feat(combos): add imageInput auto|disabled capability control

docs-site/src/content/docs/guides/combos.md

@@ -164,6 +164,14 @@ default and leaves the target's own behavior unchanged. Supported values are `lo
 `high`, `xhigh`, `max`, and `ultra`; omit the field or set it to `null` to leave effort entirely to
 the caller and target.
 
+## Image / multimodal capability
+
+By default a combo publishes the **intersection** of its targets' input modalities (image is
+enabled only when every target advertises it). Set `imageInput: "disabled"` to force text-only
+even when every target supports images — the catalog drops `image` from `inputModalities`, and
+image-bearing requests are rejected with HTTP 400 before any target is called. `"auto"` (or
+omitting the field) keeps the automatic intersection.
+
 ## Encrypted v2 sub-agent tasks
 
 There is one important limitation for Codex v2 sub-agents ([issue #92](https://github.com/lidge-jun/opencodex/issues/92)).
@@ -256,6 +264,7 @@ Combos are stored in the top-level `combos` object, keyed by combo id:
 | `strategy` | No | `"failover"` | `"failover"` or `"round-robin"`. |
 | `stickyLimit` | No | `1` | Integer from 1 to 100 successful requests per round-robin selection. |
 | `defaultEffort` | No | `null` | `low`, `medium`, `high`, `xhigh`, `max`, or `ultra`; applied only when the caller omits effort and the target advertises support. |
+| `imageInput` | No | `"auto"` | `"auto"` or `"disabled"`. `"disabled"` drops image from published modalities and rejects image-bearing requests before dispatch. Cannot enable image when a target lacks it. |
 | `alias` | No | none | Optional trimmed public model id; use the alias rules above. An empty value is stored as no alias. |
 
 ## Troubleshooting

docs-site/src/content/docs/reference/configuration/routing.md

@@ -63,6 +63,7 @@ namespace, and cannot use reserved bare native families such as `gpt-*`, `o1-*`,
 | `strategy?` | `"failover" \| "round-robin"` | `"failover"` | Selection strategy. Target order is failover priority; weights shape smooth weighted round-robin. |
 | `stickyLimit?` | `number` | `1` | Successful requests retained in one round-robin batch. Range 1–100. |
 | `defaultEffort?` | `"low" \| "medium" \| "high" \| "xhigh" \| "max" \| "ultra" \| null` | unset | Applied only when the caller omits effort and the selected target advertises the requested rung. |
+| `imageInput?` | `"auto" \| "disabled"` | `"auto"` | `"disabled"` drops image from the published modalities and rejects image-bearing requests before dispatch. Cannot enable image when a target lacks it. |
 | `alias?` | `string` | — | Optional public model id in place of the canonical picker slug. |
 
 ```json

docs-site/src/content/docs/zh-cn/guides/combos.md

@@ -131,6 +131,10 @@ combo 失败分为 **跳转** 失败和 **终止** 失败。
 
 当目标能力未知，或者不包含配置的 effort 时，opencodex 会省略默认值，并保持目标自身行为不变。支持的值是 `low`、`medium`、`high`、`xhigh`、`max` 和 `ultra`；省略该字段或将其设为 `null`，就会把 effort 完全交给调用方和目标。
 
+## 图片 / 多模态能力
+
+默认情况下，combo 会发布其目标 **input modalities 的交集**（只有当每个目标都声明支持图片时，图片才会启用）。设置 `imageInput: "disabled"` 可在目标均支持图片时仍强制仅文本——目录会从 `inputModalities` 中去掉 `image`，带图请求会在分发前以 HTTP 400 拒绝。`"auto"`（或省略该字段）保持自动交集。
+
 ## 加密的 v2 子代理任务
 
 对于 Codex v2 子代理，有一个重要限制（[issue #92](https://github.com/lidge-jun/opencodex/issues/92)）。原生父进程只能把新启动 worker 的任务，以为原生 ChatGPT 后端生成的密文形式发送出去。外部 provider 无法读取那段负载。
@@ -210,6 +214,7 @@ combo 会存储在顶层的 `combos` 对象中，并以 combo id 作为键：
 | `strategy` | 否 | `"failover"` | `"failover"` 或 `"round-robin"`。 |
 | `stickyLimit` | 否 | `1` | 每次轮询选择可连续处理的成功请求数，范围为 1 到 100。 |
 | `defaultEffort` | 否 | `null` | `low`、`medium`、`high`、`xhigh`、`max` 或 `ultra`；仅当调用方省略 effort 且目标声明支持时才会应用。 |
+| `imageInput` | 否 | `"auto"` | `"auto"` 或 `"disabled"`。`"disabled"` 会从对外能力中去掉图片，并在分发前拒绝带图请求；不能在目标不支持时强开图片。 |
 | `alias` | 否 | 无 | 可选的、已修剪的公开模型 id；使用上面的别名规则。空值会以“无别名”形式存储。 |
 
 ## 故障排查

docs-site/src/content/docs/zh-cn/reference/configuration/routing.md

@@ -57,6 +57,7 @@ selector 校验、冲突规则和隐私说明见[提供方配置](/reference/con
 | `strategy?` | `"failover" \| "round-robin"` | `"failover"` | 选择策略。目标顺序表示故障切换优先级；权重会影响平滑加权轮询。 |
 | `stickyLimit?` | `number` | `1` | 在单个轮询批次中保留的成功请求数。范围 1–100。 |
 | `defaultEffort?` | `"low" \| "medium" \| "high" \| "xhigh" \| "max" \| "ultra" \| null` | unset | 仅在调用方省略 effort 且所选目标声明了请求的档位时应用。 |
+| `imageInput?` | `"auto" \| "disabled"` | `"auto"` | `"disabled"` 会从对外能力中去掉图片，并在分发前拒绝带图请求；不能在目标不支持时强开图片。 |
 | `alias?` | `string` | — | 可选的公开 model id，用于替代规范化的选择器 slug。 |
 
 ```json

gui/src/combo-capabilities.ts

@@ -0,0 +1,14 @@
+import type { ComboTarget } from "./combo-workspace-data";
+import type { ModelOption } from "./components/combo-workspace-types";
+
+/** Whether every selected target advertises image input (incomplete rows fail closed). */
+export function comboImagesSupported(targets: ComboTarget[], models: ModelOption[]): boolean {
+  if (targets.length === 0) return false;
+  return targets.every((target) => {
+    const provider = target.provider.trim();
+    const modelId = target.model.trim();
+    if (!provider || !modelId) return false;
+    const model = models.find((row) => row.provider === provider && row.id === modelId);
+    return !!model?.inputModalities?.includes("image");
+  });
+}

gui/src/combo-workspace-data.ts

@@ -55,6 +55,11 @@ export function newComboTarget(partial: Partial<ComboTarget> = {}): ComboTarget
   };
 }
 
+
+function normalizeImageInput(value: unknown): "auto" | "disabled" {
+  return value === "disabled" ? "disabled" : "auto";
+}
+
 export interface ComboItem {
   id: string;
   /** Wire id shown to clients, e.g. combo/free */
@@ -64,6 +69,7 @@ export interface ComboItem {
   strategy: ComboStrategy;
   stickyLimit: number;
   defaultEffort: ComboEffort | null;
+  imageInput?: "auto" | "disabled";
   targets: ComboTarget[];
 }
 
@@ -153,6 +159,7 @@ export function parseComboList(payload: unknown): ComboItem[] {
       strategy: normalizeStrategy(r.strategy),
       stickyLimit: normalizeStickyLimit(r.stickyLimit),
       defaultEffort: normalizeDefaultEffort(r.defaultEffort),
+      imageInput: normalizeImageInput(r.imageInput),
       targets,
     });
   }
@@ -210,6 +217,7 @@ export function draftEquals(a: ComboItem, b: ComboItem): boolean {
     || a.strategy !== b.strategy
     || a.stickyLimit !== b.stickyLimit
     || a.defaultEffort !== b.defaultEffort
+    || (a.imageInput ?? "auto") !== (b.imageInput ?? "auto")
   ) return false;
   if (a.targets.length !== b.targets.length) return false;
   return a.targets.every((t, i) => {
@@ -226,6 +234,7 @@ export function toPutBody(item: ComboItem, options: { renameFrom?: string } = {}
     strategy: ComboStrategy;
     stickyLimit?: number;
     defaultEffort: ComboEffort | null;
+    imageInput?: "disabled";
     alias?: string;
   };
 } {
@@ -238,6 +247,7 @@ export function toPutBody(item: ComboItem, options: { renameFrom?: string } = {}
         : { provider: target.provider.trim(), model: target.model.trim() }),
       strategy: item.strategy,
       defaultEffort: item.defaultEffort,
+      ...(item.imageInput === "disabled" ? { imageInput: "disabled" as const } : {}),
       ...(item.strategy === "round-robin" ? { stickyLimit: item.stickyLimit } : {}),
       ...(item.alias && item.alias.trim() ? { alias: item.alias.trim() } : {}),
     },
@@ -326,6 +336,7 @@ export function emptyDraft(id = ""): ComboItem {
     strategy: "failover",
     stickyLimit: 1,
     defaultEffort: null,
+    imageInput: "auto",
     targets: [newComboTarget()],
   };
 }

gui/src/components/combo-workspace-add-modal.tsx

@@ -10,7 +10,7 @@ import { IconX } from "../icons";
 import { useT } from "../i18n/shared";
 import { Notice } from "../ui";
 import type { ModelOption, ProviderOption } from "./combo-workspace-types";
-import { EffortSelect, StrategySeg, TargetEditor } from "./combo-workspace-controls";
+import { ComboCapabilities, EffortSelect, StrategySeg, TargetEditor } from "./combo-workspace-controls";
 import { clampedNumberInput } from "./combo-workspace-utils";
 
 export function AddComboModal({
@@ -204,6 +204,13 @@ export function AddComboModal({
               {draft.strategy === "failover" ? t("cws.targets.failoverHint") : t("cws.targets.roundRobinHint")}
             </p>
           </div>
+          <ComboCapabilities
+            targets={draft.targets}
+            models={models}
+            imageInput={draft.imageInput ?? "auto"}
+            disabled={busy}
+            onChange={(patch) => setDraft((d) => ({ ...d, ...patch }))}
+          />
         </div>
         <div className="cwi-modal-actions">
           <button type="button" className="btn btn-ghost" onClick={requestClose} disabled={busy}>{t("common.cancel")}</button>

gui/src/components/combo-workspace-controls.tsx

@@ -1,8 +1,10 @@
 import { useState } from "react";
 import type { ComboEffort, ComboStrategy, ComboTarget } from "../combo-workspace-data";
+import { comboImagesSupported } from "../combo-capabilities";
 import { COMBO_EFFORTS, newComboTarget } from "../combo-workspace-data";
 import { IconArrowDown, IconArrowUp, IconGrip, IconPlus, IconTrash } from "../icons";
 import { useT } from "../i18n/shared";
+import { Switch } from "../ui";
 import { formatProviderDisplayName } from "../provider-icons";
 import type { ModelOption, ProviderOption } from "./combo-workspace-types";
 import { clampedNumberInput, enabledProviders, modelsForProvider } from "./combo-workspace-utils";
@@ -83,6 +85,49 @@ export function EffortSelect({
   );
 }
 
+
+export function ComboCapabilities({
+  targets,
+  models,
+  imageInput,
+  disabled,
+  onChange,
+}: {
+  targets: ComboTarget[];
+  models: ModelOption[];
+  imageInput: "auto" | "disabled";
+  disabled?: boolean;
+  onChange: (patch: { imageInput?: "auto" | "disabled" }) => void;
+}) {
+  const t = useT();
+  const imagesSupported = comboImagesSupported(targets, models);
+  // Default: checked (auto) when supported; force off when any target lacks image.
+  const effectiveOn = imagesSupported && imageInput !== "disabled";
+
+  return (
+    <section className="cwi-capabilities" aria-label={t("cws.capabilities")}>
+      <span className="field-label">{t("cws.capabilities")}</span>
+      <div className="cwi-capability-row">
+        <div>
+          <span className="cwi-capability-label">{t("cws.capability.imageInput")}</span>
+          <p className="muted cwi-capability-hint">
+            {imagesSupported ? t("cws.capability.imageInputHint") : t("cws.capability.imageInputUnavailable")}
+          </p>
+        </div>
+        <Switch
+          on={effectiveOn}
+          onClick={() => {
+            if (!imagesSupported) return;
+            onChange({ imageInput: imageInput === "auto" ? "disabled" : "auto" });
+          }}
+          disabled={disabled || !imagesSupported}
+          label={t("cws.capability.imageInput")}
+        />
+      </div>
+    </section>
+  );
+}
+
 export function TargetEditor({
   targets,
   strategy,

gui/src/components/combo-workspace-detail-panel.tsx

@@ -11,7 +11,7 @@ import { IconChevron, IconTrash } from "../icons";
 import { useT } from "../i18n/shared";
 import { Notice } from "../ui";
 import type { ModelOption, ProviderOption } from "./combo-workspace-types";
-import { EffortSelect, StrategySeg, TargetEditor } from "./combo-workspace-controls";
+import { ComboCapabilities, EffortSelect, StrategySeg, TargetEditor } from "./combo-workspace-controls";
 import { clampedNumberInput } from "./combo-workspace-utils";
 
 type DetailTab = "config" | "about";
@@ -52,7 +52,7 @@ export function DetailPanel({
   const [msg, setMsg] = useState<{ ok: boolean; text: string } | null>(null);
   const [copied, setCopied] = useState(false);
   const dirty = !draftEquals(draft, baseline);
-  const baselineSyncKey = `${baseline.id}:${baseline.alias ?? ""}:${baseline.strategy}:${baseline.stickyLimit}:${baseline.defaultEffort}:${baseline.targets.map((t) => `${t.provider}/${t.model}:${t.weight ?? 1}`).join(",")}`;
+  const baselineSyncKey = `${baseline.id}:${baseline.alias ?? ""}:${baseline.strategy}:${baseline.stickyLimit}:${baseline.defaultEffort}:${baseline.imageInput ?? "auto"}:${baseline.targets.map((t) => `${t.provider}/${t.model}:${t.weight ?? 1}`).join(",")}`;
   const effortMap = useMemo(() => {
     const map = new Map<string, string[] | undefined>();
     for (const model of models) {
@@ -267,6 +267,13 @@ export function DetailPanel({
                 {draft.strategy === "failover" ? t("cws.targets.failoverHint") : t("cws.targets.roundRobinHint")}
               </p>
             </div>
+            <ComboCapabilities
+              targets={draft.targets}
+              models={models}
+              imageInput={draft.imageInput ?? "auto"}
+              disabled={busy}
+              onChange={(patch) => updateDraft((d) => ({ ...d, ...patch }))}
+            />
           </div>
         ) : (
           <section className="pwi-section">

gui/src/components/combo-workspace-types.ts

@@ -13,6 +13,7 @@ export type ModelOption = {
   id: string;
   namespaced?: string;
   reasoningEfforts?: string[];
+  inputModalities?: string[];
 };
 
 export interface ComboWorkspaceProps {

gui/src/i18n/de.ts

@@ -1734,6 +1734,10 @@ export const de: Record<TKey, string> = {
   "cws.field.defaultEffort": "Standard-Reasoning",
   "cws.field.defaultEffortNone": "Keine (Ziel-Standard)",
   "cws.field.defaultEffortHint": "Nur verwendet, wenn der Client keinen Reasoning-Aufwand sendet. Optionen sind die Schnittmenge der beworbenen Aufwände der gewählten Ziele.",
+  "cws.capability.imageInputUnavailable": "Erst verfügbar, wenn jedes gewählte Ziel Bildeingabe unterstützt.",
+  "cws.capability.imageInputHint": "Standardmäßig aktiv, wenn jedes Ziel Bilder unterstützt. Ausschalten für nur Text.",
+  "cws.capability.imageInput": "Bild / multimodal",
+  "cws.capabilities": "Fähigkeiten",
   "cws.field.defaultEffortUnsupported": "Dieser Aufwand liegt nicht in der gemeinsamen Leiter der Ziele — er wird zur Anfragezeit ignoriert oder angepasst.",
   "cws.field.defaultEffortUnsupportedOption": "nicht in der Schnittmenge",
   "cws.targets": "Ziele",

gui/src/i18n/en.ts

@@ -1768,6 +1768,10 @@ export const en = {
   "cws.field.defaultEffort": "Default reasoning",
   "cws.field.defaultEffortNone": "None (target default)",
   "cws.field.defaultEffortHint": "Used only when the client omits reasoning effort. Options are the intersection of the selected targets' advertised efforts; targets without catalog effort metadata offer none.",
+  "cws.capability.imageInputUnavailable": "Unavailable until every selected target supports image input.",
+  "cws.capability.imageInputHint": "On by default when every target supports images. Turn off to accept text only.",
+  "cws.capability.imageInput": "Image / multimodal",
+  "cws.capabilities": "Capabilities",
   "cws.field.defaultEffortUnsupported": "This effort is not in the targets' common ladder — it will be ignored or snapped at request time.",
   "cws.field.defaultEffortUnsupportedOption": "not in intersection",
   "cws.targets": "Targets",

gui/src/i18n/ja.ts

@@ -1802,6 +1802,10 @@ export const ja: Record<TKey, string> = {
   "cws.field.defaultEffort": "デフォルトの推論",
   "cws.field.defaultEffortNone": "なし(ターゲットのデフォルト)",
   "cws.field.defaultEffortHint": "クライアントが推論負荷を省略した場合のみ使用されます。選択肢は選択ターゲットが広告する負荷の交差です。",
+  "cws.capability.imageInputUnavailable": "選択した全ターゲットが画像入力に対応すると有効になります。",
+  "cws.capability.imageInputHint": "全ターゲットが画像対応なら既定でオン。オフにするとテキストのみ。",
+  "cws.capability.imageInput": "画像 / マルチモーダル",
+  "cws.capabilities": "能力",
   "cws.field.defaultEffortUnsupported": "この負荷はターゲット共通の階段にありません — リクエスト時に無視またはスナップされます。",
   "cws.field.defaultEffortUnsupportedOption": "交差に含まれない",
   "cws.targets": "ターゲット",

gui/src/i18n/ko.ts

@@ -1761,6 +1761,10 @@ export const ko: Record<TKey, string> = {
   "cws.field.defaultEffort": "기본 추론 수준",
   "cws.field.defaultEffortNone": "없음 (대상 기본값)",
   "cws.field.defaultEffortHint": "클라이언트가 추론 수준을 생략한 경우에만 사용합니다. 옵션은 선택한 대상이 광고하는 수준의 교집합입니다.",
+  "cws.capability.imageInputUnavailable": "선택한 모든 대상이 이미지 입력을 지원해야 사용할 수 있습니다.",
+  "cws.capability.imageInputHint": "모든 대상이 이미지를 지원하면 기본으로 켜집니다. 끄면 텍스트만 허용합니다.",
+  "cws.capability.imageInput": "이미지 / 멀티모달",
+  "cws.capabilities": "기능",
   "cws.field.defaultEffortUnsupported": "이 수준은 대상의 공통 사다리에 없습니다 — 요청 시 무시되거나 스냅됩니다.",
   "cws.field.defaultEffortUnsupportedOption": "교집합에 없음",
   "cws.targets": "대상",

gui/src/i18n/ru.ts

@@ -1844,6 +1844,10 @@ export const ru: Record<TKey, string> = {
   "cws.field.defaultEffort": "Рассуждения по умолчанию",
   "cws.field.defaultEffortNone": "Нет (по умолчанию для цели)",
   "cws.field.defaultEffortHint": "Используется, только если клиент не указал уровень рассуждений. Варианты — пересечение заявленных уровней выбранных целей.",
+  "cws.capability.imageInputUnavailable": "Доступно, когда все выбранные цели поддерживают ввод изображений.",
+  "cws.capability.imageInputHint": "Включено по умолчанию, если все цели поддерживают изображения. Выключите, чтобы принимать только текст.",
+  "cws.capability.imageInput": "Изображения / мультимодальность",
+  "cws.capabilities": "Возможности",
   "cws.field.defaultEffortUnsupported": "Этот уровень не входит в общую лестницу целей — при запросе он будет проигнорирован или снижен.",
   "cws.field.defaultEffortUnsupportedOption": "нет в пересечении",
   "cws.targets": "Цели",

gui/src/i18n/zh.ts

@@ -1754,6 +1754,10 @@ export const zh: Record<TKey, string> = {
   "cws.field.defaultEffort": "默认推理级别",
   "cws.field.defaultEffortNone": "无（使用目标默认）",
   "cws.field.defaultEffortHint": "仅在客户端未指定推理级别时使用。选项为所选目标已公布努力级别的交集。",
+  "cws.capability.imageInputUnavailable": "所有已选目标均支持图片输入后才可用。",
+  "cws.capability.imageInputHint": "所有目标均支持图片时默认开启；关闭后仅接受文本。",
+  "cws.capability.imageInput": "图片 / 多模态",
+  "cws.capabilities": "能力",
   "cws.field.defaultEffortUnsupported": "该级别不在目标的公共阶梯中 — 请求时会被忽略或就近映射。",
   "cws.field.defaultEffortUnsupportedOption": "不在交集中",
   "cws.targets": "目标",