converse-mcp-server 3.7.1 → 4.1.0

package/src/providers/gemini-cli.js

@@ -18,11 +18,10 @@
  *   interactively (`agy`) via Google OAuth. The first interactive login also
  *   establishes workspace trust for the user's home directory.
  *
- * The provider registry key remains 'gemini-cli' and the user-facing alias
- * remains 'gemini' for routing/normalization stability. Only three user-facing
- * model names are exposed: gemini (= gemini:flash), gemini:flash, gemini:pro.
- * Flash is the default: Gemini 3.8 Flash is the current-generation model agy
- * lists first, while 3.1 Pro remains the only Pro tier Antigravity offers.
+ * The provider registry key remains 'gemini-cli'; its namespaces are `gemini:`
+ * and `agy:`. Flash is the default: Gemini 3.8 Flash is the current-generation
+ * model agy lists first, while 3.1 Pro remains the only Pro tier Antigravity
+ * offers.
  */
 
 import { existsSync, mkdirSync, writeFileSync, rmSync } from 'node:fs';
@@ -32,6 +31,7 @@ import { randomUUID } from 'node:crypto';
 import { debugLog, debugError } from '../utils/console.js';
 import { ProviderError, ErrorCodes, StopReasons } from './interface.js';
 import { clampReasoningEffort } from '../utils/reasoningEffort.js';
+import { findCatalogEntry } from '../utils/modelCatalog.js';
 
 // Prompts at or below this length pass directly as the -p argv value (fast
 // path). Larger prompts are written to a file and -p carries a bootstrap
@@ -53,13 +53,15 @@ const POST_KILL_GRACE_MS = 5000;
 const PTY_COLS = 1000;
 
 /**
- * Supported Gemini models. Only three user-facing names are exposed; each maps
- * to an agy display-name base that gets a reasoning-effort suffix appended at
- * spawn time. All are text-only (print mode has no image input channel).
+ * Supported Gemini models, keyed by the same model IDs the Google API provider
+ * uses so a bare name resolves to one model whichever provider serves it. Each
+ * maps to an agy display-name base that gets a reasoning-effort suffix
+ * appended at spawn time. All are text-only (print mode has no image input
+ * channel).
  */
 const SUPPORTED_MODELS = {
-  gemini: {
-    modelName: 'gemini',
+  'gemini-3.8-flash': {
+    modelName: 'gemini-3.8-flash',
     friendlyName: 'Gemini 3.8 Flash (via Antigravity CLI)',
     contextWindow: 1048576,
     maxOutputTokens: 65536,
@@ -69,28 +71,23 @@ const SUPPORTED_MODELS = {
     supportsThinking: true,
     timeout: DEFAULT_TIMEOUT_MS,
     description:
-      'Gemini 3.8 Flash via Antigravity CLI (agy) - requires Antigravity Google OAuth login',
-    aliases: ['gemini-cli'],
+      'Gemini 3.8 Flash via Antigravity CLI (agy, default) - requires Antigravity Google OAuth login',
+    aliases: [
+      'flash',
+      'gemini-3.8',
+      'gemini3.8',
+      'gemini-3.8-flash-latest',
+      'flash-3.8',
+      'flash3.8',
+      'gemini-flash-3.8',
+      'gemini flash 3.8',
+      '3.8-flash',
+    ],
     // agy display-name base; reasoning_effort selects the parenthesized variant
     agyModelBase: 'Gemini 3.8 Flash',
   },
-  'gemini:flash': {
-    modelName: 'gemini:flash',
-    friendlyName: 'Gemini 3.8 Flash (via Antigravity CLI)',
-    contextWindow: 1048576,
-    maxOutputTokens: 65536,
-    supportsStreaming: true,
-    supportsImages: false,
-    supportsWebSearch: false,
-    supportsThinking: true,
-    timeout: DEFAULT_TIMEOUT_MS,
-    description:
-      'Gemini 3.8 Flash via Antigravity CLI (agy) - explicit alias of `gemini`',
-    aliases: ['flash'],
-    agyModelBase: 'Gemini 3.8 Flash',
-  },
-  'gemini:pro': {
-    modelName: 'gemini:pro',
+  'gemini-3.1-pro-preview': {
+    modelName: 'gemini-3.1-pro-preview',
     friendlyName: 'Gemini 3.1 Pro (via Antigravity CLI)',
     contextWindow: 1048576,
     maxOutputTokens: 65536,
@@ -101,11 +98,26 @@ const SUPPORTED_MODELS = {
     timeout: DEFAULT_TIMEOUT_MS,
     description:
       'Gemini 3.1 Pro via Antigravity CLI (agy) - requires Antigravity Google OAuth login',
-    aliases: ['pro'],
+    aliases: [
+      'pro',
+      'gemini-pro',
+      'gemini pro',
+      'gemini-3.1-pro',
+      'gemini-3.1',
+      'gemini3.1',
+      '3.1-pro',
+      'gemini-3',
+      'gemini3',
+      'gemini-3-pro',
+      'gemini-3-pro-preview',
+      '3-pro',
+    ],
     agyModelBase: 'Gemini 3.1 Pro',
   },
 };
 
+const DEFAULT_MODEL = 'gemini-3.8-flash';
+
 /**
  * Custom error class for Gemini CLI (agy) provider errors
  */
@@ -197,45 +209,24 @@ function effortSuffix(base, reasoningEffort) {
 }
 
 /**
- * Resolve a user-facing model name + reasoning_effort to the agy display name
- * passed via --model. Strips the gemini: prefix (case-insensitive), maps the
- * alias, and appends the effort suffix. Full agy display names pass through
- * verbatim so power users aren't blocked.
- * @param {string} model - e.g. 'gemini', 'gemini:flash', or a full agy name
+ * Resolve a model name + reasoning_effort to the agy display name passed via
+ * --model. The router hands over canonical catalog IDs; aliases are accepted
+ * for direct callers.
+ * @param {string} [model] - Catalog ID or alias; defaults to DEFAULT_MODEL
  * @param {string} [reasoningEffort]
  * @returns {string} agy --model value, e.g. 'Gemini 3.8 Flash (High)'
+ * @throws {GeminiCliProviderError} When the name is not in the catalog
  */
 export function resolveAgyModel(model, reasoningEffort) {
-  const raw = typeof model === 'string' ? model.trim() : '';
-
-  // Full agy display-name passthrough (already contains a parenthesized variant)
-  if (/\(.*\)\s*$/.test(raw) && /gemini/i.test(raw)) {
-    return raw;
-  }
-
-  let name = raw;
-  if (name.toLowerCase().startsWith('gemini:')) {
-    name = name.slice('gemini:'.length).trim();
-  }
-
-  const nameLower = name.toLowerCase();
-
-  // Determine the agy base
-  let base;
-  if (
-    !nameLower ||
-    nameLower === 'gemini' ||
-    nameLower === 'gemini-cli' ||
-    nameLower === 'flash'
-  ) {
-    base = SUPPORTED_MODELS.gemini.agyModelBase;
-  } else if (nameLower === 'pro') {
-    base = SUPPORTED_MODELS['gemini:pro'].agyModelBase;
-  } else {
-    // Unknown suffix: pass through verbatim (power-user agy display name)
-    return raw;
+  const name = typeof model === 'string' && model.trim() ? model : DEFAULT_MODEL;
+  const entry = findCatalogEntry(SUPPORTED_MODELS, name);
+  if (!entry) {
+    throw new GeminiCliProviderError(
+      `Unknown Antigravity model "${model}"`,
+      ErrorCodes.MODEL_NOT_FOUND,
+    );
   }
-
+  const base = entry.agyModelBase;
   return `${base} ${effortSuffix(base, reasoningEffort)}`;
 }
 
@@ -654,7 +645,7 @@ async function* createStreamingGenerator(fullText, userFacingModel) {
  * provider errors.
  */
 async function executeAgy(messages, options) {
-  const { model = 'gemini', reasoning_effort, signal, timeout } = options;
+  const { model = DEFAULT_MODEL, reasoning_effort, signal, timeout } = options;
 
   const prompt = buildPrompt(messages);
   const agyModel = resolveAgyModel(model, reasoning_effort);
@@ -699,7 +690,7 @@ export const geminiCliProvider = {
    * @returns {Promise<Object>|AsyncGenerator} Response or stream generator
    */
   async invoke(messages, options = {}) {
-    const { model = 'gemini', stream = false, signal } = options;
+    const { model = DEFAULT_MODEL, stream = false, signal } = options;
 
     if (signal?.aborted) {
       throw new GeminiCliProviderError('Request cancelled', 'CANCELLED');
@@ -729,9 +720,13 @@ export const geminiCliProvider = {
     };
   },
 
+  defaultModel: DEFAULT_MODEL,
+
   /**
    * Validate configuration. agy uses OAuth (no env keys); always true.
-   * Availability is determined by isAvailable (binary presence).
+   * Availability is determined by isAvailable (binary presence). agy keeps
+   * its login outside any file converse can check, so a logged-out agy
+   * surfaces at invoke time and bare-name/auto routing fails over.
    */
   validateConfig(_config) {
     return true;
@@ -752,37 +747,9 @@ export const geminiCliProvider = {
   },
 
   /**
-   * Get model configuration for a specific model (alias-aware, prefix-aware).
+   * Get model configuration for a specific model (alias-aware).
    */
   getModelConfig(modelName) {
-    if (typeof modelName !== 'string') return null;
-
-    const name = modelName.toLowerCase().trim();
-
-    // Full agy display-name passthrough → matching tier config. Any 3.x Flash
-    // (agy also lists 3.6/3.7) shares the Flash config; any 3.x Pro the Pro one.
-    if (/gemini 3\.\d+ flash/i.test(modelName)) {
-      return SUPPORTED_MODELS['gemini:flash'];
-    }
-    if (/gemini 3\.\d+ pro/i.test(modelName)) {
-      return SUPPORTED_MODELS['gemini:pro'];
-    }
-
-    // Exact key match (gemini, gemini:flash, gemini:pro)
-    if (SUPPORTED_MODELS[name]) {
-      return SUPPORTED_MODELS[name];
-    }
-
-    // Alias match
-    for (const config of Object.values(SUPPORTED_MODELS)) {
-      if (
-        config.aliases &&
-        config.aliases.some((alias) => alias.toLowerCase() === name)
-      ) {
-        return config;
-      }
-    }
-
-    return null;
+    return findCatalogEntry(SUPPORTED_MODELS, modelName);
   },
 };

package/src/providers/google.js

@@ -434,7 +434,11 @@ async function retryWithBackoff(fn, maxRetries = 4) {
 /**
  * Main Google provider implementation
  */
+const DEFAULT_MODEL = 'gemini-3.1-pro-preview';
+
 export const googleProvider = {
+  defaultModel: DEFAULT_MODEL,
+
   /**
    * Unified provider interface: invoke messages with options
    * @param {Array} messages - Array of message objects with role and content
@@ -443,7 +447,7 @@ export const googleProvider = {
    */
   async invoke(messages, options = {}) {
     const {
-      model = 'gemini-2.5-flash',
+      model = DEFAULT_MODEL,
       maxTokens = null,
       stream = false,
       reasoning_effort = 'medium',

package/src/providers/mistral.js

@@ -319,7 +319,11 @@ function extractRateLimitInfo(headers) {
 /**
  * Main Mistral provider implementation
  */
+const DEFAULT_MODEL = 'mistral-medium-3-5';
+
 export const mistralProvider = {
+  defaultModel: DEFAULT_MODEL,
+
   /**
    * Unified provider interface: invoke messages with options
    * @param {Array} messages - Array of message objects with role and content
@@ -328,7 +332,7 @@ export const mistralProvider = {
    */
   async invoke(messages, options = {}) {
     const {
-      model = 'mistral-medium-3-5',
+      model = DEFAULT_MODEL,
       maxTokens = null,
       stream = false,
       reasoning_effort = 'medium',

package/src/providers/openai-compatible.js

@@ -273,6 +273,7 @@ export function createOpenAICompatibleProvider(providerConfig) {
     transformStreamChunk,
     resolveModelConfig,
     defaultParams = {},
+    defaultModel = Object.keys(supportedModels)[0],
   } = providerConfig;
 
   // Create custom error class for this provider
@@ -284,12 +285,14 @@ export function createOpenAICompatibleProvider(providerConfig) {
   }
 
   return {
+    defaultModel,
+
     /**
      * Unified provider interface: invoke messages with options
      */
     async invoke(messages, options = {}) {
       const {
-        model = Object.keys(supportedModels)[0], // Default to first model
+        model = defaultModel,
         maxTokens = null,
         stream = false,
         reasoning_effort = 'medium',

package/src/providers/openai.js

@@ -547,7 +547,11 @@ function convertMessages(messages, useResponsesAPI = false) {
 /**
  * Main OpenAI provider implementation
  */
+const DEFAULT_MODEL = 'gpt-6-sol';
+
 export const openaiProvider = {
+  defaultModel: DEFAULT_MODEL,
+
   /**
    * Unified provider interface: invoke messages with options
    * @param {Array} messages - Array of message objects with role and content
@@ -556,7 +560,7 @@ export const openaiProvider = {
    */
   async invoke(messages, options = {}) {
     const {
-      model = 'gpt-6',
+      model = DEFAULT_MODEL,
       maxTokens = null,
       stream = false,
       reasoning_effort = 'medium',

package/src/providers/openrouter.js

@@ -203,7 +203,7 @@ function validateApiKey(apiKey) {
  * `X-Title`), while still accepting the legacy `openrouterreferer`/
  * `openroutertitle` config-key spellings as input.
  */
-function getCustomHeaders(config) {
+export function getCustomHeaders(config) {
   const headers = {};
 
   const referer =
@@ -550,6 +550,7 @@ export const openrouterProvider = createOpenAICompatibleProvider({
   baseURL: 'https://openrouter.ai/api/v1',
   providerName: 'OpenRouter',
   supportedModels: SUPPORTED_MODELS,
+  defaultModel: 'z-ai/glm-5.2',
   validateApiKey,
   transformRequest,
   transformResponse,

package/src/providers/xai.js

@@ -212,7 +212,11 @@ function extractCitations(response) {
 /**
  * Main XAI provider implementation
  */
+const DEFAULT_MODEL = 'grok-4.5';
+
 export const xaiProvider = {
+  defaultModel: DEFAULT_MODEL,
+
   /**
    * Unified provider interface: invoke messages with options
    * @param {Array} messages - Array of message objects with role and content
@@ -221,7 +225,7 @@ export const xaiProvider = {
    */
   async invoke(messages, options = {}) {
     const {
-      model = 'grok-4.5',
+      model = DEFAULT_MODEL,
       maxTokens = null,
       stream = false,
       reasoning_effort = 'medium',

package/src/services/summarizationService.js

@@ -9,20 +9,21 @@
 import { createLogger } from '../utils/logger.js';
 import { debugLog, debugError } from '../utils/console.js';
 
-import { mapModelToProvider } from '../utils/modelRouting.js';
+import { resolveModelSpec } from '../utils/modelRouting.js';
 
 const logger = createLogger('summarization');
 
-// Default fast models for summarization tasks (prioritize GPT-5-nano for speed)
-const FAST_MODELS = {
-  openai: 'gpt-5-nano', // Fastest GPT-5 model with minimal reasoning
-  google: 'flash',
-  xai: 'grok-4.5',
-  anthropic: 'claude-3-5-haiku-latest',
-  mistral: 'mistral-small-latest',
-  deepseek: 'deepseek-v4-flash',
-  openrouter: 'z-ai/glm-5.2',
-};
+// Fast models for summarization tasks, tried in order (GPT-5-nano first for
+// speed). Namespaced so each names exactly one provider.
+export const FAST_MODELS = [
+  'openai:gpt-5-nano',
+  'google:gemini-2.5-flash',
+  'xai:grok-4.5',
+  'anthropic:claude-haiku-4-5-20251001',
+  'mistral:mistral-small-2603',
+  'deepseek:deepseek-v4-flash',
+  'openrouter:z-ai/glm-5.2',
+];
 
 export class SummarizationService {
   constructor(providers, config) {
@@ -51,16 +52,12 @@ export class SummarizationService {
 
     try {
       // Select fast model if not specified
-      const selectedModel = model || this._selectFastModel();
-      const providerName = mapModelToProvider(selectedModel, this.providers);
-      const provider = this.providers[providerName];
-
-      if (!provider || !provider.isAvailable(this.config)) {
-        debugLog(
-          `Summarization: Provider ${providerName} not available for title generation`,
-        );
+      const selected = this._resolve(model || this._selectFastModel());
+      if (!selected) {
+        debugLog('Summarization: No available model for title generation');
         return this._fallbackTitle(prompt);
       }
+      const { provider, resolvedModel: selectedModel } = selected;
 
       // Create messages for title generation
       const messages = [
@@ -112,16 +109,12 @@ export class SummarizationService {
 
     try {
       // Select fast model if not specified
-      const selectedModel = model || this._selectFastModel();
-      const providerName = mapModelToProvider(selectedModel, this.providers);
-      const provider = this.providers[providerName];
-
-      if (!provider || !provider.isAvailable(this.config)) {
-        debugLog(
-          `Summarization: Provider ${providerName} not available for streaming summary`,
-        );
+      const selected = this._resolve(model || this._selectFastModel());
+      if (!selected) {
+        debugLog('Summarization: No available model for streaming summary');
         return this._fallbackStreamingSummary(content, currentFocus);
       }
+      const { provider, resolvedModel: selectedModel } = selected;
 
       // Create messages for streaming summary
       const messages = [
@@ -175,16 +168,12 @@ export class SummarizationService {
 
     try {
       // Select fast model if not specified
-      const selectedModel = model || this._selectFastModel();
-      const providerName = mapModelToProvider(selectedModel, this.providers);
-      const provider = this.providers[providerName];
-
-      if (!provider || !provider.isAvailable(this.config)) {
-        debugLog(
-          `Summarization: Provider ${providerName} not available for final summary`,
-        );
+      const selected = this._resolve(model || this._selectFastModel());
+      if (!selected) {
+        debugLog('Summarization: No available model for final summary');
         return this._fallbackFinalSummary(content);
       }
+      const { provider, resolvedModel: selectedModel } = selected;
 
       // Create messages for final summary
       const messages = [
@@ -221,6 +210,20 @@ export class SummarizationService {
     }
   }
 
+  /**
+   * Route a model spec to its first available provider.
+   * @private
+   * @returns {{ provider: object, providerName: string, resolvedModel: string }|null}
+   */
+  _resolve(spec) {
+    const resolution = resolveModelSpec(spec, this.providers, this.config);
+    if (resolution.status !== 'ok') {
+      debugLog(`Summarization: ${resolution.error}`);
+      return null;
+    }
+    return resolution;
+  }
+
   /**
    * Select the best available fast model
    * @private
@@ -228,14 +231,10 @@ export class SummarizationService {
   _selectFastModel() {
     // If a model is configured, try to use it first
     if (this.configuredModel) {
-      const providerName = mapModelToProvider(
-        this.configuredModel,
-        this.providers,
-      );
-      const provider = this.providers[providerName];
-      if (provider && provider.isAvailable(this.config)) {
+      const resolution = resolveModelSpec(this.configuredModel, this.providers, this.config);
+      if (resolution.status === 'ok') {
         debugLog(
-          `Summarization: Using configured model ${this.configuredModel} from ${providerName}`,
+          `Summarization: Using configured model ${this.configuredModel} from ${resolution.providerName}`,
         );
         return this.configuredModel;
       }
@@ -244,13 +243,10 @@ export class SummarizationService {
       );
     }
 
-    // Check which providers are available and return the first fast model
-    for (const [providerName, fastModel] of Object.entries(FAST_MODELS)) {
-      const provider = this.providers[providerName];
-      if (provider && provider.isAvailable(this.config)) {
-        debugLog(
-          `Summarization: Selected fast model ${fastModel} from ${providerName}`,
-        );
+    // Return the first fast model whose provider is available
+    for (const fastModel of FAST_MODELS) {
+      if (resolveModelSpec(fastModel, this.providers, this.config).status === 'ok') {
+        debugLog(`Summarization: Selected fast model ${fastModel}`);
         return fastModel;
       }
     }