nova-terminal-assistant 0.1.0

package/src/core/config/ConfigManager.ts

@@ -0,0 +1,1219 @@
+// ============================================================================
+// ConfigManager - Loads and manages application configuration
+// ============================================================================
+
+import fs from 'node:fs/promises';
+import path from 'node:path';
+import os from 'node:os';
+import type { NovaConfig, CoreConfig, ModelProviderConfig, ModelConfig } from '../types/config.js';
+import { ConfigError } from '../types/errors.js';
+
+const DEFAULT_CONFIG: NovaConfig = {
+  core: {
+    defaultModel: 'glm-5',
+    defaultApprovalMode: 'default',
+    maxTurns: 100,
+    maxTokens: 8192,
+    temperature: 0.7,
+    contextWindowTarget: 128000,
+    streaming: true,
+    logLevel: 'info',
+    timeout: 300000,
+  },
+  models: {
+    providers: {
+      anthropic: {
+        type: 'anthropic',
+        models: {
+          // Claude 4 Series
+          'claude-sonnet-4-20250514': {
+            name: 'Claude Sonnet 4',
+            maxContextTokens: 200000,
+            maxOutputTokens: 64000,
+            supportsVision: true,
+            supportsTools: true,
+            supportsStreaming: true,
+            supportsThinking: true,
+            inputCostPerMToken: 3,
+            outputCostPerMToken: 15,
+          },
+          'claude-opus-4-20250514': {
+            name: 'Claude Opus 4',
+            maxContextTokens: 200000,
+            maxOutputTokens: 32000,
+            supportsVision: true,
+            supportsTools: true,
+            supportsStreaming: true,
+            supportsThinking: true,
+            inputCostPerMToken: 15,
+            outputCostPerMToken: 75,
+          },
+          // Claude 3.5 Series
+          'claude-3-5-sonnet-20241022': {
+            name: 'Claude 3.5 Sonnet',
+            maxContextTokens: 200000,
+            maxOutputTokens: 8192,
+            supportsVision: true,
+            supportsTools: true,
+            supportsStreaming: true,
+            supportsThinking: false,
+            inputCostPerMToken: 3,
+            outputCostPerMToken: 15,
+          },
+          'claude-3-5-sonnet-20240620': {
+            name: 'Claude 3.5 Sonnet (Jun 2024)',
+            maxContextTokens: 200000,
+            maxOutputTokens: 8192,
+            supportsVision: true,
+            supportsTools: true,
+            supportsStreaming: true,
+            supportsThinking: false,
+            inputCostPerMToken: 3,
+            outputCostPerMToken: 15,
+          },
+          'claude-3-5-haiku-20241022': {
+            name: 'Claude 3.5 Haiku',
+            maxContextTokens: 200000,
+            maxOutputTokens: 8192,
+            supportsVision: true,
+            supportsTools: true,
+            supportsStreaming: true,
+            supportsThinking: false,
+            inputCostPerMToken: 1,
+            outputCostPerMToken: 5,
+          },
+          // Claude 3 Series
+          'claude-3-opus-20240229': {
+            name: 'Claude 3 Opus',
+            maxContextTokens: 200000,
+            maxOutputTokens: 4096,
+            supportsVision: true,
+            supportsTools: true,
+            supportsStreaming: true,
+            supportsThinking: false,
+            inputCostPerMToken: 15,
+            outputCostPerMToken: 75,
+          },
+          'claude-3-sonnet-20240229': {
+            name: 'Claude 3 Sonnet',
+            maxContextTokens: 200000,
+            maxOutputTokens: 4096,
+            supportsVision: true,
+            supportsTools: true,
+            supportsStreaming: true,
+            supportsThinking: false,
+            inputCostPerMToken: 3,
+            outputCostPerMToken: 15,
+          },
+          'claude-3-haiku-20240307': {
+            name: 'Claude 3 Haiku',
+            maxContextTokens: 200000,
+            maxOutputTokens: 4096,
+            supportsVision: true,
+            supportsTools: true,
+            supportsStreaming: true,
+            supportsThinking: false,
+            inputCostPerMToken: 0.25,
+            outputCostPerMToken: 1.25,
+          },
+        },
+        defaultModel: 'claude-3-5-sonnet-20241022',
+      },
+      openai: {
+        type: 'openai',
+        models: {
+          // GPT-4o Series
+          'gpt-4o': {
+            name: 'GPT-4o',
+            maxContextTokens: 128000,
+            maxOutputTokens: 16384,
+            supportsVision: true,
+            supportsTools: true,
+            supportsStreaming: true,
+            supportsThinking: false,
+            inputCostPerMToken: 2.5,
+            outputCostPerMToken: 10,
+          },
+          'gpt-4o-2024-11-20': {
+            name: 'GPT-4o (Nov 2024)',
+            maxContextTokens: 128000,
+            maxOutputTokens: 16384,
+            supportsVision: true,
+            supportsTools: true,
+            supportsStreaming: true,
+            supportsThinking: false,
+            inputCostPerMToken: 2.5,
+            outputCostPerMToken: 10,
+          },
+          'gpt-4o-mini': {
+            name: 'GPT-4o Mini',
+            maxContextTokens: 128000,
+            maxOutputTokens: 16384,
+            supportsVision: true,
+            supportsTools: true,
+            supportsStreaming: true,
+            supportsThinking: false,
+            inputCostPerMToken: 0.15,
+            outputCostPerMToken: 0.6,
+          },
+          // o1 Series (reasoning)
+          'o1': {
+            name: 'o1',
+            maxContextTokens: 200000,
+            maxOutputTokens: 100000,
+            supportsVision: true,
+            supportsTools: false,
+            supportsStreaming: false,
+            supportsThinking: true,
+            inputCostPerMToken: 15,
+            outputCostPerMToken: 60,
+          },
+          'o1-mini': {
+            name: 'o1-mini',
+            maxContextTokens: 128000,
+            maxOutputTokens: 65536,
+            supportsVision: false,
+            supportsTools: false,
+            supportsStreaming: false,
+            supportsThinking: true,
+            inputCostPerMToken: 1.1,
+            outputCostPerMToken: 4.4,
+          },
+          'o3-mini': {
+            name: 'o3-mini',
+            maxContextTokens: 200000,
+            maxOutputTokens: 100000,
+            supportsVision: false,
+            supportsTools: false,
+            supportsStreaming: true,
+            supportsThinking: true,
+            inputCostPerMToken: 1.1,
+            outputCostPerMToken: 4.4,
+          },
+          // GPT-4 Turbo
+          'gpt-4-turbo': {
+            name: 'GPT-4 Turbo',
+            maxContextTokens: 128000,
+            maxOutputTokens: 4096,
+            supportsVision: true,
+            supportsTools: true,
+            supportsStreaming: true,
+            supportsThinking: false,
+            inputCostPerMToken: 10,
+            outputCostPerMToken: 30,
+          },
+        },
+        defaultModel: 'gpt-4o',
+      },
+      google: {
+        type: 'custom',
+        baseUrl: 'https://generativelanguage.googleapis.com/v1beta/openai',
+        models: {
+          'gemini-2.0-flash': {
+            name: 'Gemini 2.0 Flash',
+            maxContextTokens: 1048576,
+            maxOutputTokens: 8192,
+            supportsVision: true,
+            supportsTools: true,
+            supportsStreaming: true,
+            supportsThinking: false,
+            inputCostPerMToken: 0.1,
+            outputCostPerMToken: 0.4,
+          },
+          'gemini-2.0-flash-lite': {
+            name: 'Gemini 2.0 Flash Lite',
+            maxContextTokens: 1048576,
+            maxOutputTokens: 8192,
+            supportsVision: true,
+            supportsTools: true,
+            supportsStreaming: true,
+            supportsThinking: false,
+            inputCostPerMToken: 0.05,
+            outputCostPerMToken: 0.15,
+          },
+          'gemini-2.5-pro-preview-05-06': {
+            name: 'Gemini 2.5 Pro Preview',
+            maxContextTokens: 1048576,
+            maxOutputTokens: 65536,
+            supportsVision: true,
+            supportsTools: true,
+            supportsStreaming: true,
+            supportsThinking: true,
+            inputCostPerMToken: 1.25,
+            outputCostPerMToken: 10,
+          },
+          'gemini-2.5-flash-preview-05-20': {
+            name: 'Gemini 2.5 Flash Preview',
+            maxContextTokens: 1048576,
+            maxOutputTokens: 65536,
+            supportsVision: true,
+            supportsTools: true,
+            supportsStreaming: true,
+            supportsThinking: true,
+            inputCostPerMToken: 0.15,
+            outputCostPerMToken: 0.6,
+          },
+        },
+        defaultModel: 'gemini-2.0-flash',
+      },
+      deepseek: {
+        type: 'custom',
+        baseUrl: 'https://api.deepseek.com',
+        models: {
+          'deepseek-chat': {
+            name: 'DeepSeek V3',
+            maxContextTokens: 65536,
+            maxOutputTokens: 8192,
+            supportsVision: false,
+            supportsTools: true,
+            supportsStreaming: true,
+            supportsThinking: false,
+            inputCostPerMToken: 0.27,
+            outputCostPerMToken: 1.1,
+          },
+          'deepseek-reasoner': {
+            name: 'DeepSeek R1',
+            maxContextTokens: 65536,
+            maxOutputTokens: 8192,
+            supportsVision: false,
+            supportsTools: false,
+            supportsStreaming: true,
+            supportsThinking: true,
+            inputCostPerMToken: 0.55,
+            outputCostPerMToken: 2.19,
+          },
+        },
+        defaultModel: 'deepseek-chat',
+      },
+      ollama: {
+        type: 'ollama',
+        baseUrl: 'http://localhost:11434',
+        models: {
+          'llama3.1': {
+            name: 'Llama 3.1 8B',
+            maxContextTokens: 128000,
+            maxOutputTokens: 4096,
+            supportsVision: false,
+            supportsTools: true,
+            supportsStreaming: true,
+            supportsThinking: false,
+          },
+          'llama3.1:70b': {
+            name: 'Llama 3.1 70B',
+            maxContextTokens: 128000,
+            maxOutputTokens: 4096,
+            supportsVision: false,
+            supportsTools: true,
+            supportsStreaming: true,
+            supportsThinking: false,
+          },
+          'codellama': {
+            name: 'Code Llama',
+            maxContextTokens: 16384,
+            maxOutputTokens: 4096,
+            supportsVision: false,
+            supportsTools: true,
+            supportsStreaming: true,
+            supportsThinking: false,
+          },
+          'mistral': {
+            name: 'Mistral 7B',
+            maxContextTokens: 32768,
+            maxOutputTokens: 4096,
+            supportsVision: false,
+            supportsTools: true,
+            supportsStreaming: true,
+            supportsThinking: false,
+          },
+          'qwen2.5-coder': {
+            name: 'Qwen 2.5 Coder',
+            maxContextTokens: 131072,
+            maxOutputTokens: 8192,
+            supportsVision: false,
+            supportsTools: true,
+            supportsStreaming: true,
+            supportsThinking: false,
+          },
+          'deepseek-coder-v2': {
+            name: 'DeepSeek Coder V2',
+            maxContextTokens: 128000,
+            maxOutputTokens: 4096,
+            supportsVision: false,
+            supportsTools: true,
+            supportsStreaming: true,
+            supportsThinking: false,
+          },
+          'gemma2': {
+            name: 'Gemma 2 9B',
+            maxContextTokens: 8192,
+            maxOutputTokens: 4096,
+            supportsVision: false,
+            supportsTools: true,
+            supportsStreaming: true,
+            supportsThinking: false,
+          },
+        },
+        defaultModel: 'llama3.1',
+      },
+      // ==================== Ollama Cloud ====================
+      'ollama-cloud': {
+        type: 'ollama-cloud',
+        baseUrl: 'https://ollama.com',
+        models: {
+          'deepseek-v3.2': {
+            name: 'DeepSeek V3.2 (Cloud)',
+            maxContextTokens: 131072,
+            maxOutputTokens: 8192,
+            supportsVision: false,
+            supportsTools: true,
+            supportsStreaming: true,
+            supportsThinking: true,
+          },
+          'deepseek-v3.1:671b': {
+            name: 'DeepSeek V3.1 671B (Cloud)',
+            maxContextTokens: 131072,
+            maxOutputTokens: 8192,
+            supportsVision: false,
+            supportsTools: true,
+            supportsStreaming: true,
+            supportsThinking: true,
+          },
+          'kimi-k2.5': {
+            name: 'Kimi K2.5 (Cloud)',
+            maxContextTokens: 131072,
+            maxOutputTokens: 8192,
+            supportsVision: false,
+            supportsTools: true,
+            supportsStreaming: true,
+            supportsThinking: true,
+          },
+          'kimi-k2:1t': {
+            name: 'Kimi K2 1T (Cloud)',
+            maxContextTokens: 131072,
+            maxOutputTokens: 8192,
+            supportsVision: false,
+            supportsTools: true,
+            supportsStreaming: true,
+            supportsThinking: true,
+          },
+          'qwen3-coder:480b': {
+            name: 'Qwen3 Coder 480B (Cloud)',
+            maxContextTokens: 131072,
+            maxOutputTokens: 8192,
+            supportsVision: false,
+            supportsTools: true,
+            supportsStreaming: true,
+            supportsThinking: false,
+          },
+          'qwen3.5:397b': {
+            name: 'Qwen3.5 397B (Cloud)',
+            maxContextTokens: 131072,
+            maxOutputTokens: 8192,
+            supportsVision: false,
+            supportsTools: true,
+            supportsStreaming: true,
+            supportsThinking: false,
+          },
+          'glm-5': {
+            name: 'GLM-5 (Cloud)',
+            maxContextTokens: 128000,
+            maxOutputTokens: 8192,
+            supportsVision: false,
+            supportsTools: true,
+            supportsStreaming: true,
+            supportsThinking: true,
+          },
+          'glm-4.7': {
+            name: 'GLM-4.7 (Cloud)',
+            maxContextTokens: 128000,
+            maxOutputTokens: 8192,
+            supportsVision: false,
+            supportsTools: true,
+            supportsStreaming: true,
+            supportsThinking: false,
+          },
+          'mistral-large-3:675b': {
+            name: 'Mistral Large 3 675B (Cloud)',
+            maxContextTokens: 131072,
+            maxOutputTokens: 8192,
+            supportsVision: false,
+            supportsTools: false,
+            supportsStreaming: true,
+            supportsThinking: false,
+          },
+          'gpt-oss:120b': {
+            name: 'GPT-OSS 120B (Cloud)',
+            maxContextTokens: 131072,
+            maxOutputTokens: 8192,
+            supportsVision: false,
+            supportsTools: false,
+            supportsStreaming: true,
+            supportsThinking: false,
+          },
+          'minimax-m2.7': {
+            name: 'MiniMax M2.7 (Cloud)',
+            maxContextTokens: 131072,
+            maxOutputTokens: 8192,
+            supportsVision: false,
+            supportsTools: false,
+            supportsStreaming: true,
+            supportsThinking: false,
+          },
+        },
+        defaultModel: 'deepseek-v3.2',
+      },
+      // ==================== Chinese Providers ====================
+      qwen: {
+        type: 'custom',
+        baseUrl: 'https://dashscope.aliyuncs.com/compatible-mode/v1',
+        models: {
+          'qwen-max': {
+            name: 'Qwen Max',
+            maxContextTokens: 32768,
+            maxOutputTokens: 8192,
+            supportsVision: true,
+            supportsTools: true,
+            supportsStreaming: true,
+            supportsThinking: false,
+            inputCostPerMToken: 2,
+            outputCostPerMToken: 6,
+          },
+          'qwen-plus': {
+            name: 'Qwen Plus',
+            maxContextTokens: 131072,
+            maxOutputTokens: 8192,
+            supportsVision: false,
+            supportsTools: true,
+            supportsStreaming: true,
+            supportsThinking: false,
+            inputCostPerMToken: 0.8,
+            outputCostPerMToken: 2,
+          },
+          'qwen-turbo': {
+            name: 'Qwen Turbo',
+            maxContextTokens: 131072,
+            maxOutputTokens: 8192,
+            supportsVision: false,
+            supportsTools: true,
+            supportsStreaming: true,
+            supportsThinking: false,
+            inputCostPerMToken: 0.3,
+            outputCostPerMToken: 0.6,
+          },
+          'qwen-coder-plus': {
+            name: 'Qwen Coder Plus',
+            maxContextTokens: 131072,
+            maxOutputTokens: 8192,
+            supportsVision: false,
+            supportsTools: true,
+            supportsStreaming: true,
+            supportsThinking: false,
+            inputCostPerMToken: 1,
+            outputCostPerMToken: 3,
+          },
+          'qwen-vl-max': {
+            name: 'Qwen VL Max',
+            maxContextTokens: 32768,
+            maxOutputTokens: 8192,
+            supportsVision: true,
+            supportsTools: false,
+            supportsStreaming: true,
+            supportsThinking: false,
+            inputCostPerMToken: 2,
+            outputCostPerMToken: 6,
+          },
+        },
+        defaultModel: 'qwen-max',
+      },
+      glm: {
+        type: 'custom',
+        baseUrl: 'https://open.bigmodel.cn/api/paas/v4',
+        models: {
+          'glm-4-plus': {
+            name: 'GLM-4 Plus',
+            maxContextTokens: 128000,
+            maxOutputTokens: 4096,
+            supportsVision: false,
+            supportsTools: true,
+            supportsStreaming: true,
+            supportsThinking: false,
+            inputCostPerMToken: 0.5,
+            outputCostPerMToken: 0.5,
+          },
+          'glm-4-flash': {
+            name: 'GLM-4 Flash',
+            maxContextTokens: 128000,
+            maxOutputTokens: 4096,
+            supportsVision: false,
+            supportsTools: true,
+            supportsStreaming: true,
+            supportsThinking: false,
+            inputCostPerMToken: 0.1,
+            outputCostPerMToken: 0.1,
+          },
+          'glm-4v-plus': {
+            name: 'GLM-4V Plus',
+            maxContextTokens: 8192,
+            maxOutputTokens: 4096,
+            supportsVision: true,
+            supportsTools: false,
+            supportsStreaming: true,
+            supportsThinking: false,
+            inputCostPerMToken: 0.5,
+            outputCostPerMToken: 0.5,
+          },
+          'glm-4-long': {
+            name: 'GLM-4 Long',
+            maxContextTokens: 1000000,
+            maxOutputTokens: 4096,
+            supportsVision: false,
+            supportsTools: false,
+            supportsStreaming: true,
+            supportsThinking: false,
+            inputCostPerMToken: 0.1,
+            outputCostPerMToken: 0.1,
+          },
+        },
+        defaultModel: 'glm-4-plus',
+      },
+      moonshot: {
+        type: 'custom',
+        baseUrl: 'https://api.moonshot.cn/v1',
+        models: {
+          'moonshot-v1-128k': {
+            name: 'Moonshot V1 128K',
+            maxContextTokens: 131072,
+            maxOutputTokens: 4096,
+            supportsVision: false,
+            supportsTools: true,
+            supportsStreaming: true,
+            supportsThinking: false,
+            inputCostPerMToken: 0.6,
+            outputCostPerMToken: 1.6,
+          },
+          'moonshot-v1-32k': {
+            name: 'Moonshot V1 32K',
+            maxContextTokens: 32768,
+            maxOutputTokens: 4096,
+            supportsVision: false,
+            supportsTools: true,
+            supportsStreaming: true,
+            supportsThinking: false,
+            inputCostPerMToken: 0.24,
+            outputCostPerMToken: 0.6,
+          },
+          'moonshot-v1-8k': {
+            name: 'Moonshot V1 8K',
+            maxContextTokens: 8192,
+            maxOutputTokens: 4096,
+            supportsVision: false,
+            supportsTools: true,
+            supportsStreaming: true,
+            supportsThinking: false,
+            inputCostPerMToken: 0.12,
+            outputCostPerMToken: 0.3,
+          },
+        },
+        defaultModel: 'moonshot-v1-128k',
+      },
+      baichuan: {
+        type: 'custom',
+        baseUrl: 'https://api.baichuan-ai.com/v1',
+        models: {
+          'baichuan4': {
+            name: 'Baichuan 4',
+            maxContextTokens: 131072,
+            maxOutputTokens: 4096,
+            supportsVision: false,
+            supportsTools: true,
+            supportsStreaming: true,
+            supportsThinking: false,
+            inputCostPerMToken: 4,
+            outputCostPerMToken: 4,
+          },
+          'baichuan3-turbo': {
+            name: 'Baichuan 3 Turbo',
+            maxContextTokens: 32768,
+            maxOutputTokens: 4096,
+            supportsVision: false,
+            supportsTools: true,
+            supportsStreaming: true,
+            supportsThinking: false,
+            inputCostPerMToken: 0.4,
+            outputCostPerMToken: 0.4,
+          },
+        },
+        defaultModel: 'baichuan4',
+      },
+      minimax: {
+        type: 'custom',
+        baseUrl: 'https://api.minimax.chat/v1',
+        models: {
+          'abab6.5s': {
+            name: 'MiniMax abab6.5s',
+            maxContextTokens: 32768,
+            maxOutputTokens: 8192,
+            supportsVision: false,
+            supportsTools: true,
+            supportsStreaming: true,
+            supportsThinking: false,
+            inputCostPerMToken: 1,
+            outputCostPerMToken: 1,
+          },
+        },
+        defaultModel: 'abab6.5s',
+      },
+      yi: {
+        type: 'custom',
+        baseUrl: 'https://api.lingyiwanwu.com/v1',
+        models: {
+          'yi-large': {
+            name: 'Yi Large',
+            maxContextTokens: 16384,
+            maxOutputTokens: 4096,
+            supportsVision: false,
+            supportsTools: true,
+            supportsStreaming: true,
+            supportsThinking: false,
+            inputCostPerMToken: 2,
+            outputCostPerMToken: 2,
+          },
+          'yi-medium': {
+            name: 'Yi Medium',
+            maxContextTokens: 16384,
+            maxOutputTokens: 4096,
+            supportsVision: false,
+            supportsTools: true,
+            supportsStreaming: true,
+            supportsThinking: false,
+            inputCostPerMToken: 0.35,
+            outputCostPerMToken: 0.35,
+          },
+          'yi-coder': {
+            name: 'Yi Coder',
+            maxContextTokens: 16384,
+            maxOutputTokens: 4096,
+            supportsVision: false,
+            supportsTools: true,
+            supportsStreaming: true,
+            supportsThinking: false,
+            inputCostPerMToken: 1.5,
+            outputCostPerMToken: 1.5,
+          },
+        },
+        defaultModel: 'yi-large',
+      },
+      siliconflow: {
+        type: 'custom',
+        baseUrl: 'https://api.siliconflow.cn/v1',
+        models: {
+          'deepseek-ai/DeepSeek-V3': {
+            name: 'DeepSeek V3 (SiliconFlow)',
+            maxContextTokens: 65536,
+            maxOutputTokens: 8192,
+            supportsVision: false,
+            supportsTools: true,
+            supportsStreaming: true,
+            supportsThinking: false,
+            inputCostPerMToken: 0.27,
+            outputCostPerMToken: 0.11,
+          },
+          'Qwen/Qwen2.5-72B-Instruct': {
+            name: 'Qwen 2.5 72B (SiliconFlow)',
+            maxContextTokens: 131072,
+            maxOutputTokens: 8192,
+            supportsVision: false,
+            supportsTools: true,
+            supportsStreaming: true,
+            supportsThinking: false,
+            inputCostPerMToken: 0.7,
+            outputCostPerMToken: 0.7,
+          },
+          'THUDM/glm-4-9b-chat': {
+            name: 'GLM-4 9B (SiliconFlow)',
+            maxContextTokens: 128000,
+            maxOutputTokens: 4096,
+            supportsVision: false,
+            supportsTools: true,
+            supportsStreaming: true,
+            supportsThinking: false,
+            inputCostPerMToken: 0,
+            outputCostPerMToken: 0,
+          },
+        },
+        defaultModel: 'deepseek-ai/DeepSeek-V3',
+      },
+      // ==================== International Providers ====================
+      groq: {
+        type: 'custom',
+        baseUrl: 'https://api.groq.com/openai/v1',
+        models: {
+          'llama-3.3-70b-versatile': {
+            name: 'Llama 3.3 70B (Groq)',
+            maxContextTokens: 131072,
+            maxOutputTokens: 32768,
+            supportsVision: false,
+            supportsTools: true,
+            supportsStreaming: true,
+            supportsThinking: false,
+            inputCostPerMToken: 0.59,
+            outputCostPerMToken: 0.79,
+          },
+          'llama-3.1-8b-instant': {
+            name: 'Llama 3.1 8B (Groq)',
+            maxContextTokens: 131072,
+            maxOutputTokens: 8192,
+            supportsVision: false,
+            supportsTools: true,
+            supportsStreaming: true,
+            supportsThinking: false,
+            inputCostPerMToken: 0.05,
+            outputCostPerMToken: 0.08,
+          },
+          'mixtral-8x7b-32768': {
+            name: 'Mixtral 8x7B (Groq)',
+            maxContextTokens: 32768,
+            maxOutputTokens: 8192,
+            supportsVision: false,
+            supportsTools: true,
+            supportsStreaming: true,
+            supportsThinking: false,
+            inputCostPerMToken: 0.24,
+            outputCostPerMToken: 0.24,
+          },
+          'gemma2-9b-it': {
+            name: 'Gemma 2 9B (Groq)',
+            maxContextTokens: 8192,
+            maxOutputTokens: 4096,
+            supportsVision: false,
+            supportsTools: true,
+            supportsStreaming: true,
+            supportsThinking: false,
+            inputCostPerMToken: 0.05,
+            outputCostPerMToken: 0.08,
+          },
+        },
+        defaultModel: 'llama-3.3-70b-versatile',
+      },
+      mistral: {
+        type: 'custom',
+        baseUrl: 'https://api.mistral.ai/v1',
+        models: {
+          'mistral-large-latest': {
+            name: 'Mistral Large',
+            maxContextTokens: 131072,
+            maxOutputTokens: 8192,
+            supportsVision: false,
+            supportsTools: true,
+            supportsStreaming: true,
+            supportsThinking: false,
+            inputCostPerMToken: 2,
+            outputCostPerMToken: 6,
+          },
+          'codestral-latest': {
+            name: 'Codestral',
+            maxContextTokens: 32768,
+            maxOutputTokens: 8192,
+            supportsVision: false,
+            supportsTools: true,
+            supportsStreaming: true,
+            supportsThinking: false,
+            inputCostPerMToken: 0.3,
+            outputCostPerMToken: 0.9,
+          },
+          'mistral-small-latest': {
+            name: 'Mistral Small',
+            maxContextTokens: 131072,
+            maxOutputTokens: 8192,
+            supportsVision: false,
+            supportsTools: true,
+            supportsStreaming: true,
+            supportsThinking: false,
+            inputCostPerMToken: 0.2,
+            outputCostPerMToken: 0.6,
+          },
+        },
+        defaultModel: 'mistral-large-latest',
+      },
+      together: {
+        type: 'custom',
+        baseUrl: 'https://api.together.xyz/v1',
+        models: {
+          'meta-llama/Meta-Llama-3.1-70B-Instruct-Turbo': {
+            name: 'Llama 3.1 70B (Together)',
+            maxContextTokens: 131072,
+            maxOutputTokens: 8192,
+            supportsVision: false,
+            supportsTools: true,
+            supportsStreaming: true,
+            supportsThinking: false,
+            inputCostPerMToken: 0.88,
+            outputCostPerMToken: 0.88,
+          },
+          'mistralai/Mixtral-8x7B-Instruct-v0.1': {
+            name: 'Mixtral 8x7B (Together)',
+            maxContextTokens: 32768,
+            maxOutputTokens: 8192,
+            supportsVision: false,
+            supportsTools: true,
+            supportsStreaming: true,
+            supportsThinking: false,
+            inputCostPerMToken: 0.24,
+            outputCostPerMToken: 0.24,
+          },
+        },
+        defaultModel: 'meta-llama/Meta-Llama-3.1-70B-Instruct-Turbo',
+      },
+      perplexity: {
+        type: 'custom',
+        baseUrl: 'https://api.perplexity.ai',
+        models: {
+          'sonar-pro': {
+            name: 'Sonar Pro',
+            maxContextTokens: 200000,
+            maxOutputTokens: 8192,
+            supportsVision: false,
+            supportsTools: false,
+            supportsStreaming: true,
+            supportsThinking: false,
+            inputCostPerMToken: 3,
+            outputCostPerMToken: 15,
+          },
+          'sonar': {
+            name: 'Sonar',
+            maxContextTokens: 127000,
+            maxOutputTokens: 8192,
+            supportsVision: false,
+            supportsTools: false,
+            supportsStreaming: true,
+            supportsThinking: false,
+            inputCostPerMToken: 1,
+            outputCostPerMToken: 1,
+          },
+        },
+        defaultModel: 'sonar-pro',
+      },
+    },
+    aliases: {
+      // Claude aliases
+      'claude': 'claude-3-5-sonnet-20241022',
+      'sonnet': 'claude-3-5-sonnet-20241022',
+      'sonnet4': 'claude-sonnet-4-20250514',
+      'opus': 'claude-3-opus-20240229',
+      'opus4': 'claude-opus-4-20250514',
+      'haiku': 'claude-3-5-haiku-20241022',
+      // OpenAI aliases
+      'gpt4': 'gpt-4o',
+      'gpt4mini': 'gpt-4o-mini',
+      'o1': 'o1',
+      'o1mini': 'o1-mini',
+      'o3mini': 'o3-mini',
+      // Google aliases
+      'gemini': 'gemini-2.0-flash',
+      'gemflash': 'gemini-2.0-flash',
+      'gempro': 'gemini-2.5-pro-preview-05-06',
+      // DeepSeek aliases
+      'deepseek': 'deepseek-chat',
+      'dsr1': 'deepseek-reasoner',
+      // Chinese provider aliases
+      'qwen': 'qwen-max',
+      'qwenplus': 'qwen-plus',
+      'qwen-turbo': 'qwen-turbo',
+      'qwen-coder': 'qwen-coder-plus',
+      'glm': 'glm-4-plus',
+      'glm4': 'glm-4-plus',
+      'moonshot': 'moonshot-v1-128k',
+      'baichuan': 'baichuan4',
+      'yi': 'yi-large',
+      // International provider aliases
+      'groq': 'llama-3.3-70b-versatile',
+      'mistral': 'mistral-large-latest',
+      'codestral': 'codestral-latest',
+      'sonar': 'sonar-pro',
+      // Ollama aliases
+      'local': 'llama3.1',
+      // Ollama Cloud aliases
+      'cloud': 'deepseek-v3.2',
+      'deepseek-cloud': 'deepseek-v3.2',
+      'kimi': 'kimi-k2.5',
+      'qwen-cloud': 'qwen3-coder:480b',
+      'glm-cloud': 'glm-5',
+    },
+  },
+  security: {
+    sandbox: 'restricted',
+    checkpoints: true,
+    blockedCommands: ['rm -rf /', 'mkfs', 'dd if=', ':(){ :|:& };:'],
+    allowedCommands: ['ls', 'cat', 'pwd', 'echo', 'which', 'node', 'npm', 'pnpm'],
+  },
+  fileFilter: {
+    ignorePatterns: ['node_modules/**', '.git/**', 'dist/**', 'build/**', '*.log', '.env', '.env.local'],
+    maxFileSize: 10 * 1024 * 1024,
+    maxBatchSize: 100,
+  },
+  telemetry: {
+    enabled: false,
+    track: { usage: false, errors: true, performance: false },
+  },
+  extensions: {
+    enabled: [],
+  },
+  preferences: {
+    theme: 'auto',
+    language: 'en',
+  },
+};
+
+export class ConfigManager {
+  private config: NovaConfig | null = null;
+  private configPath: string;
+  private projectConfigPath: string | null = null;
+
+  constructor() {
+    this.configPath = path.join(os.homedir(), '.nova', 'config.yaml');
+  }
+
+  /** Load configuration from file, merging with defaults */
+  async load(projectDir?: string): Promise<NovaConfig> {
+    this.config = JSON.parse(JSON.stringify(DEFAULT_CONFIG));
+
+    // Load global config
+    try {
+      const globalContent = await fs.readFile(this.configPath, 'utf-8');
+      const globalConfig = this.parseConfig(globalContent);
+      this.config = this.mergeConfig(this.config, globalConfig);
+    } catch {
+      // Global config doesn't exist yet, that's fine
+    }
+
+    // Load project config
+    if (projectDir) {
+      const projectPaths = [
+        path.join(projectDir, '.nova', 'config.yaml'),
+        path.join(projectDir, '.nova', 'config.json'),
+        path.join(projectDir, '.nova.json'),
+      ];
+
+      for (const p of projectPaths) {
+        try {
+          const content = await fs.readFile(p, 'utf-8');
+          const projectConfig = p.endsWith('.json') ? JSON.parse(content) : this.parseConfig(content);
+          this.config = this.mergeConfig(this.config, projectConfig);
+          this.projectConfigPath = p;
+          break;
+        } catch {
+          continue;
+        }
+      }
+    }
+
+    this.applyEnvOverrides();
+    return this.config;
+  }
+
+  /** Get the loaded configuration */
+  getConfig(): NovaConfig {
+    if (!this.config) {
+      throw new ConfigError('Configuration not loaded. Call load() first.');
+    }
+    return this.config;
+  }
+
+  /** Get core config */
+  getCoreConfig(): CoreConfig {
+    return this.getConfig().core;
+  }
+
+  /** Get model configuration */
+  getModelConfig(modelId: string): { provider: ModelProviderConfig; model: ModelConfig } | null {
+    const config = this.getConfig();
+
+    // Handle provider/model format (e.g., "coding-plan-alibaba/glm-5")
+    let targetProvider: string | null = null;
+    let targetModelId: string = modelId;
+
+    if (modelId.includes('/')) {
+      const [providerKey, modelKey] = modelId.split('/');
+      targetProvider = providerKey;
+      targetModelId = modelKey;
+    }
+
+    // Resolve alias for the model ID (not for provider/model format)
+    const resolvedId = targetProvider ? targetModelId : (config.models.aliases?.[modelId] || modelId);
+
+    // If provider was specified, look only in that provider
+    if (targetProvider) {
+      const provider = config.models.providers[targetProvider];
+      if (provider) {
+        const model = provider.models[targetModelId];
+        if (model) {
+          return {
+            provider: { ...provider, name: targetProvider },
+            model,
+          };
+        }
+      }
+      return null; // Provider specified but not found or model not in provider
+    }
+
+    // Otherwise, search across all providers
+    for (const [providerName, provider] of Object.entries(config.models.providers)) {
+      const model = provider.models[resolvedId];
+      if (model) {
+        return {
+          provider: { ...provider, name: providerName },
+          model,
+        };
+      }
+    }
+
+    return null;
+  }
+
+  /**
+   * Dynamically register a model to an existing provider.
+   * Useful for Ollama runtime model discovery and custom provider model addition.
+   * Returns true if the model was newly registered, false if it already existed.
+   */
+  registerModel(
+    providerName: string,
+    modelId: string,
+    modelConfig: ModelConfig,
+    providerOverrides?: Partial<ModelProviderConfig>
+  ): boolean {
+    if (!this.config) return false;
+
+    if (!this.config.models.providers[providerName]) {
+      // Create a new provider entry
+      this.config.models.providers[providerName] = {
+        type: providerOverrides?.type || 'custom',
+        models: {},
+        ...(providerOverrides?.baseUrl && { baseUrl: providerOverrides.baseUrl }),
+        ...(providerOverrides?.defaultModel && { defaultModel: providerOverrides.defaultModel }),
+      };
+    }
+
+    const provider = this.config.models.providers[providerName];
+    if (provider.models[modelId]) return false; // Already exists
+
+    provider.models[modelId] = modelConfig;
+    if (!provider.defaultModel) {
+      provider.defaultModel = modelId;
+    }
+    return true;
+  }
+
+  /**
+   * Dynamically register a complete provider with models.
+   */
+  registerProvider(providerName: string, providerConfig: ModelProviderConfig): boolean {
+    if (!this.config) return false;
+    this.config.models.providers[providerName] = providerConfig;
+    return true;
+  }
+
+  /**
+   * Add a model alias.
+   */
+  registerAlias(alias: string, targetModelId: string): void {
+    if (!this.config) return;
+    if (!this.config.models.aliases) {
+      this.config.models.aliases = {};
+    }
+    this.config.models.aliases[alias] = targetModelId;
+  }
+
+  /** Get config file path */
+  getConfigPath(): string {
+    return this.configPath;
+  }
+
+  /** Set the default model */
+  setDefaultModel(modelId: string): void {
+    if (!this.config) return;
+    this.config.core.defaultModel = modelId;
+  }
+
+  /** Save configuration */
+  async save(config: Partial<NovaConfig>): Promise<void> {
+    const dir = path.dirname(this.configPath);
+    await fs.mkdir(dir, { recursive: true });
+    const content = JSON.stringify(config, null, 2);
+    await fs.writeFile(this.configPath, content, 'utf-8');
+  }
+
+  /** Deep merge two configurations */
+  private mergeConfig(base: NovaConfig, source: Partial<NovaConfig>): NovaConfig {
+    const result = { ...base };
+    for (const [key, value] of Object.entries(source)) {
+      if (value !== null && typeof value === 'object' && !Array.isArray(value)) {
+        (result as any)[key] = { ...((base as any)[key] || {}), ...value };
+      } else if (value !== undefined) {
+        (result as any)[key] = value;
+      }
+    }
+    return result;
+  }
+
+  /** Apply environment variable overrides */
+  private applyEnvOverrides(): void {
+    if (!this.config) return;
+
+    const overrides: Record<string, (v: string) => void> = {
+      NOVA_MODEL: (v) => { this.config!.core.defaultModel = v; },
+      NOVA_LOG_LEVEL: (v) => { this.config!.core.logLevel = v as any; },
+      NOVA_MAX_TURNS: (v) => { this.config!.core.maxTurns = parseInt(v, 10); },
+      NOVA_MAX_TOKENS: (v) => { this.config!.core.maxTokens = parseInt(v, 10); },
+      NOVA_TEMPERATURE: (v) => { this.config!.core.temperature = parseFloat(v); },
+      NOVA_APPROVAL_MODE: (v) => { this.config!.core.defaultApprovalMode = v as any; },
+    };
+
+    for (const [envKey, setter] of Object.entries(overrides)) {
+      const value = process.env[envKey];
+      if (value) setter(value);
+    }
+  }
+
+  /** Simple config parser (YAML-like or JSON) */
+  private parseConfig(content: string): Record<string, unknown> {
+    try {
+      // Try JSON first
+      return JSON.parse(content);
+    } catch {
+      // Fallback: simple line-based parsing
+      const result: Record<string, unknown> = {};
+      const lines = content.split('\n');
+      let currentKey = '';
+
+      for (const line of lines) {
+        if (line.trim().startsWith('#') || !line.trim()) continue;
+        const indent = line.search(/\S/);
+
+        if (indent === 0) {
+          const [key, ...valueParts] = line.trim().split(':');
+          const value = valueParts.join(':').trim();
+          currentKey = key;
+          result[key] = value ? this.parseValue(value) : {};
+        } else if (currentKey && typeof result[currentKey] === 'object') {
+          const [key, ...valueParts] = line.trim().split(':');
+          const value = valueParts.join(':').trim();
+          (result[currentKey] as Record<string, unknown>)[key] = this.parseValue(value);
+        }
+      }
+      return result;
+    }
+  }
+
+  /** Parse a scalar value */
+  private parseValue(value: string): unknown {
+    if (value === 'true') return true;
+    if (value === 'false') return false;
+    if (value === 'null' || value === '~') return null;
+    if (/^-?\d+$/.test(value)) return parseInt(value, 10);
+    if (/^-?\d+\.\d+$/.test(value)) return parseFloat(value);
+    if ((value.startsWith('"') && value.endsWith('"')) || (value.startsWith("'") && value.endsWith("'"))) {
+      return value.slice(1, -1);
+    }
+    return value;
+  }
+}