fix: remove host model warnings from profile selectors (#3521)

apps/web/components/task-create-dialog-options.test.tsx

@@ -58,19 +58,6 @@ const AGENT_WITH_GPT: AvailableAgent = {
   },
 } as unknown as AvailableAgent;
 
-const AGENT_WITH_OPUS_VARIATION: AvailableAgent = {
-  name: "omp-acp",
-  available: true,
-  model_config: {
-    default_model: "opus[1m]",
-    available_models: [{ id: "opus[1m]", name: "Opus (1m)" }],
-    current_model_id: "opus[1m]",
-    available_modes: [],
-    supports_dynamic_models: false,
-    status: "ok",
-  },
-} as unknown as AvailableAgent;
-
 const GONE_MODEL = "claude-gone";
 const DATA_DISABLED = "data-disabled";
 const MODEL_PROBE_WARNING_TEST_ID = "agent-profile-model-probe-warning";
@@ -118,14 +105,6 @@ function renderOptions(profiles: AgentProfileOption[]) {
   return screen.getByTestId("option-0");
 }
 
-function getModelProbeWarning() {
-  return screen.getByTestId(MODEL_PROBE_WARNING_TEST_ID);
-}
-
-function getModelProbeWarningLabel() {
-  return getModelProbeWarning().getAttribute("aria-label");
-}
-
 beforeEach(() => {
   vi.clearAllMocks();
   setAvailableAgents([AGENT_WITH_GPT]);
@@ -226,80 +205,122 @@ describe("useAgentProfileOptions recent-use ordering", () => {
   });
 });
 
-describe("useAgentProfileOptions executor-authoritative model hint", () => {
+describe("useAgentProfileOptions model-independent labels", () => {
-  it("keeps a profile whose start model is absent from the host probe selectable", () => {
+  // @covers AC-AGENTS-NO-SILENT-MODEL-FALLBACK-003.1
-    const option = renderOptions([profileOption({ model: GONE_MODEL })]);
+  it.each([
-    expect(option.getAttribute(DATA_DISABLED)).toBeNull();
+    ["exact", "gpt-5", ["gpt-5"]],
-    expect(getModelProbeWarningLabel()).toContain(GONE_MODEL);
+    ["missing", GONE_MODEL, ["gpt-5"]],
-  });
+    ["unique variation", "opus", ["opus[1m]"]],
-
+    ["multiple variations", "opus", ["opus[1m]", "opus[270k]"]],
-  it("keeps a profile with an available start model selectable", () => {
+    ["legacy effort IDs", "gpt-6-astra", ["gpt-6-astra[low]", "gpt-6-astra[high]"]],
-    const option = renderOptions([profileOption({ model: "gpt-5" })]);
+    ["bracketed request", "opus[1m]", ["opus[270k]"]],
-    expect(option.getAttribute(DATA_DISABLED)).toBeNull();
+    ["empty catalog", GONE_MODEL, []],
-  });
+    ["provider default", "", ["gpt-5"]],
-
+  ])("does not show host model advisories in either label: %s", (_, model, models) => {
-  it("names a unique advertised variation without disabling the profile", () => {
+    setAvailableAgents([
-    setAvailableAgents([AGENT_WITH_OPUS_VARIATION]);
+      {
-    const option = renderOptions([profileOption({ model: "opus" })]);
+        ...AGENT_WITH_GPT,
-
+        model_config: {
-    expect(option.getAttribute(DATA_DISABLED)).toBeNull();
+          ...AGENT_WITH_GPT.model_config,
-    expect(getModelProbeWarningLabel()).toContain("opus[1m]");
+          available_models: (models as string[]).map((id) => ({ id, name: id })),
-  });
+        },
-
+      },
-  it("keeps a profile with an empty (agent default) model selectable", () => {
-    const option = renderOptions([profileOption({ model: "" })]);
-    expect(option.getAttribute(DATA_DISABLED)).toBeNull();
-  });
-
-  it("keeps a gone-model profile with a fallback selectable", () => {
-    const option = renderOptions([profileOption({ model: GONE_MODEL, fallback_model: "gpt-5" })]);
-    expect(option.getAttribute(DATA_DISABLED)).toBeNull();
-    expect(option.getAttribute("data-reason")).toBeNull();
-
-    expect(getModelProbeWarningLabel()).toContain(GONE_MODEL);
-  });
-
-  it("keeps a profile selectable when both saved models are absent from the host probe", () => {
-    const option = renderOptions([
-      profileOption({ model: GONE_MODEL, fallback_model: "other-gone" }),
     ]);
-    expect(option.getAttribute(DATA_DISABLED)).toBeNull();
+    const profile = profileOption({ model: model as string });
-    expect(getModelProbeWarningLabel()).toContain(GONE_MODEL);
+    const { result } = renderHook(() => useAgentProfileOptions([profile]));
-  });
+    const option = result.current[0]!;
-
+    expect(option.disabled).toBeUndefined();
-  it("keeps auto-fallback profiles selectable and shows the host probe hint", () => {
+    expect(option.disabledReason).toBeUndefined();
-    const option = renderOptions([profileOption({ model: GONE_MODEL, auto_fallback: true })]);
-    expect(option.getAttribute(DATA_DISABLED)).toBeNull();
-    expect(getModelProbeWarningLabel()).toContain(GONE_MODEL);
-  });
-
-  it("renders the host probe hint as one compact warning trigger", () => {
-    const option = renderOptions([profileOption({ model: GONE_MODEL })]);
-
-    expect(getModelProbeWarning()).toBeTruthy();
-    expect(option.textContent).not.toContain(
-      "The host probe did not advertise claude-gone. The selected executor will decide the model at launch.",
-    );
-  });
-
-  it("uses a non-interactive warning indicator in the selected trigger", () => {
-    const { result } = renderHook(() =>
-      useAgentProfileOptions([profileOption({ model: GONE_MODEL })]),
-    );
-
     render(
       <TooltipProvider>
-        <div>{result.current[0]?.renderTriggerLabel?.()}</div>
+        <div data-testid="option-label">{option.renderLabel()}</div>
+        <div data-testid="selected-label">{option.renderTriggerLabel?.()}</div>
       </TooltipProvider>,
     );
-
+    for (const label of ["option-label", "selected-label"]) {
-    expect(screen.queryByTestId(MODEL_PROBE_WARNING_TEST_ID)).toBeNull();
+      const element = screen.getByTestId(label);
-    expect(screen.getByTitle(/claude-gone/)).toBeTruthy();
+      expect(element.textContent).toContain("hybrid");
+      expect(element.querySelector("button, .tabler-icon-alert-triangle")).toBeNull();
+      expect(element.querySelector('[title*="host probe"]')).toBeNull();
+    }
   });
 
-  it("does not gate when the agent model list is unknown (probe not landed)", () => {
+  // @covers AC-AGENTS-NO-SILENT-MODEL-FALLBACK-003.3
+  it.each(["auth_required", "not_installed", "failed"] as const)(
+    "preserves %s health indicators when the saved model is absent",
+    (capability_status) => {
+      setAvailableAgents([]);
+      const profile = profileOption({
+        model: GONE_MODEL,
+        capability_status,
+        capability_error: "Agent needs attention",
+      });
+      const { result } = renderHook(() => useAgentProfileOptions([profile]));
+      const option = result.current[0]!;
+      render(
+        <TooltipProvider>
+          <div>{option.renderLabel()}</div>
+          <div>{option.renderTriggerLabel?.()}</div>
+        </TooltipProvider>,
+      );
+      expect(screen.getAllByTitle("Agent needs attention")).toHaveLength(2);
+      expect(screen.queryByTestId(MODEL_PROBE_WARNING_TEST_ID)).toBeNull();
+    },
+  );
+
+  // @covers AC-AGENTS-NO-SILENT-MODEL-FALLBACK-003.6
+  it.each([false, true])(
+    "preserves saved profile settings with auto fallback %s",
+    (auto_fallback) => {
+      const profile = profileOption({
+        model: GONE_MODEL,
+        fallback_model: "other-gone",
+        auto_fallback,
+        cli_passthrough: true,
+      });
+      const saved = structuredClone(profile);
+      const option = renderOptions([profile]);
+      expect(option.getAttribute(DATA_DISABLED)).toBeNull();
+      expect(option.getAttribute("data-reason")).toBeNull();
+      expect(screen.queryByTestId(MODEL_PROBE_WARNING_TEST_ID)).toBeNull();
+      expect(option.querySelector(".tabler-icon-terminal-2")).not.toBeNull();
+      expect(profile).toEqual(saved);
+    },
+  );
+
+  it("keeps labels stable when a pending host catalog changes", () => {
     setAvailableAgents([]);
-    const option = renderOptions([profileOption({ model: GONE_MODEL })]);
+    const profile = profileOption({ model: GONE_MODEL });
-    expect(option.getAttribute(DATA_DISABLED)).toBeNull();
+    const { rerender } = render(<OptionsProbe profiles={[profile]} />);
+    const initialLabel = screen.getByTestId("option-0").innerHTML;
+    setAvailableAgents([AGENT_WITH_GPT]);
+    rerender(<OptionsProbe profiles={[profile]} />);
+    expect(screen.getByTestId("option-0").innerHTML).toBe(initialLabel);
+    expect(screen.queryByTestId(MODEL_PROBE_WARNING_TEST_ID)).toBeNull();
   });
 });
+
+it("refreshes capability health from the host snapshot without inspecting model IDs", () => {
+  setAvailableAgents([
+    {
+      ...AGENT_WITH_GPT,
+      available: false,
+      model_config: {
+        ...AGENT_WITH_GPT.model_config,
+        error: "Agent unavailable",
+      },
+    },
+  ]);
+  const { result } = renderHook(() =>
+    useAgentProfileOptions([profileOption({ model: GONE_MODEL })]),
+  );
+  const option = result.current[0]!;
+  render(
+    <TooltipProvider>
+      <div>{option.renderLabel()}</div>
+      <div>{option.renderTriggerLabel?.()}</div>
+    </TooltipProvider>,
+  );
+  expect(screen.getAllByTitle("Agent unavailable")).toHaveLength(2);
+  expect(screen.queryByTestId(MODEL_PROBE_WARNING_TEST_ID)).toBeNull();
+});

apps/web/components/task-create-dialog-options.tsx

@@ -1,21 +1,12 @@
 "use client";
 
-import { useEffect, useMemo, useState } from "react";
+import { useEffect, useMemo } from "react";
 import { useTranslation } from "react-i18next";
 import { useAppStore, useAppStoreApi } from "@/components/state-provider";
 import { t } from "@/lib/i18n";
-import { IconAlertTriangle, IconGitBranch, IconTerminal2 } from "@tabler/icons-react";
+import { IconGitBranch, IconTerminal2 } from "@tabler/icons-react";
 import { Badge } from "@kandev/ui/badge";
-import {
-  Drawer,
-  DrawerContent,
-  DrawerDescription,
-  DrawerHeader,
-  DrawerTitle,
-  DrawerTrigger,
-} from "@kandev/ui/drawer";
 import { ScrollOnOverflow } from "@kandev/ui/scroll-on-overflow";
-import { Tooltip, TooltipContent, TooltipTrigger } from "@kandev/ui/tooltip";
 import type {
   LocalRepository,
   Repository,
@@ -23,17 +14,17 @@ import type {
   Executor,
   ExecutorProfile,
 } from "@/lib/types/http";
-import type { AvailableAgent } from "@/lib/types/http-agents";
 import type { AgentProfileOption } from "@/lib/state/slices";
 import { useAvailableAgents } from "@/hooks/domains/settings/use-available-agents";
 import { useFeature } from "@/hooks/domains/features/use-feature";
-import { isSelectableAgentProfile } from "@/lib/state/slices/settings/types";
+import {
+  isSelectableAgentProfile,
+  refreshProfileCapabilities,
+} from "@/lib/state/slices/settings/types";
 import { formatUserHomePath, truncateRepoPath } from "@/lib/utils";
 import { getExecutorIcon } from "@/lib/executor-icons";
 import { AgentLogo } from "@/components/agent-logo";
 import { getCapabilityWarning } from "@/lib/capability-warning";
-import { findUniqueModelVariation } from "@/lib/model-variation";
-import { useTouchDrawer } from "@/hooks/use-compact-task-chrome";
 import { buildBranchKeywords } from "./branch-picker-options";
 import {
   ensureAgentProfileRecentUseLoaded,
@@ -50,50 +41,6 @@ type OptionItem = {
   disabledReason?: string;
 };
 
-function ModelProbeWarning({ note }: { note: string }) {
-  const usesTouchDrawer = useTouchDrawer();
-  const [drawerOpen, setDrawerOpen] = useState(false);
-  const trigger = (
-    <button
-      type="button"
-      className="inline-flex min-h-11 min-w-8 shrink-0 cursor-help items-center justify-center rounded-sm border-0 bg-transparent p-0 focus-visible:outline-none focus-visible:ring-2 focus-visible:ring-ring"
-      aria-label={note}
-      aria-expanded={usesTouchDrawer ? drawerOpen : undefined}
-      aria-haspopup={usesTouchDrawer ? "dialog" : undefined}
-      data-testid="agent-profile-model-probe-warning"
-      onPointerDown={(event) => event.stopPropagation()}
-      onClick={(event) => event.stopPropagation()}
-    >
-      <IconAlertTriangle className="size-3.5 text-amber-500" aria-hidden />
-    </button>
-  );
-
-  if (usesTouchDrawer) {
-    return (
-      <Drawer open={drawerOpen} onOpenChange={setDrawerOpen}>
-        <DrawerTrigger asChild>{trigger}</DrawerTrigger>
-        <DrawerContent>
-          <DrawerHeader>
-            <DrawerTitle className="sr-only">{note}</DrawerTitle>
-            <DrawerDescription>{note}</DrawerDescription>
-          </DrawerHeader>
-        </DrawerContent>
-      </Drawer>
-    );
-  }
-
-  return (
-    <Tooltip>
-      <TooltipTrigger asChild>{trigger}</TooltipTrigger>
-      <TooltipContent side="top">{note}</TooltipContent>
-    </Tooltip>
-  );
-}
-
-function ModelProbeWarningIndicator({ note }: { note: string }) {
-  return <IconAlertTriangle className="size-3.5 text-amber-500" title={note} aria-hidden />;
-}
-
 export function useRepositoryOptions(
   repositories: Repository[],
   discoveredRepositories: LocalRepository[],
@@ -185,19 +132,15 @@ export function useBranchOptions(branchOptionsRaw: Branch[]) {
   }, [branchOptionsRaw]);
 }
 
-// advertisedModelIDs returns the currently advertised model IDs for an agent
-// from the host-utility probe cache (empty when the probe has not landed).
-function advertisedModelIDs(availableAgents: AvailableAgent[], agentName: string): string[] {
-  const agent = availableAgents.find((a) => a.name === agentName);
-  return agent?.model_config?.available_models?.map((m) => m.id) ?? [];
-}
-
 export function useAgentProfileOptions(
   agentProfiles: AgentProfileOption[],
   context?: AgentProfileRecentUseContext,
 ): OptionItem[] {
   const { t } = useTranslation();
-  const { items: availableAgents } = useAvailableAgents();
+  // Keep capability discovery alive for every selector surface. The host
+  // catalog supplies health status only; it never participates in model-ID
+  // matching or selector eligibility.
+  const availableAgents = useAvailableAgents();
   const dynamicRoutingEnabled = useFeature("dynamicAgentRouting");
   const storeApi = useAppStoreApi();
   const recentUseLoaded = useAppStore((state) => !context || state.agentProfileRecentUse.loaded);
@@ -209,9 +152,13 @@ export function useAgentProfileOptions(
     void ensureAgentProfileRecentUseLoaded(storeApi);
   }, [context, recentUseLoaded, storeApi]);
   return useMemo(() => {
+    const profilesWithCapabilities = refreshProfileCapabilities(
+      agentProfiles,
+      availableAgents.items,
+    );
     // Disabled profiles stay in the store (existing sessions keep their
     // labels) but are never offered as a choice for new work.
-    const selectable = agentProfiles.filter((profile) =>
+    const selectable = profilesWithCapabilities.filter((profile) =>
       isSelectableAgentProfile(profile, dynamicRoutingEnabled),
     );
     const orderedProfiles = context
@@ -223,26 +170,7 @@ export function useAgentProfileOptions(
       const profileLabel = parts[1] ?? "";
       const isPassthrough = profile.cli_passthrough === true;
       const warning = getCapabilityWarning(profile.capability_status, profile.capability_error);
-      // The host-utility probe is an editing hint only. The selected
+      const renderProfileLabel = () => (
-      // executor owns the launch-time model catalog, so a host-only mismatch
-      // must never remove a profile from the task selector.
-      const advertised = advertisedModelIDs(availableAgents, profile.agent_name);
-      const startModelGone = Boolean(
-        profile.model && advertised.length > 0 && !advertised.includes(profile.model),
-      );
-      const uniqueVariation = startModelGone
-        ? findUniqueModelVariation(profile.model ?? "", advertised)
-        : null;
-      let modelProbeNote: string | undefined;
-      if (startModelGone) {
-        modelProbeNote = uniqueVariation
-          ? t("settings:profileStartModelUniqueVariationOnHost", {
-              model: profile.model,
-              variation: uniqueVariation,
-            })
-          : t("settings:profileStartModelNotAdvertisedOnHost", { model: profile.model });
-      }
-      const renderProfileLabel = (modelProbeWarning: React.ReactNode) => (
         <span className="flex min-w-0 flex-1 flex-col gap-1">
           <span className="flex shrink-0 items-center justify-between gap-2">
             <span className="flex shrink-0 items-center gap-1.5">
@@ -251,7 +179,6 @@ export function useAgentProfileOptions(
               {warning && (
                 <warning.Icon className={`size-3.5 ${warning.color}`} title={warning.title} />
               )}
-              {modelProbeWarning}
             </span>
             <span className="flex shrink-0 items-center gap-1.5">
               {isPassthrough && (
@@ -274,15 +201,11 @@ export function useAgentProfileOptions(
         label: profile.label,
         disabled: undefined,
         disabledReason: undefined,
-        renderLabel: () =>
+        renderLabel: renderProfileLabel,
-          renderProfileLabel(modelProbeNote ? <ModelProbeWarning note={modelProbeNote} /> : null),
+        renderTriggerLabel: renderProfileLabel,
-        renderTriggerLabel: () =>
-          renderProfileLabel(
-            modelProbeNote ? <ModelProbeWarningIndicator note={modelProbeNote} /> : null,
-          ),
       };
     });
-  }, [agentProfiles, availableAgents, context, dynamicRoutingEnabled, recentProfileIds, t]);
+  }, [agentProfiles, availableAgents.items, context, dynamicRoutingEnabled, recentProfileIds, t]);
 }
 
 export function useExecutorOptions(executors: Executor[]): OptionItem[] {

apps/web/e2e/tests/session/model-mismatch-warning-helpers.ts

@@ -43,3 +43,16 @@ export async function readModelSelectionWarnings(
   const { messages } = await apiClient.listSessionMessages(sessionId);
   return messages.filter((message) => message.metadata?.kind === "model_selection_warning");
 }
+
+export async function createExecutorOnlyModelProfile(apiClient: ApiClient): Promise<AgentProfile> {
+  const { agents } = await apiClient.listAgents();
+  const agent = agents.find((item) => item.name === "mock-agent");
+  if (!agent) throw new Error("The E2E fixture must provide a mock agent");
+  return apiClient.createAgentProfile(agent.id, "Opus High", {
+    model: AMBIGUOUS_MODEL_VARIATIONS[0],
+    config_options: { effort: "high" },
+    fallback_model: "mock-smart",
+    auto_fallback: false,
+    env_vars: [{ key: MODEL_CATALOG_ENV, value: "ambiguous" }],
+  });
+}

apps/web/e2e/tests/settings/mobile-no-silent-model-fallback.spec.ts

@@ -2,15 +2,18 @@ import { expect, test } from "../../fixtures/test-base";
 import { KanbanPage } from "../../pages/kanban-page";
 import { assertNoDocumentHorizontalOverflow } from "../../helpers/layout-assertions";
 import {
+  createExecutorOnlyModelProfile,
   createModelVariationProfile,
   createMismatchedProfile,
-  MODEL_VARIATION_BASE,
-  UNIQUE_MODEL_VARIATION,
-  UNADVERTISED_MODEL,
 } from "../session/model-mismatch-warning-helpers";
+import { launchExecutorOnlyModelProfile } from "./profile-model-selection-helpers";
 
 test.describe("executor-authoritative model selection on mobile", () => {
-  test("keeps the host-mismatched profile reachable by touch", async ({ testPage, apiClient }) => {
+  test("keeps the host-mismatched profile reachable by touch", async ({
+    testPage,
+    apiClient,
+    prCapture,
+  }) => {
     const profile = await createMismatchedProfile(apiClient, "Mobile host mismatch profile");
     try {
       const kanban = new KanbanPage(testPage);
@@ -29,23 +32,25 @@ test.describe("executor-authoritative model selection on mobile", () => {
         .getByRole("option", { name: profile.name, exact: false });
       await expect(option).toBeVisible();
       await expect(option).toBeEnabled();
-      const warning = option.getByTestId("agent-profile-model-probe-warning");
+      await expect(option.getByTestId("agent-profile-model-probe-warning")).toHaveCount(0);
-      await expect(warning).toBeVisible();
+      await expect(option.locator(".tabler-icon-alert-triangle")).toHaveCount(0);
-      const warningText = `The host probe did not advertise ${UNADVERTISED_MODEL}. The selected executor will decide the model at launch.`;
+      await prCapture.screenshot("mobile-profile-options", {
-      await expect(dialog).not.toContainText(warningText);
+        caption: "Saved profiles remain selectable without host-model warnings.",
-      await warning.tap();
+      });
-      await expect(
+      await option.tap();
-        testPage
+      await expect(testPage.getByRole("listbox")).not.toBeVisible();
-          .locator('[data-slot="drawer-content"][data-state="open"]')
+      await expect(selector).toContainText(profile.name);
-          .filter({ hasText: warningText }),
+      await prCapture.screenshot("mobile-selected-profile", {
-      ).toBeVisible();
+        caption: "The selected profile keeps its name without a model advisory.",
-      await assertNoDocumentHorizontalOverflow(testPage, "mobile model mismatch selector");
+      });
+      await expect(selector.locator("button, .tabler-icon-alert-triangle")).toHaveCount(0);
+      await assertNoDocumentHorizontalOverflow(testPage, "mobile profile selection");
     } finally {
       await apiClient.deleteAgentProfile(profile.id, true).catch(() => {});
     }
   });
 
-  test("opens the unique variation advisory by touch without overflow", async ({
+  test("selects a unique-variation profile by touch without model help", async ({
     testPage,
     apiClient,
   }) => {
@@ -70,18 +75,28 @@ test.describe("executor-authoritative model selection on mobile", () => {
         .getByRole("option", { name: profile.name, exact: false });
       await expect(option).toBeVisible();
       await expect(option).toBeEnabled();
-      const warning = option.getByTestId("agent-profile-model-probe-warning");
+      await expect(option.getByTestId("agent-profile-model-probe-warning")).toHaveCount(0);
-      await expect(warning).toBeVisible();
+      await expect(option.locator(".tabler-icon-alert-triangle")).toHaveCount(0);
-      const warningText = `The host probe found one possible variation of ${MODEL_VARIATION_BASE}: ${UNIQUE_MODEL_VARIATION}. The selected executor will decide the model at launch.`;
+      await option.tap();
-      await warning.tap();
+      await expect(testPage.getByRole("listbox")).not.toBeVisible();
-      await expect(
+      await expect(selector).toContainText(profile.name);
-        testPage
+      await expect(selector.locator("button, .tabler-icon-alert-triangle")).toHaveCount(0);
-          .locator('[data-slot="drawer-content"][data-state="open"]')
+      await assertNoDocumentHorizontalOverflow(testPage, "mobile profile selection");
-          .filter({ hasText: warningText }),
-      ).toBeVisible();
-      await assertNoDocumentHorizontalOverflow(testPage, "mobile unique variation advisory");
     } finally {
       await apiClient.deleteAgentProfile(profile.id, true).catch(() => {});
     }
   });
+  // @covers AC-AGENTS-NO-SILENT-MODEL-FALLBACK-003.5
+  // @covers AC-AGENTS-NO-SILENT-MODEL-FALLBACK-003.6
+  test("launches a host-mismatched profile on its requested executor model without a warning", async ({
+    testPage,
+    apiClient,
+  }) => {
+    const profile = await createExecutorOnlyModelProfile(apiClient);
+    try {
+      await launchExecutorOnlyModelProfile(testPage, apiClient, profile, true);
+    } finally {
+      await apiClient.deleteAgentProfile(profile.id, true);
+    }
+  });
 });

apps/web/e2e/tests/settings/no-silent-model-fallback.spec.ts

@@ -1,15 +1,14 @@
 import { expect, test } from "../../fixtures/test-base";
 import { KanbanPage } from "../../pages/kanban-page";
 import {
+  createExecutorOnlyModelProfile,
   createModelVariationProfile,
   createMismatchedProfile,
-  MODEL_VARIATION_BASE,
-  UNIQUE_MODEL_VARIATION,
-  UNADVERTISED_MODEL,
 } from "../session/model-mismatch-warning-helpers";
+import { launchExecutorOnlyModelProfile } from "./profile-model-selection-helpers";
 
 test.describe("executor-authoritative model selection", () => {
-  test("keeps a host-mismatched profile selectable", async ({ testPage, apiClient }) => {
+  test("keeps a host-mismatched profile selectable", async ({ testPage, apiClient, prCapture }) => {
     const profile = await createMismatchedProfile(apiClient, "Host mismatch selectable profile");
     try {
       const kanban = new KanbanPage(testPage);
@@ -28,24 +27,27 @@ test.describe("executor-authoritative model selection", () => {
         .getByRole("option", { name: profile.name, exact: false });
       await expect(option).toBeVisible();
       await expect(option).toBeEnabled();
-      const warning = option.getByTestId("agent-profile-model-probe-warning");
+      await expect(option.getByTestId("agent-profile-model-probe-warning")).toHaveCount(0);
-      await expect(warning).toBeVisible();
+      await expect(option.locator(".tabler-icon-alert-triangle")).toHaveCount(0);
-      const warningText = `The host probe did not advertise ${UNADVERTISED_MODEL}. The selected executor will decide the model at launch.`;
+      await prCapture.screenshot("desktop-profile-options", {
-      await expect(dialog).not.toContainText(warningText);
+        caption: "Saved profiles remain selectable without host-model warnings.",
-      await warning.hover();
+      });
-      await expect(
+      const search = testPage.locator("[cmdk-input]");
-        testPage
+      await search.fill(profile.name);
-          .locator('[data-slot="tooltip-content"]:not([data-state="closed"])')
+      await search.press("Enter");
-          .filter({ hasText: warningText }),
+      await expect(testPage.getByRole("listbox")).not.toBeVisible();
-      ).toBeVisible();
+      await expect(selector).toContainText(profile.name);
-      await option.click();
+      await prCapture.screenshot("desktop-selected-profile", {
+        caption: "The selected profile keeps its name without a model advisory.",
+      });
+      await expect(selector.locator(".tabler-icon-alert-triangle")).toHaveCount(0);
       await expect(selector.locator("button")).toHaveCount(0);
     } finally {
       await apiClient.deleteAgentProfile(profile.id, true).catch(() => {});
     }
   });
 
-  test("names one host-advertised variation while keeping the profile selectable", async ({
+  test("selects a unique-variation profile without a host model advisory", async ({
     testPage,
     apiClient,
   }) => {
@@ -70,18 +72,27 @@ test.describe("executor-authoritative model selection", () => {
         .getByRole("option", { name: profile.name, exact: false });
       await expect(option).toBeVisible();
       await expect(option).toBeEnabled();
-      const warning = option.getByTestId("agent-profile-model-probe-warning");
+      await expect(option.getByTestId("agent-profile-model-probe-warning")).toHaveCount(0);
-      await expect(warning).toBeVisible();
+      await expect(option.locator(".tabler-icon-alert-triangle")).toHaveCount(0);
-      const warningText = `The host probe found one possible variation of ${MODEL_VARIATION_BASE}: ${UNIQUE_MODEL_VARIATION}. The selected executor will decide the model at launch.`;
-      await warning.hover();
-      await expect(
-        testPage
-          .locator('[data-slot="tooltip-content"]:not([data-state="closed"])')
-          .filter({ hasText: warningText }),
-      ).toBeVisible();
       await option.click();
+      await expect(testPage.getByRole("listbox")).not.toBeVisible();
+      await expect(selector).toContainText(profile.name);
+      await expect(selector.locator(".tabler-icon-alert-triangle")).toHaveCount(0);
     } finally {
       await apiClient.deleteAgentProfile(profile.id, true).catch(() => {});
     }
   });
+  // @covers AC-AGENTS-NO-SILENT-MODEL-FALLBACK-003.5
+  // @covers AC-AGENTS-NO-SILENT-MODEL-FALLBACK-003.6
+  test("launches a host-mismatched profile on its requested executor model without a warning", async ({
+    testPage,
+    apiClient,
+  }) => {
+    const profile = await createExecutorOnlyModelProfile(apiClient);
+    try {
+      await launchExecutorOnlyModelProfile(testPage, apiClient, profile, false);
+    } finally {
+      await apiClient.deleteAgentProfile(profile.id, true);
+    }
+  });
 });

apps/web/e2e/tests/settings/profile-model-selection-helpers.ts

@@ -0,0 +1,103 @@
+import { expect, type Page } from "@playwright/test";
+import type { AgentProfile } from "../../../lib/types/http-agents";
+import type { ApiClient } from "../../helpers/api-client";
+import { assertNoDocumentHorizontalOverflow } from "../../helpers/layout-assertions";
+import { waitForSessionDone } from "../../helpers/session";
+import { KanbanPage } from "../../pages/kanban-page";
+import { SessionPage } from "../../pages/session-page";
+import { readModelSelectionWarnings } from "../session/model-mismatch-warning-helpers";
+
+export async function launchExecutorOnlyModelProfile(
+  page: Page,
+  apiClient: ApiClient,
+  profile: AgentProfile,
+  mobile: boolean,
+) {
+  await expect
+    .poll(async () => {
+      const { agents } = await apiClient.listAvailableAgents();
+      return agents.find((agent) => agent.name === "mock-agent")?.model_config.status;
+    })
+    .toBe("ok");
+  const { agents } = await apiClient.listAvailableAgents();
+  const hostModels = agents
+    .find((agent) => agent.name === "mock-agent")!
+    .model_config.available_models.map((model) => model.id);
+  expect(hostModels.length).toBeGreaterThan(0);
+  expect(hostModels).not.toContain(profile.model);
+
+  const kanban = new KanbanPage(page);
+  await kanban.goto();
+  await page.reload();
+  if (mobile) await page.getByRole("button", { name: "Add task" }).tap();
+  else await kanban.createTaskButton.first().click();
+  const dialog = page.getByTestId("create-task-dialog");
+  const selector = dialog.getByTestId("agent-profile-selector");
+  await selector.click();
+  const option = page.getByRole("listbox").getByRole("option", { name: profile.name });
+  await expect(option).toBeEnabled();
+  await expect(option.locator("button, .tabler-icon-alert-triangle")).toHaveCount(0);
+  if (mobile) await option.tap();
+  else {
+    const search = page.locator("[cmdk-input]");
+    await search.fill(profile.name);
+    await search.press("Enter");
+  }
+  await expect(page.getByRole("listbox")).not.toBeVisible();
+  await expect(selector).toContainText(profile.name);
+  await expect(selector.locator("button, .tabler-icon-alert-triangle")).toHaveCount(0);
+  await dialog.getByTestId("task-title-input").fill("Use my saved model");
+  await dialog.getByTestId("task-description-input").fill("/e2e:simple-message");
+  await dialog.getByTestId("submit-start-agent").click();
+  if (mobile) await kanban.taskCardByTitle("Use my saved model").tap();
+  await expect(page).toHaveURL(/\/t\/[^/?]+/);
+  const taskId = new URL(page.url()).pathname.split("/")[2];
+  await expect.poll(async () => (await apiClient.listTaskSessions(taskId)).sessions.length).toBe(1);
+  const { sessions } = await apiClient.listTaskSessions(taskId);
+  const sessionId = sessions[0].id;
+  expect(sessions[0].agent_profile_id).toBe(profile.id);
+  await waitForSessionDone(apiClient, taskId, sessionId, "Waiting for requested executor model");
+
+  const session = new SessionPage(page);
+  await session.waitForLoad();
+  await assertRequestedModel(page, apiClient, taskId, sessionId, profile);
+  await page.reload();
+  await session.waitForLoad();
+  await assertRequestedModel(page, apiClient, taskId, sessionId, profile);
+  if (mobile) await assertNoDocumentHorizontalOverflow(page, "requested model after reload");
+}
+
+async function assertRequestedModel(
+  page: Page,
+  apiClient: ApiClient,
+  taskId: string,
+  sessionId: string,
+  profile: AgentProfile,
+) {
+  const { sessions } = await apiClient.listTaskSessions(taskId);
+  const current = sessions.find((session) => session.id === sessionId)!;
+  expect(current.metadata?.runtime_config).toMatchObject({
+    model: profile.model,
+    config_options: { effort: "high" },
+  });
+  expect(current.metadata?.acp_model_state).toMatchObject({
+    current_model_id: profile.model,
+    models: expect.arrayContaining([expect.objectContaining({ model_id: profile.model })]),
+  });
+  expect(await readModelSelectionWarnings(apiClient, sessionId)).toHaveLength(0);
+  const modelState = current.metadata?.acp_model_state as
+    | { models?: Array<{ model_id?: string; name?: string }> }
+    | undefined;
+  const requestedModelName = modelState?.models?.find(
+    (model) => model.model_id === profile.model,
+  )?.name;
+  expect(requestedModelName).toBeTruthy();
+  const modelTrigger = page.getByRole("button", { name: "Session model settings" });
+  await expect(modelTrigger).toContainText(requestedModelName!);
+  await expect(modelTrigger).toContainText("High");
+  const saved = await apiClient.getAgentProfile(profile.id);
+  expect(saved.model).toBe(profile.model);
+  expect(saved.config_options).toEqual(profile.config_options);
+  expect(saved.fallback_model).toBe(profile.fallback_model);
+  expect(saved.auto_fallback).toBe(profile.auto_fallback);
+}

apps/web/src/locales/en/settings.json

@@ -503,8 +503,6 @@
   "previewSound": "Preview sound",
   "profileForTasksCreatedByAgents": "Profile for Tasks Created by Agents",
   "programming": "Programming",
-  "profileStartModelNotAdvertisedOnHost": "The host probe did not advertise {{model}}. The selected executor will decide the model at launch.",
-  "profileStartModelUniqueVariationOnHost": "The host probe found one possible variation of {{model}}: {{variation}}. The selected executor will decide the model at launch.",
   "modelVariationAdvisory": "The host catalog has one possible variation of {{model}}: {{variation}}. The saved model remains unchanged until launch.",
   "promptAdd": "Add prompt",
   "promptEditorAriaLabel": "Prompt editor",

apps/web/src/locales/pseudo/settings.json

@@ -503,8 +503,6 @@
   "previewSound": "Ƥŕēvĩēŵ śōũńď",
   "profileForTasksCreatedByAgents": "Ƥŕōƒĩĺē ƒōŕ Ţàśķś Ćŕēàţēď ƀŷ Àĝēńţś",
   "programming": "Ƥŕōĝŕàḿḿĩńĝ",
-  "profileStartModelNotAdvertisedOnHost": "Ţĥē ĥōśţ ƥŕōƀē ďĩď ńōţ àďvēŕţĩśē {{model}}. Ţĥē śēĺēćţēď ēxēćũţōŕ ŵĩĺĺ ďēćĩďē ţĥē ḿōďēĺ àţ ĺàũńćĥ.",
-  "profileStartModelUniqueVariationOnHost": "Ţĥē ĥōśţ ƥŕōƀē ƒōũńď ōńē ƥōśśĩƀĺē vàŕĩàţĩōń ōƒ {{model}}: {{variation}}. Ţĥē śēĺēćţēď ēxēćũţōŕ ŵĩĺĺ ďēćĩďē ţĥē ḿōďēĺ àţ ĺàũńćĥ.",
   "modelVariationAdvisory": "Ţĥē ĥōśţ ćàţàĺōĝ ĥàś ōńē ƥōśśĩƀĺē vàŕĩàţĩōń ōƒ {{model}}: {{variation}}. Ţĥē śàvēď ḿōďēĺ ŕēḿàĩńś ũńćĥàńĝēď ũńţĩĺ ĺàũńćĥ.",
   "promptAdd": "Àďď ƥŕōḿƥţ",
   "promptEditorAriaLabel": "Ƥŕōḿƥţ ēďĩţōŕ",

apps/web/src/locales/pt-pt/settings.json

@@ -750,8 +750,6 @@
   "taskBehavior": "Comportamento das tarefas",
   "terminalAndEditors": "Terminal e editores",
   "workspacesAndAccess": "Workspaces e acesso",
-  "profileStartModelNotAdvertisedOnHost": "A sondagem do anfitrião não anunciou {{model}}. O executor selecionado decidirá o modelo no arranque.",
-  "profileStartModelUniqueVariationOnHost": "A sondagem do anfitrião encontrou uma única variação possível de {{model}}: {{variation}}. O executor selecionado decidirá o modelo no arranque.",
   "modelVariationAdvisory": "O catálogo do anfitrião tem uma única variação possível de {{model}}: {{variation}}. O modelo guardado permanece inalterado até ao arranque.",
   "navUnits": "Unidades da organização",
   "unitsTitle": "Unidades da organização",

apps/web/src/locales/zh-cn/settings.json

@@ -750,8 +750,6 @@
   "keyValueAddItem": "添加条目",
   "keyValueKeyPlaceholder": "键",
   "keyValueValuePlaceholder": "值",
-  "profileStartModelNotAdvertisedOnHost": "主机探测未公布 {{model}}。所选执行器将在启动时决定使用的模型。",
-  "profileStartModelUniqueVariationOnHost": "主机探测发现 {{model}} 的一个可能变体：{{variation}}。所选执行器将在启动时决定使用的模型。",
   "modelVariationAdvisory": "主机目录中有一个可能的 {{model}} 变体：{{variation}}。保存的模型在启动前保持不变。",
   "navUnits": "组织单元",
   "unitsTitle": "组织单元",

apps/web/src/locales/zh-hk/settings.json

@@ -750,8 +750,6 @@
   "keyValueAddItem": "新增條目",
   "keyValueKeyPlaceholder": "鍵",
   "keyValueValuePlaceholder": "值",
-  "profileStartModelNotAdvertisedOnHost": "主機探測未公佈 {{model}}。所選執行器將在啓動時決定使用的模型。",
-  "profileStartModelUniqueVariationOnHost": "主機探測發現 {{model}} 的一個可能變體：{{variation}}。所選執行器將在啓動時決定使用的模型。",
   "modelVariationAdvisory": "主機目錄中有一個可能的 {{model}} 變體：{{variation}}。已儲存的模型在啟動前保持不變。",
   "navUnits": "組織單元",
   "unitsTitle": "組織單元",

apps/web/src/locales/zh-tw/settings.json

@@ -750,8 +750,6 @@
   "keyValueAddItem": "新增條目",
   "keyValueKeyPlaceholder": "鍵",
   "keyValueValuePlaceholder": "值",
-  "profileStartModelNotAdvertisedOnHost": "主機探測未公佈 {{model}}。所選執行器將在啟動時決定使用的模型。",
-  "profileStartModelUniqueVariationOnHost": "主機探測發現 {{model}} 的一個可能變體：{{variation}}。所選執行器將在啟動時決定使用的模型。",
   "modelVariationAdvisory": "主機目錄中有一個可能的 {{model}} 變體：{{variation}}。已儲存的模型在啟動前保持不變。",
   "navUnits": "組織單元",
   "unitsTitle": "組織單元",

docs/plans/profile-selector-model-warnings/plan.md

@@ -0,0 +1,203 @@
+---
+created: 2026-09-08
+status: completed
+requirements:
+  - REQ-AGENTS-NO-SILENT-MODEL-FALLBACK-001
+  - REQ-AGENTS-NO-SILENT-MODEL-FALLBACK-002
+  - REQ-AGENTS-NO-SILENT-MODEL-FALLBACK-003
+system_design:
+  - ../../specs/agents/system-design/no-silent-model-fallback-01.md
+  - ../../specs/agents/system-design/no-silent-model-fallback-02.md
+legacy_specs: []
+---
+
+# Implementation Plan: Profile selector model warnings
+
+## Overview
+
+Remove host model advisories from saved-profile selectors. Keep profile editing
+diagnostics and actual executor model-selection warnings available where users
+can act on them.
+
+The agent system owns this change because it owns profile model intent and
+executor model authority. Amend the existing
+[requirements](../../specs/agents/requirements/no-silent-model-fallback.md);
+do not create a separate UI specification.
+
+One work order delivers the shared selector change with unit, desktop, mobile,
+and runtime-warning regression coverage. The user authorized implementation,
+PR creation, a 15-minute wait, and PR fixup through `@implement`.
+Deployment remains outside this package.
+
+## Evidence and classification
+
+This is a change to intended product behavior. The former system design
+explicitly required the amber selector icon; the implementation follows that
+rule. Requirement 002.7 now locates the host advisory in the profile editor.
+Requirement 003 defines selector behavior and preservation checks.
+
+Read-only investigation on 2026-09-08 established:
+
+- The live instance used backend port 38429 and database
+  `/root/.kandev/data/kandev.db`.
+- Profile `4fe051a9-f5cc-422b-bd0c-05048c22699e`, named Astra High, stored
+  `gpt-6-astra` and `settings.config_options.reasoning_effort = high`.
+- Session `cfac5c57-f9a8-4b89-bc4c-2db8a0d6e20f` persisted that model and
+  effort in its settled runtime configuration.
+- At 22:29:53 +01:00, logs reported that start-model policy applied Astra and
+  published `current_model_id = gpt-6-astra`.
+- A fresh `acpdbg probe --timeout 45s codex-acp` completed with exit code 0.
+  Codex ACP 1.10.0 exposed 31 legacy model/effort IDs and six canonical model
+  configuration values. Both representations included Astra.
+- Both live capability and available-agent APIs contained the canonical Astra
+  ID at inspection time.
+
+The confirmed warning mechanism is the exact membership check in
+`useAgentProfileOptions`: a nonempty browser host catalog that omits the
+saved ID produces an amber advisory in both render paths. It never checks the
+eventual executor outcome. A host catalog with several bracketed variants also
+takes this path because no single variation matches.
+
+The browser catalog at screenshot time was not captured. Staleness and legacy
+ID representation are possible triggers, not established cache defects. The
+design does not depend on choosing between those hypotheses.
+
+The temporary probe capture was
+`/tmp/kandev-model-warning-probe-OL5GGn/codex-acp-probe-20260908-213045.jsonl`.
+It is optional local evidence, not a dependency for implementation. Do not
+commit raw protocol frames or copy the live database into test fixtures.
+
+## Scope
+
+### In scope
+
+- Remove host model advisory icons, help triggers, and selected-label indicators.
+- Apply the result through the existing shared profile-option hook.
+- Preserve capability health indicators, eligibility, recent-use ordering,
+  profile labels, saved configuration, and passthrough indicators.
+- Preserve profile-editor model diagnostics and persisted task warnings.
+- Remove unused selector translations and explain warning placement in public
+  agent-profile documentation when the implementation ships.
+- Cover desktop keyboard selection, mobile row taps, successful model launch,
+  actual fallback, and reload.
+
+### Out of scope
+
+- Host probe refresh, browser cache invalidation, or new polling.
+- ACP normalization changes or provider-specific bracket parsing.
+- Model/fallback policy, runtime warning metadata, database schema, or migrations.
+- Runtime model-selector redesign, profile-editor redesign, Office routing,
+  and unrelated capability-warning accessibility changes.
+- New flags, telemetry, screenshots for public docs, or provider network calls.
+
+## Technical approach
+
+In `apps/web/components/task-create-dialog-options.tsx`, remove the
+model-catalog comparison and the two model-warning renderers. Keep the
+capability hook active, and use its snapshot only to refresh health fields on
+profile options. Keep the existing option interface and shared label rendering
+compatible with consumers.
+
+The shared hook serves these existing callers:
+
+- `components/task-create-dialog-computed.ts`
+- `components/task/new-subtask-dialog.tsx`
+- `components/task/new-session-dialog.tsx`
+- `components/quick-chat/quick-chat-setup.tsx`
+- `components/automations/config-section.tsx`
+- `app/office/setup/agent-profile-setup-controls.tsx`
+
+These callers require no separate presentation branches. In particular,
+Office keeps its surrounding compatibility filtering.
+
+Do not remove `getCapabilityWarning`, model fields from profile state, or the
+editor's `findUniqueModelVariation` helper. Remove only the two unused
+`settings` selector keys from every locale, including pseudo. Keep
+`modelVariationAdvisory` and runtime-warning translations.
+
+The profile editor already prefers structured model configuration options in
+`components/settings/profile-model-fields.tsx`. Preserve that canonical source;
+this task does not introduce another model-ID resolver.
+
+Public documentation is owned by the same work order: add a short explanation
+under “Host probes and executor model catalogs” in
+`docs/public/agents-and-profiles.md`. This is an explanation section within
+the existing agent-profile guide.
+
+## Tests
+
+The following regression scenarios define the acceptance coverage.
+
+| Acceptance | Test file and scenario |
+| --- | --- |
+| 003.1 | `components/task-create-dialog-options.test.tsx`: “does not show host model advisories in option or selected labels”, parameterized over exact, missing, unique, multiple, bracketed, empty, pending, and changed host catalogs. |
+| 003.2, 003.6 | Same file: preserve eligibility, ordering, profile label and saved values; retain existing disabled-profile and recent-use cases. |
+| 003.3 | Same file: “preserves capability health warnings independently of model membership”, for auth_required, not_installed, failed, and healthy status. |
+| 003.4, 002.7 | `components/settings/profile-form-fields.test.tsx` and `lib/model-variation.test.ts`: retain missing-model, unique-variation, and fallback-control coverage. |
+| 003.5, 001.4 | `components/task/chat/messages/status-message.test.tsx`: retain structured model-selection warning rendering. |
+| 001.5, 003.6 | Browser reload checks and saved-profile API comparison in the successful-launch scenario below. |
+
+The first red test must render a healthy-capability profile with a missing host
+model and assert no advisory in either label. It fails on the current
+implementation because the dropdown renders a warning button and the selected
+label renders a titled icon. Absence of the old test ID alone is insufficient.
+
+## E2E tests
+
+| Project / file under `apps/web/e2e/tests` | Outcome and acceptance |
+| --- | --- |
+| chromium / `settings/no-silent-model-fallback.spec.ts` | Replace tooltip assertions with warning-free missing/unique profile selection. Check option and selected label, keyboard selection, and no nested warning control. Covers 003.1–003.2. |
+| mobile-chrome / `settings/mobile-no-silent-model-fallback.spec.ts` | Replace warning-drawer taps with row selection. Verify selection closes the picker, updates its label, and leaves the create flow usable without horizontal overflow. Covers 003.1–003.2. |
+| Both settings files above | Add “launches a host-mismatched profile on its requested executor model without a warning”. Submit through the UI, verify effective model and zero model-selection warnings after completion and reload, then compare saved profile configuration. Covers 003.5–003.6 and 001.5. |
+| chromium / `session/model-mismatch-warning.spec.ts` | Preserve actual fallback, unique-variation, ambiguous-catalog, single-warning, and reload coverage. Covers 001.4 and the preservation side of 003.5. |
+| mobile-chrome / `session/mobile-model-mismatch-warning.spec.ts` | Preserve visible actual fallback warning and reload without horizontal overflow. Covers 001.4 and 003.5. |
+
+For the successful-launch fixture, extend
+`session/model-mismatch-warning-helpers.ts` with a disposable mock profile
+whose exact model is `opus[270k]` and whose profile environment selects the
+existing `MOCK_AGENT_MODEL_CATALOG=ambiguous` fixture. The default E2E host
+catalog includes `opus[1m]`, so it cannot serve as the omitted model. Confirm that the host
+catalog omits that ID while the executor advertises it. Assert those
+preconditions so an ordinary exact-match launch cannot pass as the regression.
+Use existing model options for valid saved configuration values; no live Codex
+account is needed.
+
+## Work orders
+
+- [x] [Task 01: Remove host model advisories from profile selectors](task-01-remove-host-model-advisories.md) — done, sequential, no dependencies.
+
+## Verification results
+
+Planning checks on 2026-09-08:
+
+- Specification-linter tests: 30 passed.
+- Full specification lint: passed after reducing the amended design to its size limit.
+- Whitespace check: passed.
+
+Implementation checks on 2026-09-08:
+
+- Red: 11 component assertions and two selector cases per browser project failed
+  on the former model-warning controls.
+- Green: 56 targeted unit tests, six desktop browser tests, and four mobile
+  browser tests passed.
+- Typecheck, targeted ESLint, localization checks, and the localization ratchet passed.
+- Public-documentation validator: 61 tests passed and 46 pages validated.
+- Desktop and mobile screenshots show selectable profiles without a model
+  advisory or an empty warning control.
+
+Task 01 records the commands, fixture correction, and mobile interaction detail.
+
+## Risks
+
+- Removing all warning icons would hide agent health errors; tests must distinguish
+  those indicators from model advisories.
+- Removing only the dropdown button leaves the selected-label warning intact.
+- Capability polling is shared with the selector hook. Keep it independent from
+  model-ID matching so health status remains current in selector-only surfaces.
+- Old mobile tests deliberately open the warning drawer. Replace those steps
+  with completed row selection rather than deleting the tests.
+- The saved screenshot does not establish a stale-cache defect. Do not expand
+  implementation into speculative discovery changes.
+- Runtime unique-variation behavior is already covered by current code and
+  requirements, although its ADR remains proposed. This package does not
+  change that policy or its decision status.

docs/plans/profile-selector-model-warnings/task-01-remove-host-model-advisories.md

@@ -0,0 +1,230 @@
+---
+id: "01-remove-host-model-advisories"
+title: "Remove host model advisories from profile selectors"
+status: done
+wave: 1
+depends_on: []
+plan: "plan.md"
+requirements:
+  - REQ-AGENTS-NO-SILENT-MODEL-FALLBACK-001
+  - REQ-AGENTS-NO-SILENT-MODEL-FALLBACK-002
+  - REQ-AGENTS-NO-SILENT-MODEL-FALLBACK-003
+acceptance_criteria:
+  - AC-AGENTS-NO-SILENT-MODEL-FALLBACK-001.4
+  - AC-AGENTS-NO-SILENT-MODEL-FALLBACK-001.5
+  - AC-AGENTS-NO-SILENT-MODEL-FALLBACK-002.7
+  - AC-AGENTS-NO-SILENT-MODEL-FALLBACK-003.1
+  - AC-AGENTS-NO-SILENT-MODEL-FALLBACK-003.2
+  - AC-AGENTS-NO-SILENT-MODEL-FALLBACK-003.3
+  - AC-AGENTS-NO-SILENT-MODEL-FALLBACK-003.4
+  - AC-AGENTS-NO-SILENT-MODEL-FALLBACK-003.5
+  - AC-AGENTS-NO-SILENT-MODEL-FALLBACK-003.6
+system_design:
+  - ../../specs/agents/system-design/no-silent-model-fallback-01.md
+  - ../../specs/agents/system-design/no-silent-model-fallback-02.md
+---
+
+# Task 01: Remove host model advisories from profile selectors
+
+## Summary
+
+Remove the host-catalog warning from profile options and their selected labels.
+Keep agent health indicators, editor diagnostics, and actual model-selection
+warnings. Deliver the change through the shared hook with focused regression
+tests and updated public guidance.
+
+## In scope
+
+- Own the shared hook, associated component tests, selector-copy cleanup, and
+  existing desktop/mobile model-selection E2E files.
+- Add a successful exact-model launch with differing host and executor catalogs,
+  using disposable mock profiles and existing catalog fixtures.
+- Keep the existing actual-fallback browser tests as preservation checks.
+- Update the agent-profile guide's explanation of warning placement.
+- Update this work order and the plan's results after the exact checks pass.
+
+## Out of scope
+
+- Production Go changes, ACP normalization, discovery cache/polling, schema,
+  migrations, model-selection policy, or runtime warning metadata.
+- Caller-specific warning flags, new overlays, new settings, or profile rewrites.
+- Broad verification and repository-wide review. The user's later `@implement`
+  request separately authorized commit, push, PR creation, and PR fixup.
+
+## Acceptance
+
+1. Both label render paths pass the plan's host-catalog matrix without model
+   advisories. Existing health indicators, eligibility, ordering, and saved
+   profile configuration remain intact.
+2. Desktop keyboard and mobile touch flows select and launch the requested
+   profile. A successful exact-model launch has no model-selection warning;
+   actual fallback still displays one persisted warning after reload.
+3. Unused selector translations are removed consistently. The profile editor
+   keeps its advisories and the public guide explains the shipped behavior.
+
+## Implementation sequence
+
+1. Read the linked requirements and design section 8, scoped web guidance,
+   existing source, and test helpers. Mark this work order in progress only
+   after the user requests implementation.
+2. Run the new component absence assertions against the current implementation
+   and record their expected failure. Cover both render paths, not only the old
+   warning-button test ID.
+3. Update the desktop/mobile selector scenarios for the intended behavior and
+   run the focused files against the current production build to record red.
+   Keep actual-fallback tests intact.
+4. Remove the host model comparison, warning renderers, and exclusive imports.
+   Preserve the hook interface, health warning renderer, and caller data loading.
+5. Complete the successful executor-model fixture, saved-value checks, and
+   localization cleanup. Update public guidance and run the commands below.
+6. Mark done only after targeted checks pass. Record outcomes and any remaining
+   limitations in this file and the manifest.
+
+## Verification
+
+Run from the repository root. Install once in a fresh worktree:
+
+```bash
+rtk pnpm --dir apps install --frozen-lockfile
+```
+
+Use the first command for red and green component checks. The listed preservation
+suites cover editor and chat behavior without a broad audit:
+
+```bash
+rtk pnpm --dir apps/web exec vitest run components/task-create-dialog-options.test.tsx components/settings/profile-form-fields.test.tsx lib/model-variation.test.ts components/task/chat/messages/status-message.test.tsx
+rtk pnpm --dir apps/web exec eslint components/task-create-dialog-options.tsx components/task-create-dialog-options.test.tsx
+rtk pnpm --dir apps/web run typecheck
+rtk pnpm --dir apps/web run i18n:check
+rtk pnpm --dir apps/web run i18n:ratchet
+```
+
+Run projects sequentially. The managed runner builds production assets and
+isolates its instance. Check discovery counts; zero tests is not a pass.
+
+```bash
+rtk pnpm --dir apps/web e2e:run --project chromium tests/settings/no-silent-model-fallback.spec.ts tests/session/model-mismatch-warning.spec.ts
+rtk pnpm --dir apps/web e2e:run --project mobile-chrome tests/settings/mobile-no-silent-model-fallback.spec.ts tests/session/mobile-model-mismatch-warning.spec.ts
+```
+
+Inspect the rendered mobile selection state or its test screenshot for row
+spacing, viewport containment, and the absence of an empty warning hit area.
+Tests must verify selection completion and no document horizontal overflow.
+
+After public-documentation changes:
+
+```bash
+rtk node --test scripts/validate-public-docs.test.mjs
+rtk node scripts/validate-public-docs.mjs
+rtk git diff --check
+```
+
+Audit removed controls and keys:
+
+```bash
+rtk rg -n 'agent-profile-model-probe-warning|profileStartModelNotAdvertisedOnHost|profileStartModelUniqueVariationOnHost' apps/web
+```
+
+Only intentional negative assertions may retain the removed control identity.
+Production references to the two translation keys must be absent. Keep
+`modelVariationAdvisory` and existing runtime-warning keys.
+
+## Files likely touched
+
+- `apps/web/components/task-create-dialog-options.tsx`
+- `apps/web/components/task-create-dialog-options.test.tsx`
+- `apps/web/e2e/tests/settings/no-silent-model-fallback.spec.ts`
+- `apps/web/e2e/tests/settings/mobile-no-silent-model-fallback.spec.ts`
+- `apps/web/e2e/tests/settings/profile-model-selection-helpers.ts`
+- `apps/web/e2e/tests/session/model-mismatch-warning-helpers.ts`
+- `apps/web/src/locales/{en,pseudo,pt-pt,zh-cn,zh-hk,zh-tw}/settings.json`
+- `docs/public/agents-and-profiles.md`
+
+Preservation inputs, normally unchanged:
+
+- `apps/web/lib/capability-warning.ts`
+- `apps/web/components/settings/profile-model-fields.tsx`
+- `apps/web/components/settings/profile-form-fields.test.tsx`
+- `apps/web/lib/model-variation.ts` and its test file
+- `apps/web/components/task/chat/messages/status-message.test.tsx`
+- `apps/web/e2e/tests/session/model-mismatch-warning.spec.ts`
+- `apps/web/e2e/tests/session/mobile-model-mismatch-warning.spec.ts`
+- The six shared-hook consumers listed in the manifest
+
+## Dependencies
+
+None.
+
+## Risks
+
+The selected-label path can retain a warning after the option path is fixed.
+Blanket icon removal can hide capability errors. Fixture setup must prove
+different host and executor catalogs, and cleanup must delete only the
+disposable profiles created by the tests.
+
+Keep capability polling active through the shared selector hook. Apply its
+snapshot to health fields only. Do not reintroduce model-ID matching.
+
+## Parallelism
+
+`sequential`. No subagents are authorized.
+
+## Inputs
+
+- [Requirements](../../specs/agents/requirements/no-silent-model-fallback.md):
+  001.4–001.5, 002.7, and 003.
+- [Design part 1](../../specs/agents/system-design/no-silent-model-fallback-01.md):
+  section 8, runtime warning contract, and tests.
+- [Design part 2](../../specs/agents/system-design/no-silent-model-fallback-02.md):
+  warning-removal and evidence risks.
+- [Executor authority ADR](../../decisions/2026-08-15-executor-authoritative-model-selection.md).
+- [Variation resolution ADR](../../decisions/2026-09-07-unique-model-variation-resolution.md).
+- Existing option/selected-label renderers and their component harness.
+- Existing model-mismatch helpers and mock catalog environment fixture.
+- Mobile exemplar and composition rationale in design section 8.
+
+## Results
+
+Completed on 2026-09-08.
+
+- Removed host-catalog membership checks and both model-warning render paths
+  from the shared hook. Kept capability polling active and applied its snapshot
+  to health fields only. Eligibility and recent-use behavior remain unchanged.
+- Removed the two selector-only keys from all six locale catalogs. Editor
+  diagnostics and runtime-warning code remain unchanged.
+- The red component run failed 11 of 23 assertions. The old production build
+  failed both selector cases in each browser project on the warning controls.
+- The four targeted unit suites passed all 56 tests.
+- The managed desktop run passed six tests with `--host --project chromium`.
+  The mobile run passed four tests with `--host --no-build --project mobile-chrome`.
+  Both used the exact file lists in Verification. The mobile run reused the
+  unchanged production assets from the desktop run.
+- The successful-launch fixture proves that the host omits `opus[270k]` and the
+  executor advertises it. It verifies the model, effort, zero warnings, and
+  unchanged saved profile after completion and reload.
+- The first mobile launch run failed because it assumed desktop navigation.
+  The corrected test taps the created task card before it checks the session.
+- Targeted ESLint passed for the component, its test, and all four changed or
+  new E2E files. Typecheck, `i18n:check`, and `i18n:ratchet` passed.
+- Public-documentation validation passed: 61 validator tests and 46 pages.
+- Fresh desktop/mobile screenshots show both the option list and selected label.
+  The mobile screenshots show usable row spacing with no empty warning control.
+
+### PR fixup remediation
+
+- Greptile's P1 finding was valid: the shared selector hook also kept capability
+  revalidation alive for selector-only surfaces. The hook now calls
+  `useAvailableAgents` and applies `refreshProfileCapabilities` to the local
+  option input. The host catalog supplies health status only. It cannot remove,
+  disable, or rewrite a profile because of a model ID.
+- The remediation reran the focused desktop project (six passed) and mobile
+  project (four passed) after rebuilding the desktop production assets.
+- The E2E launch helper now derives the expected model display name from the
+  executor's settled ACP model state. It no longer couples the assertion to a
+  fixture display string.
+- Claude's trigger-label suggestion was already satisfied by the parameterized
+  test loop, which checks positive profile text and warning absence in both
+  render paths.
+
+No backend changes, schema changes, live-instance restarts, or profile rewrites
+were required. The screenshot-time browser catalog remains unknown.

docs/public/agents-and-profiles.md

@@ -258,7 +258,10 @@ stale browser action does not replace a newer route decision.
 
 The model list shown while editing a profile comes from a host probe. It is an
 editing hint, not a launch gate. A profile remains selectable when its saved
-model is missing from that host list.
+model is missing from that host list. Profile selectors do not show a model
+warning for this difference. Inspect the model list in profile settings for
+discovery details. Authentication, installation, and probe-failure indicators
+remain visible on profile selectors.
 
 At task launch, the selected executor's ACP catalog is authoritative. For
 profiles without automatic fallback, Kandev follows the four-step order above.

docs/specs/agents/README.md

@@ -50,7 +50,8 @@ surface shared by task and Office consumers.
 - [Agent MCP Timeout Budgets](requirements/mcp-timeout-budgets.md)
 - [Mock-agent slow command duration syntax](requirements/mock-agent-slow-duration.md)
 - [Native Code Review](requirements/native-code-review.md)
-- [No Silent Model Fallback](requirements/no-silent-model-fallback.md)
+- [No Silent Model Fallback](requirements/no-silent-model-fallback.md), including
+  profile-selector warning placement and executor-authoritative task warnings.
 - [Copy agent configuration to isolated executors](requirements/portable-agent-configuration.md)
 - [Disable an Agent Profile](requirements/profile-disable.md)
 - [Agent Profile Recent Use](requirements/profile-recent-use.md)

docs/specs/agents/requirements/no-silent-model-fallback.md

@@ -82,11 +82,44 @@ my session to an unrelated default model.
   variation, it shall keep the saved model unchanged and shall persist a
   warning that identifies the inferred effective model.
 - **AC-AGENTS-NO-SILENT-MODEL-FALLBACK-002.7:** The host profile probe shall
-  present unique-variation resolution as an advisory. It shall not become the
+  present unique-variation resolution as an advisory in the profile editor.
-  launch authority or disable the profile.
+  It shall not become the launch authority or add a model warning to a profile
+  selector. It shall not disable the profile.
 - **AC-AGENTS-NO-SILENT-MODEL-FALLBACK-002.8:** The same resolution order shall
   apply at initial launch, context reset, and workspace rebind.
 
+### REQ-AGENTS-NO-SILENT-MODEL-FALLBACK-003: Show model warnings at the point of action
+
+**Intent:** Let users select a saved profile without treating a host discovery
+difference as evidence of an executor failure.
+
+#### Acceptance criteria
+
+- **AC-AGENTS-NO-SILENT-MODEL-FALLBACK-003.1:** A host model-catalog difference
+  shall not add an icon, message, tooltip, or help action to a profile selector.
+  This applies to its options and its selected label, including missing models,
+  one or several model variations, and empty or pending catalogs.
+- **AC-AGENTS-NO-SILENT-MODEL-FALLBACK-003.2:** On desktop and mobile, an
+  otherwise eligible profile shall remain selectable by keyboard or pointer.
+  On touch devices, tapping its row shall select it without opening model help.
+- **AC-AGENTS-NO-SILENT-MODEL-FALLBACK-003.3:** Existing authentication,
+  installation, and capability-probe failure indicators shall remain visible.
+  Their presence shall not depend on whether the saved model is advertised.
+- **AC-AGENTS-NO-SILENT-MODEL-FALLBACK-003.4:** The profile editor shall retain
+  its missing-model treatment, unique-variation advisory, and fallback controls.
+- **AC-AGENTS-NO-SILENT-MODEL-FALLBACK-003.5:** When the executor applies the
+  requested model successfully, a host catalog difference shall not cause a
+  model-selection warning in task chat. Actual fallback warnings shall retain
+  the persistence and visibility defined by requirement 001.
+- **AC-AGENTS-NO-SILENT-MODEL-FALLBACK-003.6:** Selecting a profile shall not
+  rewrite its saved model, reasoning configuration, or fallback settings.
+
+## Compatibility
+
+The September 2026 selector amendment removes host model advisories from
+profile selection. It preserves executor model selection and profile editing.
+Delivery is tracked in the [implementation package](../../../plans/profile-selector-model-warnings/plan.md).
+
 ## Out of scope
 
 - Ranking multiple variations by context size, speed, price, or display order.
@@ -94,6 +127,7 @@ my session to an unrelated default model.
 - Inferring a base model from an already bracketed request.
 - Rewriting the profile model to the current executor's advertised variation.
 - Changing Office post-start provider routing or mid-turn model switching.
+- Changing host discovery refresh, cache invalidation, or ACP model normalization.
 
 ## System design
 

docs/specs/agents/system-design/no-silent-model-fallback-01.md

@@ -4,6 +4,7 @@ system: agents
 requirements:
   - REQ-AGENTS-NO-SILENT-MODEL-FALLBACK-001
   - REQ-AGENTS-NO-SILENT-MODEL-FALLBACK-002
+  - REQ-AGENTS-NO-SILENT-MODEL-FALLBACK-003
 created: 2026-08-23
 owners:
   - kandev
@@ -21,6 +22,7 @@ resolution from a bare requested model to one advertised bracketed variation.
 | --- | --- |
 | `REQ-AGENTS-NO-SILENT-MODEL-FALLBACK-001` | [Migrated source detail](#migrated-source-detail) |
 | `REQ-AGENTS-NO-SILENT-MODEL-FALLBACK-002` | [Unique model-variation resolution](#unique-model-variation-resolution) |
+| `REQ-AGENTS-NO-SILENT-MODEL-FALLBACK-003` | [Profile selection and warning placement](#8-profile-selection-and-warning-placement) |
 
 ## Migrated source detail
 
@@ -139,7 +141,7 @@ over `fallback_model`):
 | Session start, model selection unsupported | Continue on the agent default and persist a warning. | Same | Same |
 | Mid-session model/auth failure (office run, post-start) | Unchanged ADR behavior: office re-dispatches via the workspace routing chain (`routingerr.Decide(ContextOffice)`; availability codes → `DecisionFallback`). The profile's model policy does **not** gate office fallback — the workspace routing configuration is the office authorization owner. | Same as default-on-mismatch: `fallback_model` is a session-start policy, not an office routing input. | Legacy: re-dispatch to next candidate in the provider order (unchanged). |
 | Boot reconciliation | Never overwrite a gone start model (keep it; UI shows it red). Same for a gone `fallback_model`. | Same | Same (reconciler is mode-independent). |
-| New-task / new-agent profile picker | Profile selectable with a host-catalog warning. | Profile selectable with a host-catalog warning. | Profile selectable with a host-catalog warning. |
+| New-task / new-agent profile picker | Profile selectable without a host model advisory. | Same. | Same. |
 | Model picker (profile editor, session toolbar) | Gone models greyed out, unselectable, visible. | Same. | Same. |
 
 `SetModel` failures that mean "this agent does not support model selection"
@@ -227,17 +229,17 @@ table. This calculation is advisory because the host catalog can differ from
 the selected executor catalog.
 
 When the host advertises one variation, the task profile option remains
-selectable and identifies the possible target in its existing amber advisory.
+selectable without a host model advisory. The profile editor keeps the saved
-The profile editor keeps the saved bare ID visible and describes the unique
+bare ID visible and describes the unique
 host variation instead of reporting that the model is simply gone. Selecting
 the advertised variation remains an explicit way to update the profile.
 
-When the host advertises zero or multiple variations, the current missing-model
+When the host advertises zero or multiple variations, the profile editor's
-presentation remains. The runtime can still reach a different decision from
+missing-model presentation remains. The runtime can still reach a different decision from
 the selected executor catalog.
 
-The task-create advisory keeps its existing desktop tooltip and coarse-pointer
+The task-create selector has no host model tooltip or help drawer.
-drawer. The profile editor uses its existing responsive selector. The durable
+The profile editor uses its existing responsive selector. The durable
 chat warning uses the existing status-message layout on desktop and mobile.
 This change adds no nested scroll area or hover-only information.
 
@@ -496,27 +498,57 @@ All new copy is externalized via `t()` into the `settings` i18n namespace
   (`apps/web/src/locales/{en,pseudo,pt-pt,zh-cn}/settings.json`) — the i18n ratchet
 judges added lines even in unmigrated files.
 
-### 8. Profile picker warnings (new-task / new-agent)
+### 8. Profile selection and warning placement
 
 `apps/web/lib/state/slices/settings/types.ts` — `AgentProfileOption` gains
 `model`, `fallbackModel`, `autoFallback` (populated in
 `toAgentProfileOption`).
 
 `apps/web/components/task-create-dialog-options.tsx`
-(`useAgentProfileOptions`) can compute a host-catalog difference.
+(`useAgentProfileOptions`) derives profile labels and eligibility without
-This difference is advisory only.
+comparing the saved model with the host model catalog. Remove
+`advertisedModelIDs`, `ModelProbeWarning`, and `ModelProbeWarningIndicator` from
+this module, plus imports used only by those helpers. The hook no longer needs
+`useAvailableAgents` solely to produce model advisories.
 
-- Every profile remains selectable.
+Both `renderLabel` and `renderTriggerLabel` use the same profile presentation.
-- A missing host model shows one amber warning icon beside the profile name.
+Keep the option interface compatible with its consumers. Preserve profile
-- On fine pointers, hovering or focusing the warning icon reveals the full
+eligibility, recent-use ordering, names, logos, and CLI passthrough indicators.
-  localized advisory that the executor decides availability at launch. On
+Keep `getCapabilityWarning` and its authentication, installation, and probe
-  coarse pointers, tapping the icon opens the same advisory in a drawer. The
+failure states. Those states describe agent health rather than a model mismatch.
-  advisory is not shown as an always-visible secondary row in the option list.
+
-- The warning does not promise that an explicit fallback will be available.
+Consumers include task creation, new subtasks, new sessions, quick chat,
-- The warning does not change the saved profile model.
+automation configuration, and Office setup. All receive this behavior from the
+shared hook. Do not add consumer-specific warning suppression flags.
+
+Retain profile model/configuration fields and the editor's
+`findUniqueModelVariation`. After checking references, remove these unused
+selector keys from every locale:
+`profileStartModelNotAdvertisedOnHost` and
+`profileStartModelUniqueVariationOnHost`.
 
 `apps/web/app/office/setup/agent-profile-setup-controls.tsx`
-(`useSelectableProfileOptions`) uses the same advisory behavior.
+(`useSelectableProfileOptions`) retains its Office eligibility filtering around
+the shared profile options.
+
+#### Desktop and mobile composition
+
+Keep the touch-usable combobox and shared selection logic. A row tap completes
+the temporary profile choice without another help drawer or an empty hit area.
+Preserve keyboard selection, focus return, picker-owned scrolling, and safe-area
+constraints. Assert no document horizontal overflow.
+
+Use `e2e/tests/settings/mobile-no-silent-model-fallback.spec.ts` as the shipped
+flow and `components/task/mobile/mobile-picker-sheet.tsx` as the curated
+temporary-choice precedent. No picker-shell replacement is needed.
+
+#### Rationale and compatibility
+
+A host observation cannot establish the executor outcome. A neutral icon or
+better ID comparison would still report valid catalog differences during
+selection. Keep the existing executor-authority ADR and runtime behavior.
+This change needs no resolver, cache, schema, or new ADR. Explain warning
+placement in `docs/public/agents-and-profiles.md` when implementation ships.
 
 The persisted `model_selection_warning` chat message renders through
 `apps/web/components/task/chat/messages/status-message.tsx`.
@@ -566,8 +598,11 @@ Frontend (Vitest, `*.test.ts(x)`):
 
 - `model-config-selector`: disabled option not selectable; greyed class.
 - `session-models` WS handler: stale active model is kept (not cleared).
-- `useAgentProfileOptions`: every host-mismatch profile remains selectable and
+- `useAgentProfileOptions`: every otherwise eligible host-mismatch profile
-  shows an advisory warning.
+  remains selectable with no model advisory in either label. Cover canonical
+  IDs, legacy bracketed IDs, unique and multiple variations, and empty catalogs.
+- Capability health warnings, recent-use ordering, disabled-profile filtering,
+  and saved profile values remain unchanged by selector rendering.
 - Profile editor: gone start model renders red + disabled; auto-fallback keeps
   the explicit fallback choice visible but disables its controls.
 - Profile editors: fallback settings start collapsed; expanding exposes both
@@ -581,8 +616,11 @@ E2E (Playwright, `apps/web/e2e`):
 
 - Mock backend (`KANDEV_E2E_MOCK=true`): create a profile whose start model
   is not in the host catalog. Make sure that the task-create picker keeps the
-  profile selectable, shows one warning icon, and reveals the advisory warning
+  profile selectable without a host model warning in the option or selected
-  through the fine-pointer tooltip or coarse-pointer drawer.
+  label. Select the profile with keyboard on desktop and a row tap on mobile.
+- Launch a profile whose executor advertises its exact model while the host
+  omits it. Verify the selected model, no model-selection warning, and unchanged
+  profile values after reload.
 - Launch with an executor catalog that omits the profile model. Make sure that
   no model-selection call occurs, the task continues, and chat shows one warning.
 - Reload the task page. Make sure that the warning remains in chat.

docs/specs/agents/system-design/no-silent-model-fallback-02.md

@@ -4,6 +4,7 @@ system: agents
 requirements:
   - REQ-AGENTS-NO-SILENT-MODEL-FALLBACK-001
   - REQ-AGENTS-NO-SILENT-MODEL-FALLBACK-002
+  - REQ-AGENTS-NO-SILENT-MODEL-FALLBACK-003
 created: 2026-08-23
 owners:
   - kandev
@@ -21,6 +22,7 @@ variation resolution.
 | --- | --- |
 | `REQ-AGENTS-NO-SILENT-MODEL-FALLBACK-001` | [Migrated source detail](#migrated-source-detail) |
 | `REQ-AGENTS-NO-SILENT-MODEL-FALLBACK-002` | [Risks & Open Questions](#risks--open-questions) |
+| `REQ-AGENTS-NO-SILENT-MODEL-FALLBACK-003` | [Risks & Open Questions](#risks--open-questions) |
 
 ## Migrated source detail
 
@@ -36,8 +38,8 @@ variation resolution.
   workspace routing configuration, not the execution profile — and is
   documented in the behavior matrix above.
 - **Probe staleness**: the advertised list can be stale (probe cached).
-  The profile picker uses it only as a hint. The executor session catalog owns
+  The profile editor uses it only as a hint. Profile selectors do not render
-  the launch decision.
+  host model advisories. The executor session catalog owns the launch decision.
 - **Cold Claude model lists**: a valid restricted model can be absent from a
   cold bridge's initial list. Pre-session exposure lets the bridge include and
   select the configured model. If the bridge still omits it, Kandev uses the
@@ -46,7 +48,7 @@ variation resolution.
   before the initial process starts. It does not restart a live ACP bridge to
   expose a newly selected hidden model during context reset.
 - **Office vs. kanban surfaces**: both share the same agent-profile rows.
-  The advisory picker behavior covers kanban task creation and Office setup.
+  The shared selector behavior covers kanban task creation and Office setup.
   Office run-detail routing surfaces are unchanged.
 - **Collapsed controls remain legible**: the disclosure header summarizes the
   effective mode, and dirty-state decoration is applied to the disclosure
@@ -65,3 +67,9 @@ variation resolution.
 - **Legacy automatic fallback remains unchanged**: when `auto_fallback` is
   enabled and the requested model is absent, Kandev does not apply an explicit
   fallback or infer a variation. It continues with the provider default.
+- **Removal must be narrow**: authentication, missing-CLI, and failed-probe
+  indicators remain. Profile-editor advisories and persisted task warnings
+  retain their existing behavior.
+- **Evidence has limits**: a model warning does not prove a cache defect.
+  Discovery refresh or normalization changes require separate root-cause
+  evidence; this amendment does not change those contracts.