 模型 API 密钥。运行时需要 AI 模型来处理技能执行和自然语言理解。
+
+编辑运行时目录中的 `.env` 文件：
+- **Windows**: `%USERPROFILE%\.agentskills-runtime\<platform>-<arch>\release\.env`
+- **macOS/Linux**: `~/.agentskills-runtime/<platform>-<arch>/release/.env`
+
+其中 `<platform>-<arch>` 是您的系统架构，例如 `win-x64`、`darwin-x64` 或 `linux-x64`
+
+添加您的 AI 模型配置（选择一个提供商）：
+
+```ini
+# 选项 1: StepFun (阶跃星辰)
+MODEL_PROVIDER=stepfun
+MODEL_NAME=step-1-8k
+STEPFUN_API_KEY=your_stepfun_api_key_here
+STEPFUN_BASE_URL=https://api.stepfun.com/v1
+
+# 选项 2: DeepSeek
+MODEL_PROVIDER=deepseek
+MODEL_NAME=deepseek-chat
+DEEPSEEK_API_KEY=your_deepseek_api_key_here
+
+# 选项 3: 华为云 MaaS
+MODEL_PROVIDER=maas
+MAAS_API_KEY=your_maas_api_key_here
+MAAS_BASE_URL=https://api.modelarts-maas.com/v2
+MAAS_MODEL_NAME=qwen3-coder-480b-a35b-instruct
+
+# 选项 4: Sophnet
+MODEL_PROVIDER=sophnet
+SOPHNET_API_KEY=your_sophnet_api_key_here
+SOPHNET_BASE_URL=https://www.sophnet.com/api/open-apis/v1
+```
+
+> **注意**：如果没有正确配置 AI 模型，运行时将无法启动，并显示类似 "Get env variable XXX_API_KEY error" 的错误。
+
+### 3. 启动运行时
 
 ```bash
 # 后台启动（默认）
@@ -48,7 +86,7 @@ npx skills start --foreground
 npx skills start --port 3000 --host 0.0.0.0
 ```
 
-### 3. 管理技能
+### 4. 管理技能
 
 ```bash
 # 查找并安装技能
@@ -72,7 +110,7 @@ npx skills run my-skill -p '{"input": "data"}'
 
 ```bash
 npx skills install-runtime
-npx skills install-runtime --runtime-version 0.0.13
+npx skills install-runtime --runtime-version 0.0.16
 ```
 
 #### `npx skills start`
@@ -136,11 +174,23 @@ npx skills find skill --source atomgit --limit 5
 从 GitHub 或本地路径安装技能。
 
 ```bash
+# 从本地目录安装
 npx skills add ./my-skill
+
+# 从 GitHub 仓库安装
 npx skills add github.com/user/skill-repo
 npx skills add github.com/user/skill-repo --branch develop
+
+# 从多技能仓库安装（指定子目录）
+npx skills add https://github.com/user/skills-repo/tree/main/skills/my-skill
+npx skills add https://atomgit.com/user/skills-repo/tree/main/skills/skill-creator
+
+# 带选项安装
+npx skills add github.com/user/skill-repo -y  # 跳过确认
 ```
 
+> **提示**：对于包含多个技能的仓库，使用 `/tree/<分支>/<技能路径>` 格式指定具体的子目录。这样可以避免交互式选择提示。
+
 #### `npx skills list`
 
 列出已安装的技能。
@@ -653,18 +703,26 @@ onMounted(loadSkills);
 // apps/backend/package.json
 {
   "dependencies": {
-    "@opencangjie/skills": "^0.0.13"
+    "@opencangjie/skills": "^0.0.16"
   }
 }
 
 // apps/uctoo-app-client-pc/package.json
 {
   "dependencies": {
-    "@opencangjie/skills": "^0.0.13"
+    "@opencangjie/skills": "^0.0.16"
   }
 }
 ```
 
+## 相关链接
+
+- [NPM 包](https://www.npmjs.com/package/@opencangjie/skills)
+- [文档](https://atomgit.com/uctoo/agentskills-runtime#readme)
+- [仓库](https://atomgit.com/uctoo/agentskills-runtime)
+- [问题追踪](https://atomgit.com/uctoo/agentskills-runtime/issues)
+- [Python SDK](../python)
+
 ## 许可证
 
 MIT

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@opencangjie/skills",
-  "version": "0.0.37",
+  "version": "0.0.39",
   "description": "JavaScript/TypeScript SDK for AgentSkills Runtime - Install, manage, and execute AI agent skills with built-in runtime support",
   "main": "dist/index.js",
   "types": "dist/index.d.ts",
@@ -89,7 +89,7 @@
     }
   },
   "runtime": {
-    "version": "0.0.15",
+    "version": "0.0.17",
     "downloadUrl": "https://atomgit.com/UCToo/agentskills-runtime/releases",
     "platforms": {
       "win32-x64": "agentskills-runtime-win-x64.tar.gz",

package/scripts/runtime.js

@@ -53,7 +53,8 @@ function getPlatform() {
 }
 
 function getRuntimeDir() {
-  return path.join(__dirname, '..', 'runtime');
+  const homeDir = os.homedir();
+  return path.join(homeDir, '.agentskills-runtime');
 }
 
 function getRuntimePath() {
@@ -142,6 +143,27 @@ async function downloadRuntime(version = RUNTIME_VERSION) {
         fs.chmodSync(runtimePath, '755');
       }
       
+      // Handle .env file - preserve existing configuration
+      const releaseDir = getRuntimeWorkingDir();
+      const envFile = path.join(releaseDir, '.env');
+      const envExampleFile = path.join(releaseDir, '.env.example');
+      
+      if (!fs.existsSync(envFile) && fs.existsSync(envExampleFile)) {
+        fs.copyFileSync(envExampleFile, envFile);
+        console.log('Created .env file from .env.example');
+      } else if (!fs.existsSync(envFile)) {
+        const defaultEnvContent = `# AgentSkills Runtime Configuration
+# This file was auto-generated. Edit as needed.
+
+# Skill Installation Path
+SKILL_INSTALL_PATH=./skills
+`;
+        fs.writeFileSync(envFile, defaultEnvContent);
+        console.log('Created default .env file');
+      } else {
+        console.log('Preserved existing .env file');
+      }
+      
       console.log(`AgentSkills Runtime v${version} downloaded successfully from ${mirror.name}!`);
       return true;
     } catch (error) {