From d98bd8f872bd4382100553ebc55cb789b7c45331 Mon Sep 17 00:00:00 2001
From: Daniel Riccio <ricciodaniel98@gmail.com>
Date: Fri, 5 Sep 2025 11:25:34 -0500
Subject: [PATCH] fix: keep system prompt as plain string for Vertex SDK
 compatibility

The Vertex SDK does not support array-form system prompts with cache_control,
unlike the regular Anthropic SDK. This fix ensures the system parameter remains
a plain string, applying cache_control only to message content blocks.

- Modified system prompt handling to maintain string format
- Updated tests to match corrected behavior
- Maintains caching functionality for message blocks only

Fixes compatibility issue introduced in commit a3d8c0e3f6
---
 .../__tests__/anthropic-vertex.spec.ts         | 18 +++---------------
 src/api/providers/anthropic-vertex.ts          | 11 +++++------
 2 files changed, 8 insertions(+), 21 deletions(-)

diff --git a/src/api/providers/__tests__/anthropic-vertex.spec.ts b/src/api/providers/__tests__/anthropic-vertex.spec.ts
index 9d83f265c7..3e93a9e01b 100644
--- a/src/api/providers/__tests__/anthropic-vertex.spec.ts
+++ b/src/api/providers/__tests__/anthropic-vertex.spec.ts
@@ -163,13 +163,7 @@ describe("VertexHandler", () => {
 				model: "claude-3-5-sonnet-v2@20241022",
 				max_tokens: 8192,
 				temperature: 0,
-				system: [
-					{
-						type: "text",
-						text: "You are a helpful assistant",
-						cache_control: { type: "ephemeral" },
-					},
-				],
+				system: "You are a helpful assistant", // System remains as plain string
 				messages: [
 					{
 						role: "user",
@@ -364,16 +358,10 @@ describe("VertexHandler", () => {
 			expect(textChunks[0].text).toBe("Hello")
 			expect(textChunks[1].text).toBe(" world!")
 
-			// Verify cache control was added correctly
+			// Verify cache control was added correctly - system remains string, only messages have cache_control
 			expect(mockCreate).toHaveBeenCalledWith(
 				expect.objectContaining({
-					system: [
-						{
-							type: "text",
-							text: "You are a helpful assistant",
-							cache_control: { type: "ephemeral" },
-						},
-					],
+					system: "You are a helpful assistant", // System remains as plain string
 					messages: [
 						expect.objectContaining({
 							role: "user",
diff --git a/src/api/providers/anthropic-vertex.ts b/src/api/providers/anthropic-vertex.ts
index c70a15926d..1928e46a6d 100644
--- a/src/api/providers/anthropic-vertex.ts
+++ b/src/api/providers/anthropic-vertex.ts
@@ -75,12 +75,13 @@ export class AnthropicVertexHandler extends BaseProvider implements SingleComple
 		 * 1. Maximum of 4 blocks can have cache_control
 		 * 2. Only text blocks can be cached (images and other content types cannot)
 		 * 3. Cache control can only be applied to user messages, not assistant messages
+		 * 4. System prompt must remain as a plain string - Vertex SDK does not support array format
 		 *
 		 * Our caching strategy:
-		 * - Cache the system prompt (1 block)
+		 * - System prompt cannot be cached directly (must remain as plain string)
 		 * - Cache the last text block of the second-to-last user message (1 block)
 		 * - Cache the last text block of the last user message (1 block)
-		 * This ensures we stay under the 4-block limit while maintaining effective caching
+		 * This ensures compatibility with Vertex SDK while maintaining effective caching
 		 * for the most relevant context.
 		 */
 		const params: Anthropic.Messages.MessageCreateParamsStreaming = {
@@ -88,10 +89,8 @@ export class AnthropicVertexHandler extends BaseProvider implements SingleComple
 			max_tokens: maxTokens ?? ANTHROPIC_DEFAULT_MAX_TOKENS,
 			temperature,
 			thinking,
-			// Cache the system prompt if caching is enabled.
-			system: supportsPromptCache
-				? [{ text: systemPrompt, type: "text" as const, cache_control: { type: "ephemeral" } }]
-				: systemPrompt,
+			// System must remain as plain string for Vertex SDK compatibility
+			system: systemPrompt,
 			messages: supportsPromptCache ? addCacheBreakpoints(messages) : messages,
 			stream: true,
 		}