fix(tests): update model lookup references and enhance Claude executor tests

internal/registry/model_registry_safety_test.go

@@ -136,13 +136,13 @@ func TestGetAvailableModelsReturnsClonedSupportedParameters(t *testing.T) {
 }
 
 func TestLookupModelInfoReturnsCloneForStaticDefinitions(t *testing.T) {
-	first := LookupModelInfo("glm-4.6")
+	first := LookupModelInfo("claude-sonnet-4-6")
 	if first == nil || first.Thinking == nil || len(first.Thinking.Levels) == 0 {
 		t.Fatalf("expected static model with thinking levels, got %+v", first)
 	}
 	first.Thinking.Levels[0] = "mutated"
 
-	second := LookupModelInfo("glm-4.6")
+	second := LookupModelInfo("claude-sonnet-4-6")
 	if second == nil || second.Thinking == nil || len(second.Thinking.Levels) == 0 || second.Thinking.Levels[0] == "mutated" {
 		t.Fatalf("expected static lookup clone, got %+v", second)
 	}

internal/runtime/executor/claude_executor_test.go

@@ -1714,7 +1714,27 @@ func TestClaudeExecutor_ExecuteStream_AcceptEncodingOverrideCannotBypassIdentity
 	}
 }
 
-// Test case 1: String system prompt is preserved and converted to a content block
+func expectedClaudeCodeStaticPrompt() string {
+	return strings.Join([]string{
+		helps.ClaudeCodeIntro,
+		helps.ClaudeCodeSystem,
+		helps.ClaudeCodeDoingTasks,
+		helps.ClaudeCodeToneAndStyle,
+		helps.ClaudeCodeOutputEfficiency,
+	}, "\n\n")
+}
+
+func expectedForwardedSystemReminder(text string) string {
+	return fmt.Sprintf(`<system-reminder>
+As you answer the user's questions, you can use the following context from the system:
+%s
+
+IMPORTANT: this context may or may not be relevant to your tasks. You should not respond to this context unless it is highly relevant to your task.
+</system-reminder>
+`, text)
+}
+
+// Test case 1: String system prompt is preserved by forwarding it to the first user message
 func TestCheckSystemInstructionsWithMode_StringSystemPreserved(t *testing.T) {
 	payload := []byte(`{"system":"You are a helpful assistant.","messages":[{"role":"user","content":"hi"}]}`)
 
@@ -1733,42 +1753,52 @@ func TestCheckSystemInstructionsWithMode_StringSystemPreserved(t *testing.T) {
 	if !strings.HasPrefix(blocks[0].Get("text").String(), "x-anthropic-billing-header:") {
 		t.Fatalf("blocks[0] should be billing header, got %q", blocks[0].Get("text").String())
 	}
-	if blocks[1].Get("text").String() != "You are a Claude agent, built on Anthropic's Claude Agent SDK." {
+	if blocks[1].Get("text").String() != "You are Claude Code, Anthropic's official CLI for Claude." {
 		t.Fatalf("blocks[1] should be agent block, got %q", blocks[1].Get("text").String())
 	}
-	if blocks[2].Get("text").String() != "You are a helpful assistant." {
-		t.Fatalf("blocks[2] should be user system prompt, got %q", blocks[2].Get("text").String())
+	if blocks[2].Get("text").String() != expectedClaudeCodeStaticPrompt() {
+		t.Fatalf("blocks[2] should be static Claude Code prompt, got %q", blocks[2].Get("text").String())
 	}
-	if blocks[2].Get("cache_control.type").String() != "ephemeral" {
-		t.Fatalf("blocks[2] should have cache_control.type=ephemeral")
+	if blocks[2].Get("cache_control").Exists() {
+		t.Fatalf("blocks[2] should not have cache_control, got %s", blocks[2].Get("cache_control").Raw)
+	}
+
+	if got := gjson.GetBytes(out, "messages.0.content").String(); got != expectedForwardedSystemReminder("You are a helpful assistant.")+"hi" {
+		t.Fatalf("messages[0].content should include forwarded system prompt, got %q", got)
 	}
 }
 
-// Test case 2: Strict mode drops the string system prompt
+// Test case 2: Strict mode keeps only the injected Claude Code system blocks
 func TestCheckSystemInstructionsWithMode_StringSystemStrict(t *testing.T) {
 	payload := []byte(`{"system":"You are a helpful assistant.","messages":[{"role":"user","content":"hi"}]}`)
 
 	out := checkSystemInstructionsWithMode(payload, true)
 
 	blocks := gjson.GetBytes(out, "system").Array()
-	if len(blocks) != 2 {
-		t.Fatalf("strict mode should produce 2 blocks, got %d", len(blocks))
+	if len(blocks) != 3 {
+		t.Fatalf("strict mode should produce 3 injected blocks, got %d", len(blocks))
+	}
+	if got := gjson.GetBytes(out, "messages.0.content").String(); got != "hi" {
+		t.Fatalf("strict mode should not forward system prompt into messages, got %q", got)
 	}
 }
 
-// Test case 3: Empty string system prompt does not produce a spurious block
+// Test case 3: Empty string system prompt does not alter the first user message
 func TestCheckSystemInstructionsWithMode_EmptyStringSystemIgnored(t *testing.T) {
 	payload := []byte(`{"system":"","messages":[{"role":"user","content":"hi"}]}`)
 
 	out := checkSystemInstructionsWithMode(payload, false)
 
 	blocks := gjson.GetBytes(out, "system").Array()
-	if len(blocks) != 2 {
-		t.Fatalf("empty string system should produce 2 blocks, got %d", len(blocks))
+	if len(blocks) != 3 {
+		t.Fatalf("empty string system should still produce 3 injected blocks, got %d", len(blocks))
+	}
+	if got := gjson.GetBytes(out, "messages.0.content").String(); got != "hi" {
+		t.Fatalf("empty string system should not alter messages, got %q", got)
 	}
 }
 
-// Test case 4: Array system prompt is unaffected by the string handling
+// Test case 4: Array system prompt is forwarded to the first user message
 func TestCheckSystemInstructionsWithMode_ArraySystemStillWorks(t *testing.T) {
 	payload := []byte(`{"system":[{"type":"text","text":"Be concise."}],"messages":[{"role":"user","content":"hi"}]}`)
 
@@ -1778,12 +1808,15 @@ func TestCheckSystemInstructionsWithMode_ArraySystemStillWorks(t *testing.T) {
 	if len(blocks) != 3 {
 		t.Fatalf("expected 3 system blocks, got %d", len(blocks))
 	}
-	if blocks[2].Get("text").String() != "Be concise." {
-		t.Fatalf("blocks[2] should be user system prompt, got %q", blocks[2].Get("text").String())
+	if blocks[2].Get("text").String() != expectedClaudeCodeStaticPrompt() {
+		t.Fatalf("blocks[2] should be static Claude Code prompt, got %q", blocks[2].Get("text").String())
+	}
+	if got := gjson.GetBytes(out, "messages.0.content").String(); got != expectedForwardedSystemReminder("Be concise.")+"hi" {
+		t.Fatalf("messages[0].content should include forwarded array system prompt, got %q", got)
 	}
 }
 
-// Test case 5: Special characters in string system prompt survive conversion
+// Test case 5: Special characters in string system prompt survive forwarding
 func TestCheckSystemInstructionsWithMode_StringWithSpecialChars(t *testing.T) {
 	payload := []byte(`{"system":"Use <xml> tags & \"quotes\" in output.","messages":[{"role":"user","content":"hi"}]}`)
 
@@ -1793,8 +1826,8 @@ func TestCheckSystemInstructionsWithMode_StringWithSpecialChars(t *testing.T) {
 	if len(blocks) != 3 {
 		t.Fatalf("expected 3 system blocks, got %d", len(blocks))
 	}
-	if blocks[2].Get("text").String() != `Use <xml> tags & "quotes" in output.` {
-		t.Fatalf("blocks[2] text mangled, got %q", blocks[2].Get("text").String())
+	if got := gjson.GetBytes(out, "messages.0.content").String(); got != expectedForwardedSystemReminder(`Use <xml> tags & "quotes" in output.`)+"hi" {
+		t.Fatalf("forwarded system prompt text mangled, got %q", got)
 	}
 }
 
@@ -1902,8 +1935,11 @@ func TestApplyCloaking_PreservesConfiguredStrictModeAndSensitiveWordsWhenModeOmi
 	out := applyCloaking(context.Background(), cfg, auth, payload, "claude-3-5-sonnet-20241022", "key-123")
 
 	blocks := gjson.GetBytes(out, "system").Array()
-	if len(blocks) != 2 {
-		t.Fatalf("expected strict mode to keep only injected system blocks, got %d", len(blocks))
+	if len(blocks) != 3 {
+		t.Fatalf("expected strict mode to keep the 3 injected Claude Code system blocks, got %d", len(blocks))
+	}
+	if got := gjson.GetBytes(out, "messages.0.content.#").Int(); got != 1 {
+		t.Fatalf("strict mode should not prepend a forwarded system reminder block, got %d content blocks", got)
 	}
 	if got := gjson.GetBytes(out, "messages.0.content.0.text").String(); !strings.Contains(got, "\u200B") {
 		t.Fatalf("expected configured sensitive word obfuscation to apply, got %q", got)

test/thinking_conversion_test.go

@@ -2,7 +2,6 @@ package test
 
 import (
 	"fmt"
-	"strings"
 	"testing"
 	"time"