Fixed: #897

refactor(executor): remove `prompt_cache_retention` from request payloads

config.example.yaml

@@ -202,6 +202,7 @@ ws-auth: false
 # These mappings rename model IDs for both model listing and request routing.
 # Supported channels: gemini-cli, vertex, aistudio, antigravity, claude, codex, qwen, iflow.
 # NOTE: Mappings do not apply to gemini-api-key, codex-api-key, claude-api-key, openai-compatibility, vertex-api-key, or ampcode.
+# You can repeat the same name with different aliases to expose multiple client model names.
 # oauth-model-mappings:
 #   gemini-cli:
 #     - name: "gemini-2.5-pro"          # original model name under this channel

internal/cache/signature_cache.go

@@ -3,7 +3,6 @@ package cache
 import (
 	"crypto/sha256"
 	"encoding/hex"
-	"sort"
 	"sync"
 	"time"
 )
@@ -16,10 +15,7 @@ type SignatureEntry struct {
 
 const (
 	// SignatureCacheTTL is how long signatures are valid
-	SignatureCacheTTL = 1 * time.Hour
-
-	// MaxEntriesPerSession limits memory usage per session
-	MaxEntriesPerSession = 100
+	SignatureCacheTTL = 3 * time.Hour
 
 	// SignatureTextHashLen is the length of the hash key (16 hex chars = 64-bit key space)
 	SignatureTextHashLen = 16
@@ -112,43 +108,6 @@ func CacheSignature(sessionID, text, signature string) {
 	sc.mu.Lock()
 	defer sc.mu.Unlock()
 
-	// Evict expired entries if at capacity
-	if len(sc.entries) >= MaxEntriesPerSession {
-		now := time.Now()
-		for key, entry := range sc.entries {
-			if now.Sub(entry.Timestamp) > SignatureCacheTTL {
-				delete(sc.entries, key)
-			}
-		}
-		// If still at capacity, remove oldest entries
-		if len(sc.entries) >= MaxEntriesPerSession {
-			// Find and remove oldest quarter
-			oldest := make([]struct {
-				key string
-				ts  time.Time
-			}, 0, len(sc.entries))
-			for key, entry := range sc.entries {
-				oldest = append(oldest, struct {
-					key string
-					ts  time.Time
-				}{key, entry.Timestamp})
-			}
-			// Sort by timestamp (oldest first) using sort.Slice
-			sort.Slice(oldest, func(i, j int) bool {
-				return oldest[i].ts.Before(oldest[j].ts)
-			})
-
-			toRemove := len(oldest) / 4
-			if toRemove < 1 {
-				toRemove = 1
-			}
-
-			for i := 0; i < toRemove; i++ {
-				delete(sc.entries, oldest[i].key)
-			}
-		}
-	}
-
 	sc.entries[textHash] = SignatureEntry{
 		Signature: signature,
 		Timestamp: time.Now(),
@@ -170,22 +129,25 @@ func GetCachedSignature(sessionID, text string) string {
 
 	textHash := hashText(text)
 
-	sc.mu.RLock()
-	entry, exists := sc.entries[textHash]
-	sc.mu.RUnlock()
+	now := time.Now()
 
+	sc.mu.Lock()
+	entry, exists := sc.entries[textHash]
 	if !exists {
+		sc.mu.Unlock()
 		return ""
 	}
-
-	// Check if expired
-	if time.Since(entry.Timestamp) > SignatureCacheTTL {
-		sc.mu.Lock()
+	if now.Sub(entry.Timestamp) > SignatureCacheTTL {
 		delete(sc.entries, textHash)
 		sc.mu.Unlock()
 		return ""
 	}
 
+	// Refresh TTL on access (sliding expiration).
+	entry.Timestamp = now
+	sc.entries[textHash] = entry
+	sc.mu.Unlock()
+
 	return entry.Signature
 }
 

internal/config/config.go

@@ -521,7 +521,7 @@ func LoadConfigOptional(configFile string, optional bool) (*Config, error) {
 
 // SanitizeOAuthModelMappings normalizes and deduplicates global OAuth model name mappings.
 // It trims whitespace, normalizes channel keys to lower-case, drops empty entries,
-// and ensures (From, To) pairs are unique within each channel.
+// allows multiple aliases per upstream name, and ensures aliases are unique within each channel.
 func (cfg *Config) SanitizeOAuthModelMappings() {
 	if cfg == nil || len(cfg.OAuthModelMappings) == 0 {
 		return
@@ -532,7 +532,6 @@ func (cfg *Config) SanitizeOAuthModelMappings() {
 		if channel == "" || len(mappings) == 0 {
 			continue
 		}
-		seenName := make(map[string]struct{}, len(mappings))
 		seenAlias := make(map[string]struct{}, len(mappings))
 		clean := make([]ModelNameMapping, 0, len(mappings))
 		for _, mapping := range mappings {
@@ -544,15 +543,10 @@ func (cfg *Config) SanitizeOAuthModelMappings() {
 			if strings.EqualFold(name, alias) {
 				continue
 			}
-			nameKey := strings.ToLower(name)
 			aliasKey := strings.ToLower(alias)
-			if _, ok := seenName[nameKey]; ok {
-				continue
-			}
 			if _, ok := seenAlias[aliasKey]; ok {
 				continue
 			}
-			seenName[nameKey] = struct{}{}
 			seenAlias[aliasKey] = struct{}{}
 			clean = append(clean, ModelNameMapping{Name: name, Alias: alias, Fork: mapping.Fork})
 		}

internal/config/oauth_model_mappings_test.go

@@ -25,3 +25,32 @@ func TestSanitizeOAuthModelMappings_PreservesForkFlag(t *testing.T) {
 		t.Fatalf("expected second mapping to be gpt-6->g6 fork=false, got name=%q alias=%q fork=%v", mappings[1].Name, mappings[1].Alias, mappings[1].Fork)
 	}
 }
+
+func TestSanitizeOAuthModelMappings_AllowsMultipleAliasesForSameName(t *testing.T) {
+	cfg := &Config{
+		OAuthModelMappings: map[string][]ModelNameMapping{
+			"antigravity": {
+				{Name: "gemini-claude-opus-4-5-thinking", Alias: "claude-opus-4-5-20251101", Fork: true},
+				{Name: "gemini-claude-opus-4-5-thinking", Alias: "claude-opus-4-5-20251101-thinking", Fork: true},
+				{Name: "gemini-claude-opus-4-5-thinking", Alias: "claude-opus-4-5", Fork: true},
+			},
+		},
+	}
+
+	cfg.SanitizeOAuthModelMappings()
+
+	mappings := cfg.OAuthModelMappings["antigravity"]
+	expected := []ModelNameMapping{
+		{Name: "gemini-claude-opus-4-5-thinking", Alias: "claude-opus-4-5-20251101", Fork: true},
+		{Name: "gemini-claude-opus-4-5-thinking", Alias: "claude-opus-4-5-20251101-thinking", Fork: true},
+		{Name: "gemini-claude-opus-4-5-thinking", Alias: "claude-opus-4-5", Fork: true},
+	}
+	if len(mappings) != len(expected) {
+		t.Fatalf("expected %d sanitized mappings, got %d", len(expected), len(mappings))
+	}
+	for i, exp := range expected {
+		if mappings[i].Name != exp.Name || mappings[i].Alias != exp.Alias || mappings[i].Fork != exp.Fork {
+			t.Fatalf("expected mapping %d to be name=%q alias=%q fork=%v, got name=%q alias=%q fork=%v", i, exp.Name, exp.Alias, exp.Fork, mappings[i].Name, mappings[i].Alias, mappings[i].Fork)
+		}
+	}
+}

internal/runtime/executor/codex_executor.go

@@ -106,6 +106,7 @@ func (e *CodexExecutor) Execute(ctx context.Context, auth *cliproxyauth.Auth, re
 	body, _ = sjson.SetBytes(body, "model", model)
 	body, _ = sjson.SetBytes(body, "stream", true)
 	body, _ = sjson.DeleteBytes(body, "previous_response_id")
+	body, _ = sjson.DeleteBytes(body, "prompt_cache_retention")
 
 	url := strings.TrimSuffix(baseURL, "/") + "/responses"
 	httpReq, err := e.cacheHelper(ctx, from, url, req, body)
@@ -214,6 +215,7 @@ func (e *CodexExecutor) ExecuteStream(ctx context.Context, auth *cliproxyauth.Au
 	}
 	body = applyPayloadConfigWithRoot(e.cfg, model, to.String(), "", body, originalTranslated)
 	body, _ = sjson.DeleteBytes(body, "previous_response_id")
+	body, _ = sjson.DeleteBytes(body, "prompt_cache_retention")
 	body, _ = sjson.SetBytes(body, "model", model)
 
 	url := strings.TrimSuffix(baseURL, "/") + "/responses"
@@ -316,6 +318,7 @@ func (e *CodexExecutor) CountTokens(ctx context.Context, auth *cliproxyauth.Auth
 	body = ApplyReasoningEffortMetadata(body, req.Metadata, model, "reasoning.effort", false)
 	body, _ = sjson.SetBytes(body, "model", model)
 	body, _ = sjson.DeleteBytes(body, "previous_response_id")
+	body, _ = sjson.DeleteBytes(body, "prompt_cache_retention")
 	body, _ = sjson.SetBytes(body, "stream", false)
 
 	enc, err := tokenizerForCodexModel(model)

sdk/cliproxy/service.go

@@ -1237,7 +1237,7 @@ func applyOAuthModelMappings(cfg *config.Config, provider, authKind string, mode
 		fork  bool
 	}
 
-	forward := make(map[string]mappingEntry, len(mappings))
+	forward := make(map[string][]mappingEntry, len(mappings))
 	for i := range mappings {
 		name := strings.TrimSpace(mappings[i].Name)
 		alias := strings.TrimSpace(mappings[i].Alias)
@@ -1248,14 +1248,12 @@ func applyOAuthModelMappings(cfg *config.Config, provider, authKind string, mode
 			continue
 		}
 		key := strings.ToLower(name)
-		if _, exists := forward[key]; exists {
-			continue
-		}
-		forward[key] = mappingEntry{alias: alias, fork: mappings[i].Fork}
+		forward[key] = append(forward[key], mappingEntry{alias: alias, fork: mappings[i].Fork})
 	}
 	if len(forward) == 0 {
 		return models
 	}
+
 	out := make([]*ModelInfo, 0, len(models))
 	seen := make(map[string]struct{}, len(models))
 	for _, model := range models {
@@ -1267,17 +1265,8 @@ func applyOAuthModelMappings(cfg *config.Config, provider, authKind string, mode
 			continue
 		}
 		key := strings.ToLower(id)
-		entry, ok := forward[key]
-		if !ok {
-			if _, exists := seen[key]; exists {
-				continue
-			}
-			seen[key] = struct{}{}
-			out = append(out, model)
-			continue
-		}
-		mappedID := strings.TrimSpace(entry.alias)
-		if mappedID == "" {
+		entries := forward[key]
+		if len(entries) == 0 {
 			if _, exists := seen[key]; exists {
 				continue
 			}
@@ -1286,11 +1275,29 @@ func applyOAuthModelMappings(cfg *config.Config, provider, authKind string, mode
 			continue
 		}
 
-		if entry.fork {
+		keepOriginal := false
+		for _, entry := range entries {
+			if entry.fork {
+				keepOriginal = true
+				break
+			}
+		}
+		if keepOriginal {
 			if _, exists := seen[key]; !exists {
 				seen[key] = struct{}{}
 				out = append(out, model)
 			}
+		}
+
+		addedAlias := false
+		for _, entry := range entries {
+			mappedID := strings.TrimSpace(entry.alias)
+			if mappedID == "" {
+				continue
+			}
+			if strings.EqualFold(mappedID, id) {
+				continue
+			}
 			aliasKey := strings.ToLower(mappedID)
 			if _, exists := seen[aliasKey]; exists {
 				continue
@@ -1302,24 +1309,16 @@ func applyOAuthModelMappings(cfg *config.Config, provider, authKind string, mode
 				clone.Name = rewriteModelInfoName(clone.Name, id, mappedID)
 			}
 			out = append(out, &clone)
-			continue
+			addedAlias = true
 		}
 
-		uniqueKey := strings.ToLower(mappedID)
-		if _, exists := seen[uniqueKey]; exists {
-			continue
-		}
-		seen[uniqueKey] = struct{}{}
-		if mappedID == id {
+		if !keepOriginal && !addedAlias {
+			if _, exists := seen[key]; exists {
+				continue
+			}
+			seen[key] = struct{}{}
 			out = append(out, model)
-			continue
 		}
-		clone := *model
-		clone.ID = mappedID
-		if clone.Name != "" {
-			clone.Name = rewriteModelInfoName(clone.Name, id, mappedID)
-		}
-		out = append(out, &clone)
 	}
 	return out
 }

sdk/cliproxy/service_oauth_model_mappings_test.go

@@ -56,3 +56,37 @@ func TestApplyOAuthModelMappings_ForkAddsAlias(t *testing.T) {
 		t.Fatalf("expected forked model name %q, got %q", "models/g5", out[1].Name)
 	}
 }
+
+func TestApplyOAuthModelMappings_ForkAddsMultipleAliases(t *testing.T) {
+	cfg := &config.Config{
+		OAuthModelMappings: map[string][]config.ModelNameMapping{
+			"codex": {
+				{Name: "gpt-5", Alias: "g5", Fork: true},
+				{Name: "gpt-5", Alias: "g5-2", Fork: true},
+			},
+		},
+	}
+	models := []*ModelInfo{
+		{ID: "gpt-5", Name: "models/gpt-5"},
+	}
+
+	out := applyOAuthModelMappings(cfg, "codex", "oauth", models)
+	if len(out) != 3 {
+		t.Fatalf("expected 3 models, got %d", len(out))
+	}
+	if out[0].ID != "gpt-5" {
+		t.Fatalf("expected first model id %q, got %q", "gpt-5", out[0].ID)
+	}
+	if out[1].ID != "g5" {
+		t.Fatalf("expected second model id %q, got %q", "g5", out[1].ID)
+	}
+	if out[1].Name != "models/g5" {
+		t.Fatalf("expected forked model name %q, got %q", "models/g5", out[1].Name)
+	}
+	if out[2].ID != "g5-2" {
+		t.Fatalf("expected third model id %q, got %q", "g5-2", out[2].ID)
+	}
+	if out[2].Name != "models/g5-2" {
+		t.Fatalf("expected forked model name %q, got %q", "models/g5-2", out[2].Name)
+	}
+}