converse-mcp-server 2.9.0 → 2.9.2

package/.env.example

@@ -77,9 +77,6 @@ OPENROUTER_REFERER=https://github.com/FallDownTheSystem/converse
 # WARNING: Interactive policies may cause hangs in server/headless mode
 # CODEX_APPROVAL_POLICY=never
 
-# Default model when user specifies model: 'codex' (default: gpt-5-codex)
-# CODEX_DEFAULT_MODEL=gpt-5-codex
-
 # ============================================
 # Server Configuration
 # ============================================

package/README.md

@@ -353,7 +353,6 @@ CODEX_API_KEY=your_codex_api_key_here       # Optional if ChatGPT login availabl
 CODEX_SANDBOX_MODE=read-only                 # read-only (default), workspace-write, danger-full-access
 CODEX_SKIP_GIT_CHECK=true                    # true (default), false
 CODEX_APPROVAL_POLICY=never                  # never (default), untrusted, on-failure, on-request
-CODEX_DEFAULT_MODEL=gpt-5-codex              # Default: gpt-5-codex
 ```
 
 ### Configuration Options

package/docs/API.md

@@ -129,11 +129,11 @@ MCP_TRANSPORT=stdio npm start
   "continuation": {
     "id": "conv_d6a6a5ec-6900-4fd8-a4e0-1fa4f75dfc42",
     "provider": "openai",
-    "model": "gpt-4o-mini",
+    "model": "gpt-5-mini",
     "messageCount": 3
   },
   "metadata": {
-    "model": "gpt-4o-mini",
+    "model": "gpt-5-mini",
     "usage": {
       "input_tokens": 150,
       "output_tokens": 85,
@@ -356,22 +356,25 @@ conv_architecture_design/
 
 | Model | Context | Tokens | Features | Use Cases |
 |-------|---------|--------|----------|-----------|
+| `gpt-5.1` | 1M | 128K | Latest GPT | Multimodal, general purpose |
+| `gpt-5` | 1M | 64K | Advanced | Complex reasoning, analysis |
+| `gpt-5-mini` | 1M | 64K | Fast | Balanced performance/speed |
+| `gpt-5-nano` | 1M | 64K | Ultra-fast | Quick responses, simple queries |
+| `gpt-5-pro` | 1M | 128K | Pro tier | Extended capabilities |
 | `o3` | 200K | 100K | Reasoning | Logic, analysis, complex problems |
-| `o3-mini` | 200K | 100K | Fast reasoning | Balanced performance/speed |
-| `o4-mini` | 200K | 100K | Latest | General purpose, rapid reasoning |
-| `gpt-4o` | 128K | 16K | Multimodal | Vision, general chat |
-| `gpt-4o-mini` | 128K | 16K | Fast multimodal | Quick responses, images |
+| `o3-pro` | 200K | 100K | Extended reasoning | Deep analysis |
+| `o4-mini` | 200K | 100K | Fast reasoning | General purpose, rapid reasoning |
+| `gpt-4.1` | 1M | 32K | Large context | Long documents, analysis |
 
 ### Google/Gemini Models (API-based)
 
 | Model | Alias | Context | Tokens | Features | Use Cases |
 |-------|-------|---------|--------|----------|-----------|
 | `gemini-3-pro-preview` | `pro` | 1M | 64K | Thinking levels, enhanced reasoning | Complex problems, deep analysis |
-| `gemini-2.5-flash` | `flash` | 1M | 65K | Ultra-fast | Quick analysis, simple queries |
 | `gemini-2.5-pro` | `pro 2.5` | 1M | 65K | Thinking mode | Deep reasoning, architecture |
-| `gemini-2.0-flash` | `flash2` | 1M | 65K | Latest | Experimental thinking |
+| `gemini-2.5-flash` | `flash` | 1M | 65K | Ultra-fast | Quick analysis, simple queries |
 
-**Note:** The short model name `gemini` now routes to **Gemini CLI** (OAuth-based). For Google API access, use specific model names like `gemini-2.5-pro` or `gemini-2.0-flash`.
+**Note:** The short model name `gemini` now routes to **Gemini CLI** (OAuth-based). For Google API access, use specific model names like `gemini-2.5-pro` or `gemini-2.5-flash`.
 
 ### X.AI/Grok Models
 
@@ -384,11 +387,11 @@ conv_architecture_design/
 
 | Model | Alias | Context | Tokens | Features | Use Cases |
 |-------|-------|---------|--------|----------|-----------|
-| `claude-opus-4-1-20250805` | `opus-4.1`, `opus-4`, `opus` | 200K | 32K | Extended thinking, images, caching | Complex reasoning tasks |
-| `claude-sonnet-4-20250514` | `sonnet-4`, `sonnet` | 200K | 64K | Extended thinking, images, caching | High performance, balanced |
-| `claude-3-7-sonnet-20250219` | `sonnet-3.7` | 200K | 64K | Extended thinking, images, caching | Enhanced 3.x generation |
-| `claude-3-5-sonnet-20241022` | `claude-3.5-sonnet` | 200K | 8K | Images, caching | Fast and intelligent |
-| `claude-3-5-haiku-20241022` | `haiku` | 200K | 8K | Caching | Fastest, simple queries |
+| `claude-opus-4-5-20250220` | `opus-4.5`, `opus` | 200K | 32K | Extended thinking, images, caching | Most capable reasoning |
+| `claude-opus-4-1-20250805` | `opus-4.1`, `opus-4` | 200K | 32K | Extended thinking, images, caching | Complex reasoning tasks |
+| `claude-sonnet-4-5-20250929` | `sonnet-4.5`, `sonnet` | 200K | 64K | Extended thinking, images, caching | Enhanced reasoning |
+| `claude-sonnet-4-20250514` | `sonnet-4` | 200K | 64K | Extended thinking, images, caching | High performance, balanced |
+| `claude-haiku-4-5-20251001` | `haiku-4.5`, `haiku` | 200K | 64K | Extended thinking, caching | Fast and intelligent |
 
 **Prompt Caching (Always Enabled):**
 - System prompts are automatically cached for 1 hour using Anthropic's prompt caching
@@ -475,7 +478,6 @@ gemini
 - `CODEX_SANDBOX_MODE` - Filesystem access control
 - `CODEX_SKIP_GIT_CHECK` - Git repository requirement
 - `CODEX_APPROVAL_POLICY` - Command approval behavior
-- `CODEX_DEFAULT_MODEL` - Default model when using `model: 'codex'`
 
 ### Model Selection
 
@@ -541,9 +543,6 @@ Control Codex behavior through environment variables:
 - `on-failure`: Prompt when commands fail
 - `on-request`: Let model decide (may hang in headless mode)
 
-**CODEX_DEFAULT_MODEL** - Default model when `model: 'codex'`:
-- Default: `gpt-5-codex`
-
 **Authentication:**
 - Requires ChatGPT login (system-wide, persists across restarts)
 - Alternative: Set `CODEX_API_KEY` environment variable for headless deployments
@@ -557,7 +556,6 @@ CODEX_API_KEY=your_codex_api_key_here
 CODEX_SANDBOX_MODE=read-only                 # Default: read-only
 CODEX_SKIP_GIT_CHECK=true                    # Default: true
 CODEX_APPROVAL_POLICY=never                  # Default: never
-CODEX_DEFAULT_MODEL=gpt-5-codex              # Default: gpt-5-codex
 ```
 
 ## Context Processing
@@ -630,7 +628,7 @@ Response includes continuation ID:
   "continuation": {
     "id": "conv_abc123",
     "provider": "openai",
-    "model": "gpt-4o-mini",
+    "model": "gpt-5-mini",
     "messageCount": 2
   }
 }

package/docs/ARCHITECTURE.md

@@ -471,10 +471,10 @@ function generateRequestId() {
 }
 
 // All operations tagged with request ID
-logger.info('Processing chat request', { 
-  requestId, 
-  provider: 'openai', 
-  model: 'gpt-4o' 
+logger.info('Processing chat request', {
+  requestId,
+  provider: 'openai',
+  model: 'gpt-5'
 });
 ```
 

package/docs/PROVIDERS.md

@@ -9,11 +9,10 @@ This guide documents all supported AI providers in the Converse MCP Server and t
 - **Get Key**: [platform.openai.com/api-keys](https://platform.openai.com/api-keys)
 - **Environment Variable**: `OPENAI_API_KEY`
 - **Supported Models**:
-  - `gpt-5`, `gpt-5-mini`, `gpt-5-nano`, `gpt-5-pro` - GPT-5 family with advanced reasoning
-  - `o3`, `o3-mini`, `o3-pro` - Advanced reasoning models
-  - `o4-mini` - Latest fast reasoning model
+  - `gpt-5.1`, `gpt-5`, `gpt-5-mini`, `gpt-5-nano`, `gpt-5-pro` - GPT-5 family with advanced reasoning
+  - `o3`, `o3-pro`, `o4-mini` - Advanced reasoning models
   - `gpt-4.1` - Large context (1M tokens)
-  - `gpt-4o`, `gpt-4o-mini` - Multimodal models
+  - `o3-deep-research`, `o4-mini-deep-research` - Deep research models
 
 ### Google (Gemini)
 - **API Key Format**: `AIzaSy...` (varies)
@@ -23,8 +22,7 @@ This guide documents all supported AI providers in the Converse MCP Server and t
   - `gemini-3-pro-preview` (alias: `pro`) - Enhanced reasoning with thinking levels (1M context, 64K output)
   - `gemini-2.5-pro` (alias: `pro 2.5`) - Deep reasoning with thinking budget (1M context, 65K output)
   - `gemini-2.5-flash` (alias: `flash`) - Ultra-fast model with thinking budget (1M context, 65K output)
-  - `gemini-2.0-flash`, `gemini-2.0-flash-lite` - Latest generation (1M context, 65K output)
-- **Note**: The short model name `gemini` now routes to **Gemini CLI** (OAuth-based access). For Google API access, use specific model names like `gemini-2.5-pro` or `gemini-2.0-flash`.
+- **Note**: The short model name `gemini` now routes to **Gemini CLI** (OAuth-based access). For Google API access, use specific model names like `gemini-2.5-pro` or `gemini-2.5-flash`.
 
 ### X.AI (Grok)
 - **API Key Format**: `xai-...` (starts with `xai-`)
@@ -39,12 +37,11 @@ This guide documents all supported AI providers in the Converse MCP Server and t
 - **Get Key**: [console.anthropic.com](https://console.anthropic.com/)
 - **Environment Variable**: `ANTHROPIC_API_KEY`
 - **Supported Models**:
+  - `claude-opus-4-5-20250220` - Most capable with extended thinking (32K output)
   - `claude-opus-4-1-20250805` - Highest intelligence with extended thinking (32K output)
+  - `claude-sonnet-4-5-20250929` - Enhanced reasoning with extended thinking (64K output)
   - `claude-sonnet-4-20250514` - Balanced performance with extended thinking (64K output)
-  - `claude-3-7-sonnet-20250219` - Enhanced 3.x generation with thinking (64K output)
   - `claude-haiku-4-5-20251001` - Fast and intelligent with extended thinking (64K output)
-  - `claude-3-5-sonnet-20241022` - Fast and intelligent model (8K output)
-  - `claude-3-5-haiku-20241022` - Fastest Claude model (8K output)
   - `claude-3-opus-20240229`, `claude-3-sonnet-20240229`, `claude-3-haiku-20240307` - Previous generation
 
 ### Mistral
@@ -94,7 +91,6 @@ This guide documents all supported AI providers in the Converse MCP Server and t
   - `CODEX_SANDBOX_MODE` - Filesystem access control (default: read-only)
   - `CODEX_SKIP_GIT_CHECK` - Skip Git repository validation (default: true)
   - `CODEX_APPROVAL_POLICY` - Command approval behavior (default: never)
-  - `CODEX_DEFAULT_MODEL` - Default model (default: gpt-5-codex)
 - **Supported Models**:
   - `codex` - OpenAI Codex agentic coding assistant
   - Thread-based sessions with persistent context
@@ -204,7 +200,6 @@ CODEX_API_KEY=your_codex_api_key_here       # Optional if ChatGPT login availabl
 CODEX_SANDBOX_MODE=read-only                 # read-only (default), workspace-write, danger-full-access
 CODEX_SKIP_GIT_CHECK=true                    # true (default), false
 CODEX_APPROVAL_POLICY=never                  # never (default), untrusted, on-failure, on-request
-CODEX_DEFAULT_MODEL=gpt-5-codex              # Default: gpt-5-codex
 ```
 
 ### Claude Configuration (claude_desktop_config.json)
@@ -235,9 +230,9 @@ CODEX_DEFAULT_MODEL=gpt-5-codex              # Default: gpt-5-codex
 All providers support streaming responses for real-time output.
 
 ### Image Support
-- **Full Support**: OpenAI, Google, X.AI (Grok-4), Anthropic (Claude-4 series, Claude-3.5-Sonnet, Claude-3-Opus)
+- **Full Support**: OpenAI, Google, X.AI (Grok-4), Anthropic (Claude-4 series, Claude-3-Opus)
 - **Via OpenRouter**: Depends on the underlying model
-- **No Support**: DeepSeek, Mistral (except Large), Claude-3.5-Haiku, Codex
+- **No Support**: DeepSeek, Mistral (except Large), Codex
 
 ### Web Search
 - **Native Support**: OpenAI, Google, X.AI (Grok-4)
@@ -248,7 +243,7 @@ All providers support streaming responses for real-time output.
 - **Google**:
   - Gemini 3.0 Pro: Thinking levels (low/high) via `reasoning_effort` - always enabled
   - Gemini 2.5 Pro/Flash: Thinking budget (token-based) via `reasoning_effort`
-- **Anthropic**: Claude 4 and 3.7 models support extended thinking with `reasoning_effort`
+- **Anthropic**: All Claude 4 series models support extended thinking with `reasoning_effort`
 - **Codex**: Thread-based agentic reasoning with persistent context
 - **Others**: Standard inference only
 
@@ -282,14 +277,14 @@ When using the chat or consensus tools, specify models using their identifiers:
 ```javascript
 // Chat tool examples
 {
-  "model": "gpt-4o",              // OpenAI (keyword match)
+  "model": "gpt-5",               // OpenAI (keyword match)
   "model": "claude-opus-4",       // Anthropic (keyword match, auto-resolves)
   "model": "sonnet",              // Anthropic (keyword match)
   "model": "gemini-2.5-pro",      // Google (keyword match)
   "model": "grok-4",              // X.AI (keyword match)
   "model": "mistral-large",       // Mistral (keyword match)
   "model": "deepseek-chat",       // DeepSeek (keyword match)
-  "model": "anthropic/claude-3.5-sonnet",  // OpenRouter (slash format, not in Anthropic)
+  "model": "anthropic/claude-sonnet-4",   // OpenRouter (slash format, not in Anthropic)
   "model": "qwen/qwen3-coder",            // OpenRouter (static model)
   "model": "openrouter/auto"              // OpenRouter auto-selection
 }
@@ -298,7 +293,7 @@ When using the chat or consensus tools, specify models using their identifiers:
 {
   "models": [
     {"model": "o3"},
-    {"model": "claude-3-5-sonnet"},
+    {"model": "claude-sonnet-4"},
     {"model": "gemini-2.5-pro"}
   ]
 }
@@ -313,7 +308,7 @@ When using the chat or consensus tools, specify models using their identifiers:
 
 ### Model Not Found
 - Use exact model identifiers as listed above
-- Some providers support aliases (e.g., "claude" → "claude-3-5-sonnet")
+- Some providers support aliases (e.g., "claude" → "claude-sonnet-4-5")
 - Check provider documentation for model availability in your region
 
 ### Rate Limits

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "converse-mcp-server",
-  "version": "2.9.0",
+  "version": "2.9.2",
   "description": "Converse MCP Server - Converse with other LLMs with chat and consensus tools",
   "type": "module",
   "main": "src/index.js",
@@ -106,6 +106,7 @@
     "dotenv": "^17.2.3",
     "express": "^5.1.0",
     "lru-cache": "^11.2.2",
+    "nanoid": "^5.1.6",
     "openai": "^6.9.1",
     "p-limit": "^7.2.0",
     "vite": "^7.2.2"

package/src/config.js

@@ -39,8 +39,6 @@ const logger = createLogger('config');
 const CONFIG_SCHEMA = {
   // Server configuration
   server: {
-    PORT: { type: 'number', default: 3157, description: 'Server port' },
-    HOST: { type: 'string', default: 'localhost', description: 'Server host' },
     NODE_ENV: {
       type: 'string',
       default: 'development',
@@ -278,11 +276,6 @@ const CONFIG_SCHEMA = {
       description:
         'Approval policy (never | untrusted | on-failure | on-request)',
     },
-    CODEX_DEFAULT_MODEL: {
-      type: 'string',
-      default: 'gpt-5-codex',
-      description: 'Default Codex model',
-    },
   },
 
   // MCP configuration
@@ -819,13 +812,6 @@ export async function validateRuntimeConfig(config) {
     // Validate Codex configuration
     validateCodexConfig(config);
 
-    // Validate server configuration
-    if (config.server.port < 1 || config.server.port > 65535) {
-      throw new ConfigurationError(
-        `Invalid port number: ${config.server.port}`,
-      );
-    }
-
     // Validate environment
     const validEnvs = ['development', 'production', 'test'];
     if (!validEnvs.includes(config.environment.nodeEnv)) {
@@ -933,7 +919,7 @@ function logConfigurationSummary(config) {
   logger.info('Configuration loaded successfully', {
     data: {
       environment: config.environment.nodeEnv,
-      port: config.server.port,
+      transport: config.transport.mcptransport || 'stdio',
       logLevel: config.server.log_level,
       availableProviders: availableProviders.join(', ') || 'none',
       mcpServer: `${config.mcp.name} v${config.mcp.version}`,
@@ -965,3 +951,6 @@ export function getMcpClientConfig(config) {
     providers: getAvailableProviders(config),
   };
 }
+
+// Export CONFIG_SCHEMA for use by help documentation generator
+export { CONFIG_SCHEMA };

package/src/prompts/helpPrompt.js

@@ -1,11 +1,347 @@
 /**
  * Help Prompt Implementation
  *
- * Provides comprehensive help documentation for the Converse MCP Server
+ * Provides comprehensive help documentation for the Converse MCP Server.
+ * Auto-generates documentation from code structures (CONFIG_SCHEMA, tool schemas, provider models).
  */
 
 import { getProviders } from '../providers/index.js';
 import { getTools } from '../tools/index.js';
+import { CONFIG_SCHEMA } from '../config.js';
+
+/**
+ * Sample values for generating realistic tool examples.
+ * Structure comes from schema, content from curated values here.
+ */
+const SAMPLE_VALUES = {
+  prompt: 'Explain the authentication flow in this codebase',
+  files: ['src/auth.js'],
+  images: ['./diagram.png'],
+  model: 'auto',
+  models: ['codex', 'gemini', 'claude'],
+  continuation_id: 'conv_abc123',
+  temperature: 0.5,
+  reasoning_effort: 'medium',
+  verbosity: 'medium',
+  use_websearch: false,
+  async: false,
+  export: false,
+  enable_cross_feedback: true,
+  cross_feedback_prompt: null,
+  full_history: false,
+};
+
+/**
+ * Safely get models from a provider (handles CLI providers that may throw)
+ * @param {object} provider - Provider object
+ * @param {string} name - Provider name for logging
+ * @returns {object} Models map or empty object on error
+ */
+function safeGetModels(provider, name) {
+  try {
+    return provider?.getSupportedModels() || {};
+  } catch (error) {
+    // CLI providers may throw if not installed - silently skip
+    return {};
+  }
+}
+
+/**
+ * Generate environment variables documentation from CONFIG_SCHEMA.
+ * Sorted alphabetically within categories, compact one-liner format.
+ * @returns {string} Formatted markdown section
+ */
+function generateEnvironmentVariablesSection() {
+  const categoryTitles = {
+    server: 'Server Configuration',
+    transport: 'Transport Configuration',
+    apiKeys: 'API Keys (at least one required)',
+    providers: 'Provider Configuration',
+    mcp: 'MCP Configuration',
+    summarization: 'Summarization Configuration',
+    async: 'Async Configuration',
+  };
+
+  let output = '';
+
+  for (const [category, vars] of Object.entries(CONFIG_SCHEMA)) {
+    const title = categoryTitles[category] || category;
+    output += `### ${title}\n`;
+
+    // Sort variables alphabetically within category
+    const sortedVars = Object.entries(vars).sort((a, b) =>
+      a[0].localeCompare(b[0]),
+    );
+
+    for (const [varName, schema] of sortedVars) {
+      const tags = [];
+
+      // Add required tag
+      if (schema.required) {
+        tags.push('Required');
+      }
+
+      // Add secret tag for API keys
+      if (schema.secret) {
+        tags.push('Secret');
+      }
+
+      // Add default value
+      if (schema.default !== undefined && schema.default !== null) {
+        tags.push(`Default: ${JSON.stringify(schema.default)}`);
+      }
+
+      const tagStr = tags.length > 0 ? ` (${tags.join(', ')})` : '';
+      output += `- \`${varName}\`${tagStr}: ${schema.description}\n`;
+    }
+
+    output += '\n';
+  }
+
+  return output;
+}
+
+/**
+ * Generate tool example JSON from input schema using SAMPLE_VALUES.
+ * @param {string} toolName - Name of the tool
+ * @param {object} inputSchema - Tool's input schema
+ * @returns {string} Formatted JSON example in markdown code block
+ */
+function generateToolExamplesFromSchema(toolName, inputSchema) {
+  if (!inputSchema || !inputSchema.properties) {
+    return '';
+  }
+
+  const { properties, required = [] } = inputSchema;
+  const example = {};
+
+  // Include required parameters first
+  for (const name of required) {
+    if (properties[name]) {
+      example[name] = SAMPLE_VALUES[name] ?? getDefaultForType(properties[name]);
+    }
+  }
+
+  // For chat and consensus, add some optional parameters for richer examples
+  if (toolName === 'chat') {
+    // Add commonly used optional parameters
+    if (properties.model) example.model = SAMPLE_VALUES.model;
+    if (properties.files) example.files = SAMPLE_VALUES.files;
+    if (properties.temperature) example.temperature = SAMPLE_VALUES.temperature;
+  } else if (toolName === 'consensus') {
+    // Add commonly used optional parameters
+    if (properties.files) example.files = SAMPLE_VALUES.files;
+    if (properties.enable_cross_feedback)
+      example.enable_cross_feedback = SAMPLE_VALUES.enable_cross_feedback;
+    if (properties.temperature) example.temperature = 0.2; // Lower for consensus
+  } else if (toolName === 'check_status') {
+    // For check_status, just show the continuation_id
+    if (properties.continuation_id)
+      example.continuation_id = SAMPLE_VALUES.continuation_id;
+  } else if (toolName === 'cancel_job') {
+    // For cancel_job, just show the continuation_id
+    if (properties.continuation_id)
+      example.continuation_id = SAMPLE_VALUES.continuation_id;
+  }
+
+  return `\`\`\`json\n${JSON.stringify(example, null, 2)}\n\`\`\``;
+}
+
+/**
+ * Get a default value based on JSON schema type.
+ * @param {object} prop - Property schema
+ * @returns {any} Default value
+ */
+function getDefaultForType(prop) {
+  if (prop.default !== undefined) return prop.default;
+  if (prop.enum && prop.enum.length > 0) return prop.enum[0];
+
+  switch (prop.type) {
+    case 'string':
+      return 'example';
+    case 'number':
+      return 0;
+    case 'boolean':
+      return false;
+    case 'array':
+      return [];
+    case 'object':
+      return {};
+    default:
+      return null;
+  }
+}
+
+/**
+ * Generate factual model categorization lists (no subjective tips).
+ * Categories: context window, web search, thinking mode, image support.
+ * @param {object} allModels - Map of provider name to models
+ * @returns {string} Formatted markdown section with sorted model lists
+ */
+function generateModelCategories(allModels) {
+  // Flatten all models with provider info
+  const models = [];
+  for (const [providerName, providerModels] of Object.entries(allModels)) {
+    for (const [modelId, config] of Object.entries(providerModels)) {
+      models.push({
+        id: modelId,
+        provider: providerName,
+        ...config,
+      });
+    }
+  }
+
+  // Sort helper - alphabetically by model ID
+  const sortByModelId = (a, b) => a.id.localeCompare(b.id);
+
+  let output = '## Model Categories\n\n';
+
+  // Group by context window
+  output += '### Models by Context Window\n\n';
+
+  const windowGroups = {
+    '1M+ tokens': models.filter((m) => m.contextWindow >= 1000000),
+    '400K+ tokens': models.filter(
+      (m) => m.contextWindow >= 400000 && m.contextWindow < 1000000,
+    ),
+    '200K+ tokens': models.filter(
+      (m) => m.contextWindow >= 200000 && m.contextWindow < 400000,
+    ),
+    'Under 200K tokens': models.filter((m) => m.contextWindow < 200000),
+  };
+
+  for (const [groupName, groupModels] of Object.entries(windowGroups)) {
+    if (groupModels.length > 0) {
+      output += `**${groupName}:**\n`;
+      for (const m of groupModels.sort(sortByModelId)) {
+        output += `- ${m.id} (${m.provider}) - ${m.contextWindow.toLocaleString()} tokens\n`;
+      }
+      output += '\n';
+    }
+  }
+
+  // Models with Web Search
+  const webSearchModels = models
+    .filter((m) => m.supportsWebSearch)
+    .sort(sortByModelId);
+  if (webSearchModels.length > 0) {
+    output += '### Models with Web Search\n';
+    for (const m of webSearchModels) {
+      output += `- ${m.id} (${m.provider})\n`;
+    }
+    output += '\n';
+  }
+
+  // Models with Thinking Mode
+  const thinkingModels = models
+    .filter((m) => m.supportsThinking)
+    .sort(sortByModelId);
+  if (thinkingModels.length > 0) {
+    output += '### Models with Thinking Mode\n';
+    for (const m of thinkingModels) {
+      output += `- ${m.id} (${m.provider})\n`;
+    }
+    output += '\n';
+  }
+
+  // Models with Image Support
+  const imageModels = models
+    .filter((m) => m.supportsImages)
+    .sort(sortByModelId);
+  if (imageModels.length > 0) {
+    output += '### Models with Image Support\n';
+    for (const m of imageModels) {
+      output += `- ${m.id} (${m.provider})\n`;
+    }
+    output += '\n';
+  }
+
+  return output;
+}
+
+/**
+ * Generate configuration tips from tool parameter schemas.
+ * Extracts enum values and ranges directly from schema.
+ * @param {object} tools - Map of tool name to tool implementation
+ * @returns {string} Formatted markdown section
+ */
+function generateConfigurationTips(tools) {
+  let output = '## Configuration Tips\n\n';
+
+  // Get chat tool schema for parameter info
+  const chatSchema = tools.chat?.inputSchema?.properties || {};
+
+  // Temperature Settings
+  if (chatSchema.temperature) {
+    const tempSchema = chatSchema.temperature;
+    output += '### Temperature Settings\n';
+    if (tempSchema.minimum !== undefined && tempSchema.maximum !== undefined) {
+      output += `Range: ${tempSchema.minimum} to ${tempSchema.maximum}\n`;
+    }
+    output += `- **0.0-0.3**: Factual, deterministic responses\n`;
+    output += `- **0.4-0.7**: Balanced creativity and accuracy (recommended)\n`;
+    output += `- **0.8-1.2**: Creative writing, brainstorming\n`;
+    if (tempSchema.maximum && tempSchema.maximum > 1.2) {
+      output += `- **1.3-${tempSchema.maximum}**: Highly experimental, unpredictable\n`;
+    }
+    output += '\n';
+  }
+
+  // Reasoning Effort
+  if (chatSchema.reasoning_effort) {
+    const effortSchema = chatSchema.reasoning_effort;
+    output += '### Reasoning Effort (for supported models)\n';
+    if (effortSchema.enum) {
+      for (const value of effortSchema.enum) {
+        const descriptions = {
+          none: 'No reasoning, fastest response (GPT-5.1+ only)',
+          minimal: 'Quick responses with minimal reasoning',
+          low: 'Light analysis, simple problems',
+          medium: 'Balanced reasoning (default)',
+          high: 'Deep analysis, complex problems',
+          max: 'Maximum reasoning capability',
+        };
+        output += `- **${value}**: ${descriptions[value] || value}\n`;
+      }
+    }
+    output += '\n';
+  }
+
+  // Verbosity
+  if (chatSchema.verbosity) {
+    const verbSchema = chatSchema.verbosity;
+    output += '### Verbosity (for GPT-5 models)\n';
+    if (verbSchema.enum) {
+      const descriptions = {
+        low: 'Concise answers',
+        medium: 'Balanced detail (default)',
+        high: 'Thorough explanations',
+      };
+      for (const value of verbSchema.enum) {
+        output += `- **${value}**: ${descriptions[value] || value}\n`;
+      }
+    }
+    output += '\n';
+  }
+
+  // File Context
+  output += '### File Context\n';
+  output +=
+    '- Supports multiple file formats: code files, text, markdown, JSON, etc.\n';
+  output +=
+    '- Line ranges supported: `file.txt{10:50}` for lines 10-50, `file.txt{100:}` for line 100 onwards\n';
+  output += '- Files are automatically chunked if too large\n';
+  output +=
+    '- Images are base64 encoded and sent to models that support vision\n\n';
+
+  // Continuation IDs
+  output += '### Continuation IDs\n';
+  output += '- Automatically generated for new conversations\n';
+  output += '- Returned in the response for continuing conversations\n';
+  output += '- Conversations expire after 24 hours of inactivity\n\n';
+
+  return output;
+}
 
 /**
  * Generate comprehensive help content dynamically based on current providers
@@ -14,7 +350,7 @@ import { getTools } from '../tools/index.js';
 export function generateHelpContent(config = null) {
   const providers = getProviders();
 
-  // Collect all models from all providers
+  // Collect all models from all providers (including CLI providers with safe access)
   const allModels = {
     openai: providers.openai?.getSupportedModels() || {},
     google: providers.google?.getSupportedModels() || {},
@@ -23,8 +359,18 @@ export function generateHelpContent(config = null) {
     mistral: providers.mistral?.getSupportedModels() || {},
     deepseek: providers.deepseek?.getSupportedModels() || {},
     openrouter: providers.openrouter?.getSupportedModels() || {},
+    // CLI providers - use safeGetModels (may throw if CLI not installed)
+    codex: safeGetModels(providers.codex, 'codex'),
+    claude: safeGetModels(providers.claude, 'claude'),
+    'gemini-cli': safeGetModels(providers['gemini-cli'], 'gemini-cli'),
   };
 
+  // Limit OpenRouter models if dynamic models enabled (could have hundreds)
+  if (allModels.openrouter && Object.keys(allModels.openrouter).length > 20) {
+    const entries = Object.entries(allModels.openrouter).slice(0, 20);
+    allModels.openrouter = Object.fromEntries(entries);
+  }
+
   // Format provider models for display
   const formatProviderModels = (providerName, models) => {
     if (!models || Object.keys(models).length === 0) return '';
@@ -80,33 +426,9 @@ export function generateHelpContent(config = null) {
     return params.join('\n');
   };
 
-  const formatToolExample = (toolName) => {
-    if (toolName === 'chat') {
-      return `\`\`\`json
-{
-  "prompt": "Explain the code in main.js",
-  "model": "gpt-5",
-  "files": ["C:\\\\Users\\\\username\\\\project\\\\main.js"],
-  "temperature": 0.7,
-  "use_websearch": false
-}
-\`\`\``;
-    } else if (toolName === 'consensus') {
-      return `\`\`\`json
-{
-  "prompt": "Should we use microservices architecture for our new project?",
-  "models": ["codex", "gemini", "claude"],
-  "files": ["./requirements.md", "C:\\\\Users\\\\username\\\\architecture.md"],
-  "enable_cross_feedback": true,
-  "temperature": 0.3
-}
-\`\`\``;
-    }
-    return '';
-  };
-
   const toolsSection = Object.entries(tools)
     .map(([name, tool], index) => {
+      const example = generateToolExamplesFromSchema(name, tool.inputSchema);
       return `### ${index + 1}. ${name.charAt(0).toUpperCase() + name.slice(1)} Tool
 ${tool.description}
 
@@ -114,7 +436,7 @@ ${tool.description}
 ${formatToolParameters(tool.inputSchema)}
 
 **Example Usage:**
-${formatToolExample(name)}`;
+${example}`;
     })
     .join('\n\n');
 
@@ -134,54 +456,13 @@ ${formatProviderModels('Anthropic', allModels.anthropic)}
 ${formatProviderModels('Mistral', allModels.mistral)}
 ${formatProviderModels('DeepSeek', allModels.deepseek)}
 ${formatProviderModels('OpenRouter', allModels.openrouter)}
+${formatProviderModels('Codex', allModels.codex)}
+${formatProviderModels('Claude CLI', allModels.claude)}
+${formatProviderModels('Gemini CLI', allModels['gemini-cli'])}
+
+${generateModelCategories(allModels)}
 
-## Model Selection Tips
-
-### For Complex Reasoning Tasks
-- **Most Intelligent**: gpt-5, gpt-5-pro, gemini-pro, grok-4
-- **Fast & Smart**: gpt-5-mini, gpt-5-mini, o4-mini, flash
-- **Budget-Friendly**: gpt-5-nano, gpt-4o-mini, gemini-2.0-flash-lite
-
-### For Quick Responses
-- **Ultra-Fast**: gpt-5-nano, flash, gemini-2.0-flash, gpt-4o-mini
-- **Good Balance**: gpt-5-mini, o4-mini, grok-code-fast-1
-
-### For Large Context Windows
-- **1M+ Tokens**: gpt-4.1 (1M), all Gemini models (1M)
-- **400K Tokens**: gpt-5 family (gpt-5, gpt-5-mini, gpt-5-nano, gpt-5-pro)
-- **256K Tokens**: grok-4 series
-- **200K Tokens**: o3 series, o4-mini
-
-### Special Features
-- **Web Search**: gpt-5, gpt-5-mini, gpt-5-pro, o3 series, o4-mini, gpt-4 series, gemini models with grounding, grok-4
-- **Thinking Mode**: gpt-5 series (reasoning_effort), gemini models (thinking budget)
-- **Image Support**: All models except gemini-2.0-flash-lite and grok-code-fast-1
-
-## Configuration Tips
-
-### Temperature Settings
-- **0.0-0.3**: Factual, deterministic responses
-- **0.4-0.7**: Balanced creativity and accuracy (recommended)
-- **0.8-1.2**: Creative writing, brainstorming
-- **1.3-2.0**: Highly experimental, unpredictable
-
-### Reasoning Effort (for supported models)
-- **minimal**: Quick responses with minimal reasoning
-- **low**: Light analysis, simple problems
-- **medium**: Balanced reasoning (default)
-- **high**: Deep analysis, complex problems
-- **max**: Maximum reasoning capability
-
-### File Context
-- Supports multiple file formats: code files, text, markdown, JSON, etc.
-- Use git-bash style paths on Windows: \`/c/Users/username/file.txt\`
-- Files are automatically chunked if too large
-- Images are base64 encoded and sent to models that support vision
-
-### Continuation IDs
-- Automatically generated for new conversations
-- Returned in the response for continuing conversations
-- Conversations expire after 24 hours of inactivity
+${generateConfigurationTips(tools)}
 
 ## Best Practices
 
@@ -207,30 +488,15 @@ ${formatProviderModels('OpenRouter', allModels.openrouter)}
 
 ## Environment Variables
 
-### API Keys (at least one required):
-- \`OPENAI_API_KEY\`: For OpenAI models
-- \`GOOGLE_API_KEY\`: For Google Gemini models
-- \`XAI_API_KEY\`: For X.AI Grok models
-- \`ANTHROPIC_API_KEY\`: For Anthropic Claude models
-- \`MISTRAL_API_KEY\`: For Mistral models
-- \`DEEPSEEK_API_KEY\`: For DeepSeek models
-- \`OPENROUTER_API_KEY\`: For OpenRouter models
-
-### OpenRouter Configuration:
-- \`OPENROUTER_REFERER\`: OpenRouter referer header for compliance (required for OpenRouter)
-- \`OPENROUTER_TITLE\`: OpenRouter X-Title header for request tracking (optional)
-- \`OPENROUTER_DYNAMIC_MODELS\`: Enable dynamic model discovery via OpenRouter endpoints API (default: false). Must be set to true to use models in \`provider/model\` format (e.g., \`anthropic/claude-3.5-sonnet\`). When enabled, fetches actual model capabilities from API.
-
-### Server Configuration:
-- \`MAX_MCP_OUTPUT_TOKENS\`: Maximum response size (default: 25000)
-- \`LOG_LEVEL\`: Logging verbosity (debug, info, warn, error)
-- \`PORT\`: HTTP server port (default: 3157)
-- \`HTTP_ENABLED\`: Enable HTTP transport (default: true)
-- \`HTTP_RATE_LIMIT_ENABLED\`: Enable rate limiting (default: false)
-- \`HTTP_RATE_LIMIT_WINDOW\`: Rate limit window in milliseconds (default: 900000 - 15 minutes)
-- \`HTTP_RATE_LIMIT_MAX_REQUESTS\`: Maximum requests per window (default: 1000)
-
-Note: Server name and version are automatically read from package.json.
+${generateEnvironmentVariablesSection()}
+
+## CLI-Based Providers (Special Authentication)
+
+These providers use local CLI tools and don't require API keys:
+
+- **codex**: Requires ChatGPT login or CODEX_API_KEY environment variable
+- **claude**: Requires \`claude login\` command (Claude Code CLI authentication)
+- **gemini-cli**: Requires Gemini OAuth authentication via \`gemini\` CLI
 
 ## Need More Help?
 
@@ -271,11 +537,11 @@ export async function helpPromptHandler(args = {}, config = null) {
   let sectionContent = '';
 
   if (topicLower === 'tools') {
-    const toolsMatch = fullHelp.match(/## Available Tools[\s\S]*?(?=##|$)/);
+    const toolsMatch = fullHelp.match(/## Available Tools[\s\S]*?(?=## Provider Models|$)/);
     sectionContent = toolsMatch ? toolsMatch[0] : 'Tools section not found';
   } else if (topicLower === 'models' || topicLower === 'providers') {
     const modelsMatch = fullHelp.match(
-      /## Provider Models[\s\S]*?(?=## Model Selection Tips|$)/,
+      /## Provider Models[\s\S]*?(?=## Model Categories|$)/,
     );
     sectionContent = modelsMatch ? modelsMatch[0] : 'Models section not found';
   } else if (topicLower === 'parameters') {
@@ -286,7 +552,7 @@ export async function helpPromptHandler(args = {}, config = null) {
       ? paramsMatch[0]
       : 'Parameters section not found';
   } else if (topicLower === 'examples') {
-    // Extract example usage from both tools
+    // Extract example usage from all tools
     const examples = fullHelp.match(
       /\*\*Example Usage:\*\*[\s\S]*?```[\s\S]*?```/g,
     );

package/src/providers/anthropic.js

@@ -115,28 +115,6 @@ const SUPPORTED_MODELS = {
       'Claude Sonnet 4 - High intelligence and balanced performance with extended thinking',
     aliases: ['claude-4-sonnet', 'sonnet-4', 'claude-sonnet-4', 'sonnet4'],
   },
-  'claude-3-7-sonnet-20250219': {
-    modelName: 'claude-3-7-sonnet-20250219',
-    friendlyName: 'Claude 3.7 Sonnet',
-    contextWindow: 200000,
-    maxOutputTokens: 64000,
-    supportsStreaming: true,
-    supportsImages: true,
-    supportsTemperature: true,
-    supportsWebSearch: false,
-    supportsThinking: true,
-    minThinkingTokens: 1024,
-    maxThinkingTokens: 64000,
-    timeout: 300000,
-    description: 'Claude 3.7 Sonnet - Enhanced 3.x generation with thinking',
-    aliases: [
-      'claude-3.7-sonnet',
-      'sonnet-3.7',
-      'claude-3-7-sonnet',
-      'claude 3.7 sonnet',
-      'sonnet 3.7',
-    ],
-  },
   'claude-haiku-4-5-20251001': {
     modelName: 'claude-haiku-4-5-20251001',
     friendlyName: 'Claude Haiku 4.5',
@@ -165,47 +143,6 @@ const SUPPORTED_MODELS = {
       'claude-haiku',
     ],
   },
-  'claude-3-5-sonnet-20241022': {
-    modelName: 'claude-3-5-sonnet-20241022',
-    friendlyName: 'Claude 3.5 Sonnet',
-    contextWindow: 200000,
-    maxOutputTokens: 8192,
-    supportsStreaming: true,
-    supportsImages: true,
-    supportsTemperature: true,
-    supportsWebSearch: false,
-    supportsThinking: false,
-    timeout: 300000,
-    description: 'Claude 3.5 Sonnet - Fast and intelligent model',
-    aliases: [
-      'claude-3.5-sonnet',
-      'claude-3-5-sonnet-latest',
-      'claude-sonnet-3.5',
-      'sonnet-3.5',
-      'claude 3.5 sonnet',
-      'sonnet 3.5',
-    ],
-  },
-  'claude-3-5-haiku-20241022': {
-    modelName: 'claude-3-5-haiku-20241022',
-    friendlyName: 'Claude 3.5 Haiku',
-    contextWindow: 200000,
-    maxOutputTokens: 8192,
-    supportsStreaming: true,
-    supportsImages: false,
-    supportsTemperature: true,
-    supportsWebSearch: false,
-    supportsThinking: false,
-    timeout: 180000,
-    description: 'Claude 3.5 Haiku - Fastest model, best for simple queries',
-    aliases: [
-      'claude-3.5-haiku',
-      'claude-3-5-haiku-latest',
-      'haiku-3.5',
-      'claude 3.5 haiku',
-      'haiku 3.5',
-    ],
-  },
 };
 
 /**

package/src/providers/google.js

@@ -10,51 +10,6 @@ import { debugLog, debugError } from '../utils/console.js';
 
 // Define supported Gemini models with their capabilities
 const SUPPORTED_MODELS = {
-  'gemini-2.0-flash': {
-    modelName: 'gemini-2.0-flash',
-    friendlyName: 'Gemini (Flash 2.0)',
-    contextWindow: 1048576, // 1M tokens
-    maxOutputTokens: 65536,
-    supportsStreaming: true,
-    supportsImages: true,
-    supportsTemperature: true,
-    supportsThinking: false,
-    supportsWebSearch: true,
-    maxThinkingTokens: 0,
-    timeout: 300000,
-    description:
-      'Gemini 2.0 Flash (1M context) - Latest fast model, supports audio/video input and grounding',
-    aliases: [
-      'flash-2.0',
-      'flash2',
-      'flash 2.0',
-      'gemini flash 2.0',
-      'gemini-2.0-flash-latest',
-    ],
-  },
-  'gemini-2.0-flash-lite': {
-    modelName: 'gemini-2.0-flash-lite',
-    friendlyName: 'Gemini (Flash Lite 2.0)',
-    contextWindow: 1048576, // 1M tokens
-    maxOutputTokens: 65536,
-    supportsStreaming: true,
-    supportsImages: false,
-    supportsTemperature: true,
-    supportsThinking: false,
-    supportsWebSearch: true,
-    maxThinkingTokens: 0,
-    timeout: 300000,
-    description:
-      'Gemini 2.0 Flash Lite (1M context) - Lightweight fast model, text-only with grounding',
-    aliases: [
-      'flashlite',
-      'flash-lite',
-      'flash lite',
-      'flash-lite-2.0',
-      'gemini flash lite',
-      'gemini-2.0-flash-lite-latest',
-    ],
-  },
   'gemini-2.5-flash': {
     modelName: 'gemini-flash-latest',
     friendlyName: 'Gemini (Flash 2.5)',

package/src/providers/mistral.js

@@ -265,6 +265,9 @@ export const mistralProvider = {
       // eslint-disable-next-line no-unused-vars
       use_websearch = false, // Not supported by Mistral, ignored
       config,
+      // Filter out options not meant for the API
+      continuation_id, // eslint-disable-line no-unused-vars
+      continuationStore, // eslint-disable-line no-unused-vars
       ...otherOptions
     } = options;
 

package/src/providers/openai-compatible.js

@@ -297,6 +297,9 @@ export function createOpenAICompatibleProvider(providerConfig) {
         use_websearch = false,
         signal,
         config,
+        // Filter out options not meant for the API
+        continuation_id, // eslint-disable-line no-unused-vars
+        continuationStore, // eslint-disable-line no-unused-vars
         ...otherOptions
       } = options;
 

package/src/providers/openai.js

@@ -116,21 +116,6 @@ const SUPPORTED_MODELS = {
       'Strong reasoning (200K context) - Logical problems, code generation, systematic analysis',
     aliases: ['o3-2025-01-31'],
   },
-  'o3-mini': {
-    modelName: 'o3-mini',
-    friendlyName: 'OpenAI (O3-mini)',
-    contextWindow: 200000,
-    maxOutputTokens: 100000,
-    supportsStreaming: true,
-    supportsImages: true,
-    supportsTemperature: false,
-    supportsWebSearch: false, // o3-mini does not support web search
-    supportsResponsesAPI: true,
-    timeout: 300000,
-    description:
-      'Fast O3 variant (200K context) - Balanced performance/speed, moderate complexity',
-    aliases: ['o3mini', 'o3 mini', 'o3-mini-2025-01-31'],
-  },
   'o3-pro-2025-06-10': {
     modelName: 'o3-pro-2025-06-10',
     friendlyName: 'OpenAI (O3-Pro)',
@@ -176,36 +161,6 @@ const SUPPORTED_MODELS = {
       'GPT-4.1 (1M context) - Advanced reasoning model with large context window',
     aliases: ['gpt4.1', 'gpt-4.1', 'gpt 4.1', 'gpt-4.1-latest'],
   },
-  'gpt-4o': {
-    modelName: 'gpt-4o',
-    friendlyName: 'OpenAI (GPT-4o)',
-    contextWindow: 128000,
-    maxOutputTokens: 16384,
-    supportsStreaming: true,
-    supportsImages: true,
-    supportsTemperature: true,
-    supportsWebSearch: true,
-    supportsResponsesAPI: true,
-    timeout: 180000,
-    description:
-      'GPT-4o (128K context) - Multimodal flagship model with vision capabilities',
-    aliases: ['gpt4o', 'gpt 4o', '4o'],
-  },
-  'gpt-4o-mini': {
-    modelName: 'gpt-4o-mini',
-    friendlyName: 'OpenAI (GPT-4o-mini)',
-    contextWindow: 128000,
-    maxOutputTokens: 16384,
-    supportsStreaming: true,
-    supportsImages: true,
-    supportsTemperature: true,
-    supportsWebSearch: true,
-    supportsResponsesAPI: true,
-    timeout: 120000,
-    description:
-      'GPT-4o-mini (128K context) - Fast and efficient multimodal model',
-    aliases: ['gpt4o-mini', 'gpt 4o mini', '4o mini', '4o-mini'],
-  },
   'o3-deep-research-2025-06-26': {
     modelName: 'o3-deep-research-2025-06-26',
     friendlyName: 'OpenAI (O3 Deep Research)',
@@ -424,6 +379,9 @@ export const openaiProvider = {
       use_websearch = false,
       signal,
       config,
+      // Filter out options not meant for the API
+      continuation_id, // eslint-disable-line no-unused-vars
+      continuationStore, // eslint-disable-line no-unused-vars
       ...otherOptions
     } = options;
 

package/src/providers/xai.js

@@ -228,6 +228,9 @@ export const xaiProvider = {
       use_websearch = false,
       signal,
       config,
+      // Filter out options not meant for the API
+      continuation_id, // eslint-disable-line no-unused-vars
+      continuationStore, // eslint-disable-line no-unused-vars
       ...otherOptions
     } = options;
 

package/src/tools/chat.js

@@ -1029,7 +1029,7 @@ chatTool.inputSchema = {
       type: 'array',
       items: { type: 'string' },
       description:
-        'File paths to include as context (absolute or relative paths). Supports line ranges: "file.txt{10:50}" for lines 10-50, "{:20}" for first 20 lines, "{100:}" for line 100 to end. Example: ["./src/auth.js{50:100}", "./config.json"]',
+        'File paths to include as context (absolute or relative paths). Supports line ranges: file.txt{10:50}, file.txt{100:}. Example: ["./src/utils/auth.js{50:100}", "./config.json"]',
     },
     images: {
       type: 'array',

package/src/tools/consensus.js

@@ -1631,7 +1631,7 @@ consensusTool.inputSchema = {
       type: 'array',
       items: { type: 'string' },
       description:
-        'File paths for additional context (absolute or relative paths). Supports line ranges: "file.txt{10:50}" for lines 10-50, "{:20}" for first 20 lines, "{100:}" for line 100 to end. Example: ["./architecture.md{1:100}", "./requirements.txt"]',
+        'File paths for additional context (absolute or relative paths). Supports line ranges: file.txt{10:50}, file.txt{100:}. Example: ["./docs/architecture.md{1:100}", "./requirements.txt"]',
     },
     images: {
       type: 'array',