Add media fallback rectifier for text-only models

Replace unsupported Anthropic image blocks with an [Unsupported Image] marker when routed models are text-only or upstream rejects image input.

Add rectifier settings for media fallback and heuristic model detection, wire the controls into the settings UI, and cover the sanitizer and forwarder gates with regression tests.

src/components/settings/RectifierConfigPanel.tsx

@@ -15,6 +15,8 @@ export function RectifierConfigPanel() {
     enabled: true,
     requestThinkingSignature: true,
     requestThinkingBudget: true,
+    requestMediaFallback: true,
+    requestMediaHeuristic: true,
   });
   const [optimizerConfig, setOptimizerConfig] = useState<OptimizerConfig>({
     enabled: false,
@@ -111,6 +113,36 @@ export function RectifierConfigPanel() {
             }
           />
         </div>
+        <div className="flex items-center justify-between pl-4">
+          <div className="space-y-0.5">
+            <Label>{t("settings.advanced.rectifier.mediaFallback")}</Label>
+            <p className="text-xs text-muted-foreground">
+              {t("settings.advanced.rectifier.mediaFallbackDescription")}
+            </p>
+          </div>
+          <Switch
+            checked={config.requestMediaFallback}
+            disabled={!config.enabled}
+            onCheckedChange={(checked) =>
+              handleChange({ requestMediaFallback: checked })
+            }
+          />
+        </div>
+        <div className="flex items-center justify-between pl-8">
+          <div className="space-y-0.5">
+            <Label>{t("settings.advanced.rectifier.mediaHeuristic")}</Label>
+            <p className="text-xs text-muted-foreground">
+              {t("settings.advanced.rectifier.mediaHeuristicDescription")}
+            </p>
+          </div>
+          <Switch
+            checked={config.requestMediaHeuristic}
+            disabled={!config.enabled || !config.requestMediaFallback}
+            onCheckedChange={(checked) =>
+              handleChange({ requestMediaHeuristic: checked })
+            }
+          />
+        </div>
       </div>
 
       <div className="border-t pt-6 mt-6">

src/i18n/locales/en.json

@@ -367,7 +367,11 @@
         "thinkingSignature": "Thinking Signature Rectification",
         "thinkingSignatureDescription": "When an Anthropic-type provider returns thinking signature incompatibility or illegal request errors, automatically removes incompatible thinking-related blocks and retries once with the same provider",
         "thinkingBudget": "Thinking Budget Rectification",
-        "thinkingBudgetDescription": "When an Anthropic-type provider returns budget_tokens constraint errors (such as at least 1024), automatically normalizes thinking to enabled, sets thinking budget to 32000, and raises max_tokens to 64000 if needed, then retries once"
+        "thinkingBudgetDescription": "When an Anthropic-type provider returns budget_tokens constraint errors (such as at least 1024), automatically normalizes thinking to enabled, sets thinking budget to 32000, and raises max_tokens to 64000 if needed, then retries once",
+        "mediaFallback": "Unsupported Image Fallback",
+        "mediaFallbackDescription": "When a provider does not support image input (it declares text-only capability, or the upstream returns an unsupported-image error), automatically replaces image blocks with an [Unsupported Image] marker so the conversation is not interrupted",
+        "mediaHeuristic": "Heuristic Text-Only Model Detection",
+        "mediaHeuristicDescription": "Predictively strips images before sending based on a built-in model-name list. When off, only the declared-text-only and upstream-error fallback paths remain, preventing the built-in list from misclassifying a multimodal model as text-only and silently dropping images"
       },
       "optimizer": {
         "title": "Bedrock Request Optimizer",

src/i18n/locales/ja.json

@@ -367,7 +367,11 @@
         "thinkingSignature": "Thinking 署名整流",
         "thinkingSignatureDescription": "Anthropic タイプのプロバイダーが thinking 署名の非互換性や不正なリクエストエラーを返した場合、互換性のない thinking 関連ブロックを自動削除し、同じプロバイダーで 1 回リトライします",
         "thinkingBudget": "Thinking Budget 整流",
-        "thinkingBudgetDescription": "Anthropic タイプのプロバイダーが budget_tokens 制約エラー（例: 1024 以上）を返した場合、thinking を enabled に正規化し、thinking 予算を 32000 に設定し、必要に応じて max_tokens を 64000 に引き上げて 1 回リトライします"
+        "thinkingBudgetDescription": "Anthropic タイプのプロバイダーが budget_tokens 制約エラー（例: 1024 以上）を返した場合、thinking を enabled に正規化し、thinking 予算を 32000 に設定し、必要に応じて max_tokens を 64000 に引き上げて 1 回リトライします",
+        "mediaFallback": "非対応画像フォールバック",
+        "mediaFallbackDescription": "プロバイダーが画像入力に対応していない場合（テキストのみの能力が宣言されている、または上流が画像非対応エラーを返した場合）、画像ブロックを [Unsupported Image] マーカーに自動置換し、会話を中断させません",
+        "mediaHeuristic": "ヒューリスティックなテキスト専用モデル判定",
+        "mediaHeuristicDescription": "送信前に内蔵モデル名リストに基づいて画像を予測的に削除します。オフにすると「宣言済みテキスト専用」と「上流エラー時のフォールバック」のみが残り、内蔵リストがマルチモーダルモデルをテキスト専用と誤判定して画像を静かに削除するのを防ぎます"
       },
       "optimizer": {
         "title": "Bedrock リクエストオプティマイザー",

src/i18n/locales/zh-TW.json

@@ -367,7 +367,11 @@
         "thinkingSignature": "Thinking 簽章整流",
         "thinkingSignatureDescription": "當 Anthropic 類型供應商回傳 thinking 簽章不相容或非法請求等錯誤時，自動移除不相容的 thinking 相關區塊，並對同一供應商重試一次",
         "thinkingBudget": "Thinking Budget 整流",
-        "thinkingBudgetDescription": "當 Anthropic 類型供應商回傳 budget_tokens 限制錯誤（如至少 1024）時，自動將 thinking 規範為 enabled，並將 budget 設為 32000，同時在需要時將 max_tokens 設為 64000，然後重試一次"
+        "thinkingBudgetDescription": "當 Anthropic 類型供應商回傳 budget_tokens 限制錯誤（如至少 1024）時，自動將 thinking 規範為 enabled，並將 budget 設為 32000，同時在需要時將 max_tokens 設為 64000，然後重試一次",
+        "mediaFallback": "不支援圖片降級",
+        "mediaFallbackDescription": "當供應商不支援圖片輸入時（已宣告純文字能力，或上游回傳不支援圖片的錯誤），自動將圖片區塊替換為 [Unsupported Image] 標記，使對話不中斷",
+        "mediaHeuristic": "啟發式純文字模型辨識",
+        "mediaHeuristicDescription": "傳送前依內建模型名稱清單預測性地移除圖片。關閉後僅保留「已宣告純文字」與「上游錯誤兜底」兩條事實驅動路徑，避免內建清單將多模態模型誤判為純文字而靜默丟失圖片"
       },
       "optimizer": {
         "title": "Bedrock 請求最佳化工具",

src/i18n/locales/zh.json

@@ -367,7 +367,11 @@
         "thinkingSignature": "Thinking 签名整流",
         "thinkingSignatureDescription": "当 Anthropic 类型供应商返回 thinking 签名不兼容或非法请求等错误时，自动移除不兼容的 thinking 相关块并对同一供应商重试一次",
         "thinkingBudget": "Thinking Budget 整流",
-        "thinkingBudgetDescription": "当 Anthropic 类型供应商返回 budget_tokens 约束错误（如至少 1024）时，自动将 thinking 规范为 enabled 并将 budget 设为 32000，同时在需要时将 max_tokens 设为 64000，然后重试一次"
+        "thinkingBudgetDescription": "当 Anthropic 类型供应商返回 budget_tokens 约束错误（如至少 1024）时，自动将 thinking 规范为 enabled 并将 budget 设为 32000，同时在需要时将 max_tokens 设为 64000，然后重试一次",
+        "mediaFallback": "不支持图片降级",
+        "mediaFallbackDescription": "当供应商不支持图片输入时（已声明纯文本能力，或上游返回不支持图片的错误），自动将图片块替换为 [Unsupported Image] 标记，使对话不中断",
+        "mediaHeuristic": "启发式纯文本模型识别",
+        "mediaHeuristicDescription": "发送前按内置模型名列表预测性剥离图片。关闭后仅保留「已声明纯文本」与「上游报错兜底」两条事实驱动路径，避免内置列表把多模态模型误判为纯文本而静默丢图"
       },
       "optimizer": {
         "title": "Bedrock 请求优化器",

src/lib/api/settings.ts

@@ -265,6 +265,8 @@ export interface RectifierConfig {
   enabled: boolean;
   requestThinkingSignature: boolean;
   requestThinkingBudget: boolean;
+  requestMediaFallback: boolean;
+  requestMediaHeuristic: boolean;
 }
 
 export interface OptimizerConfig {