diff --git a/packages/server/src/server/agent/providers/claude/agent.test.ts b/packages/server/src/server/agent/providers/claude/agent.test.ts
index 6368e0398..333899b4f 100644
--- a/packages/server/src/server/agent/providers/claude/agent.test.ts
+++ b/packages/server/src/server/agent/providers/claude/agent.test.ts
@@ -9,6 +9,7 @@ import * as executableUtils from "../../../../utils/executable.js";
 import {
   ClaudeAgentClient,
   convertClaudeHistoryEntry,
+  normalizeClaudeAskUserQuestionRequestInput,
   normalizeClaudeAskUserQuestionUpdatedInput,
   toClaudeSdkMcpConfig,
 } from "./agent.js";
@@ -614,6 +615,37 @@ describe("ClaudeAgentSession features", () => {
 });
 
 describe("normalizeClaudeAskUserQuestionUpdatedInput", () => {
+  test("marks Claude AskUserQuestion options as allowing other answers", () => {
+    expect(
+      normalizeClaudeAskUserQuestionRequestInput("AskUserQuestion", {
+        questions: [
+          {
+            question: "Which provider should I use?",
+            header: "Provider",
+            options: [
+              { label: "Claude", description: "Use Claude Code" },
+              { label: "Codex", description: "Use Codex" },
+            ],
+            multiSelect: false,
+          },
+        ],
+      }),
+    ).toEqual({
+      questions: [
+        {
+          question: "Which provider should I use?",
+          header: "Provider",
+          options: [
+            { label: "Claude", description: "Use Claude Code" },
+            { label: "Codex", description: "Use Codex" },
+          ],
+          multiSelect: false,
+          allowOther: true,
+        },
+      ],
+    });
+  });
+
   test("maps frontend header-keyed answers to Claude question text keys", () => {
     expect(
       normalizeClaudeAskUserQuestionUpdatedInput(
@@ -746,6 +778,86 @@ describe("normalizeClaudeAskUserQuestionUpdatedInput", () => {
       await session.close();
     }
   });
+
+  test("respondToPermission maps other answer text back to Claude question keys", async () => {
+    const client = new ClaudeAgentClient({
+      logger: createTestLogger(),
+      resolveBinary: async () => "/test/claude/bin",
+    });
+    const session = await client.createSession({
+      provider: "claude",
+      cwd: process.cwd(),
+    });
+
+    const request = {
+      id: "permission-question-2",
+      provider: "claude",
+      name: "AskUserQuestion",
+      kind: "question",
+      input: normalizeClaudeAskUserQuestionRequestInput("AskUserQuestion", {
+        questions: [
+          {
+            question: "Which provider should I use?",
+            header: "Provider",
+            options: [
+              { label: "Claude", description: "Use Claude Code" },
+              { label: "Codex", description: "Use Codex" },
+            ],
+            multiSelect: false,
+          },
+        ],
+      }),
+    };
+
+    const resultPromise = new Promise<unknown>((resolve, reject) => {
+      (
+        session as unknown as {
+          pendingPermissions: Map<
+            string,
+            {
+              request: typeof request;
+              resolve: (value: unknown) => void;
+              reject: (error: Error) => void;
+            }
+          >;
+        }
+      ).pendingPermissions.set(request.id, {
+        request,
+        resolve,
+        reject,
+      });
+    });
+
+    try {
+      await session.respondToPermission(request.id, {
+        behavior: "allow",
+        updatedInput: {
+          answers: { Provider: "Use both" },
+        },
+      });
+
+      await expect(resultPromise).resolves.toEqual({
+        behavior: "allow",
+        updatedInput: {
+          questions: [
+            {
+              question: "Which provider should I use?",
+              header: "Provider",
+              options: [
+                { label: "Claude", description: "Use Claude Code" },
+                { label: "Codex", description: "Use Codex" },
+              ],
+              multiSelect: false,
+            },
+          ],
+          answers: { "Which provider should I use?": "Use both" },
+        },
+        updatedPermissions: undefined,
+      });
+    } finally {
+      await session.close();
+    }
+  });
 });
 
 describe("ClaudeAgentSession context window usage", () => {
diff --git a/packages/server/src/server/agent/providers/claude/agent.ts b/packages/server/src/server/agent/providers/claude/agent.ts
index 2af3d6893..2b9203948 100644
--- a/packages/server/src/server/agent/providers/claude/agent.ts
+++ b/packages/server/src/server/agent/providers/claude/agent.ts
@@ -104,6 +104,49 @@ function readNonEmptyString(value: unknown): string | null {
   return typeof value === "string" && value.trim().length > 0 ? value : null;
 }
 
+export function normalizeClaudeAskUserQuestionRequestInput(
+  toolName: string,
+  input: AgentMetadata,
+): AgentMetadata {
+  if (toolName !== "AskUserQuestion" || !Array.isArray(input.questions)) {
+    return input;
+  }
+
+  // Claude Code's AskUserQuestion schema says "Other" is host-provided, not a
+  // model-supplied option. Paseo's shared question UI uses allowOther for that
+  // freeform answer path.
+  return {
+    ...input,
+    questions: input.questions.map((item) => {
+      if (!isMetadata(item)) {
+        return item;
+      }
+      return {
+        ...item,
+        allowOther: true,
+      };
+    }),
+  };
+}
+
+function stripClaudeAskUserQuestionUiMetadata(input: AgentMetadata): AgentMetadata {
+  if (!Array.isArray(input.questions)) {
+    return input;
+  }
+
+  return {
+    ...input,
+    questions: input.questions.map((item) => {
+      if (!isMetadata(item) || !("allowOther" in item)) {
+        return item;
+      }
+      const itemForClaude: AgentMetadata = { ...item };
+      delete itemForClaude.allowOther;
+      return itemForClaude;
+    }),
+  };
+}
+
 export function normalizeClaudeAskUserQuestionUpdatedInput(
   updatedInput: AgentMetadata | undefined,
   fallbackInput: AgentMetadata | undefined,
@@ -114,7 +157,7 @@ export function normalizeClaudeAskUserQuestionUpdatedInput(
   // AskUserQuestion tool expects answer keys to match the full question text. Merge
   // the original request payload back in so provider callbacks that only return
   // `{ answers }` still satisfy Claude's full tool input schema.
-  const merged = { ...fallback, ...base };
+  const merged = stripClaudeAskUserQuestionUiMetadata({ ...fallback, ...base });
   const questions =
     (Array.isArray(base.questions) ? base.questions : null) ??
     (Array.isArray(fallback.questions) ? fallback.questions : null);
@@ -3760,6 +3803,7 @@ class ClaudeAgentSession implements AgentSession {
   ): Promise<PermissionResult> => {
     const requestId = `permission-${randomUUID()}`;
     const kind = resolvePermissionKind(toolName, input);
+    const requestInput = normalizeClaudeAskUserQuestionRequestInput(toolName, input);
     const metadata: AgentMetadata = {};
     if (options.toolUseID) {
       metadata.toolUseId = options.toolUseID;
@@ -3782,7 +3826,7 @@ class ClaudeAgentSession implements AgentSession {
       provider: "claude",
       name: toolName,
       kind,
-      input,
+      input: requestInput,
       detail: toolDetail,
       suggestions: options.suggestions?.map((suggestion) => ({
         ...suggestion,