feat(antigravity): add model specs and update mapping logic

src/tests/AntigravityCoreFeatures.test.ts

@@ -79,7 +79,7 @@ describe('Antigravity Core Features (Business Logic)', () => {
       const requestWithTools = {
         ...BASE_REQUEST,
         // Maps to a gemini-3 model potentially
-        model: 'gemini-3-pro-preview',
+        model: 'gemini-3.1-pro-preview',
         tools: [{ name: 'get_weather' }], // Mock tool structure
         thinking: { type: 'enabled', budget_tokens: 1000 },
       } as any;
@@ -96,7 +96,7 @@ describe('Antigravity Core Features (Business Logic)', () => {
 
       const requestWithTools = {
         ...BASE_REQUEST,
-        model: 'gemini-3-pro-preview',
+        model: 'gemini-3.1-pro-preview',
         tools: [{ name: 'get_weather' }],
         thinking: { type: 'enabled', budget_tokens: 1000 },
       } as any;
@@ -110,7 +110,7 @@ describe('Antigravity Core Features (Business Logic)', () => {
     it('should KEEP thinking mode if NO function calls exist (pure thinking)', () => {
       const requestPureThinking = {
         ...BASE_REQUEST,
-        model: 'gemini-3-pro-preview',
+        model: 'gemini-3.1-pro-preview',
         // No tools
         thinking: { type: 'enabled', budget_tokens: 1000 },
       } as any;
@@ -120,5 +120,60 @@ describe('Antigravity Core Features (Business Logic)', () => {
       // Should keep thinking config
       expect(result.request.generationConfig?.thinkingConfig).toBeDefined();
     });
+
+    it('should auto-enable thinking for gemini-3-flash even without explicit thinking config', () => {
+      const result = transformClaudeRequestIn(
+        {
+          ...BASE_REQUEST,
+          model: 'gemini-3-flash',
+          thinking: undefined,
+        } as any,
+        'test-project',
+      );
+
+      expect(result.request.generationConfig?.thinkingConfig).toBeDefined();
+      expect(result.request.generationConfig?.thinkingConfig?.includeThoughts).toBe(true);
+    });
+
+    it('should inject flash sentinel thoughtSignature for tool_use without cached signature', () => {
+      const requestWithToolUse = {
+        ...BASE_REQUEST,
+        model: 'gemini-3-flash',
+        thinking: undefined,
+        messages: [
+          {
+            role: 'assistant',
+            content: [
+              {
+                type: 'tool_use',
+                id: 'call_1',
+                name: 'get_weather',
+                input: { location: 'Beijing' },
+              },
+            ],
+          },
+        ],
+      } as any;
+
+      const result = transformClaudeRequestIn(requestWithToolUse, 'test-project');
+      const part = result.request.contents?.[0]?.parts?.[0] as any;
+      expect(part?.thoughtSignature).toBe('skip_thought_signature_validator');
+    });
+
+    it('should keep mixed tools for gemini 2.x+ when both function and web search are requested', () => {
+      const mixedToolsRequest = {
+        ...BASE_REQUEST,
+        model: 'gemini-2.0-flash-online',
+        tools: [{ name: 'get_weather' }, { name: 'web_search' }],
+      } as any;
+
+      const result = transformClaudeRequestIn(mixedToolsRequest, 'test-project');
+      const tools = result.request.tools || [];
+      const hasFunctions = tools.some((tool: any) => Array.isArray(tool.functionDeclarations));
+      const hasGoogleSearch = tools.some((tool: any) => tool.googleSearch);
+
+      expect(hasFunctions).toBe(true);
+      expect(hasGoogleSearch).toBe(true);
+    });
   });
 });

src/tests/scripts/project-id-request-validation.ts

@@ -8,6 +8,8 @@ const mockTokenManager: any = {
   getNextToken: async () => null,
   markAsRateLimited: () => undefined,
   markAsForbidden: () => undefined,
+  markFromUpstreamError: () => undefined,
+  recordParityError: () => undefined,
 };
 
 const mockGeminiClient: any = {
@@ -225,7 +227,11 @@ async function validateRuntimeAnthropicRequestFromRealTokenManager(): Promise<vo
   await realTokenManager.onModuleInit();
 
   const tokenManagerProxy = {
-    getNextToken: async (options?: { sessionKey?: string; excludeAccountIds?: string[] }) => {
+    getNextToken: async (options?: {
+      sessionKey?: string;
+      excludeAccountIds?: string[];
+      model?: string;
+    }) => {
       observedGetNextTokenOptions = options ?? null;
       selectedToken = await realTokenManager.getNextToken(options);
       return selectedToken;
@@ -234,6 +240,14 @@ async function validateRuntimeAnthropicRequestFromRealTokenManager(): Promise<vo
       realTokenManager.markAsRateLimited(accountIdOrEmail),
     markAsForbidden: (accountIdOrEmail: string) =>
       realTokenManager.markAsForbidden(accountIdOrEmail),
+    markFromUpstreamError: (args: {
+      accountIdOrEmail: string;
+      status?: number;
+      retryAfter?: string;
+      body?: string;
+      model?: string;
+    }) => realTokenManager.markFromUpstreamError(args),
+    recordParityError: () => realTokenManager.recordParityError(),
   };
 
   try {