v1.9.7: ask_user_input elicitation tool

apps/server/src/services/inference.ts

@@ -665,6 +665,12 @@ async function executeToolPhase(
     model: session.model,
   });
 
+  // Batch 9.7: ask_user_input pauses the loop. The tool row is still inserted
+  // (the answer endpoint needs a target row to UPDATE), but tool_results is
+  // pre-stamped with output=null as a "pending" sentinel and no tool_result
+  // frame goes out — the card renders from the tool_call frame alone. Mixed
+  // batches still execute the other tools normally.
+  let pausingForUserInput = false;
   await Promise.all(
     toolCalls.map(async (tc) => {
       const [toolRow] = await ctx.sql<{ id: string }[]>`
@@ -673,6 +679,16 @@ async function executeToolPhase(
         RETURNING id
       `;
       const toolMessageId = toolRow!.id;
+      if (tc.name === 'ask_user_input') {
+        pausingForUserInput = true;
+        const sentinel = { tool_call_id: tc.id, output: null, truncated: false };
+        await ctx.sql`
+          UPDATE messages
+          SET tool_results = ${ctx.sql.json(sentinel as never)}
+          WHERE id = ${toolMessageId}
+        `;
+        return;
+      }
       const tres = await executeToolCall(projectRoot, tc);
       const stored = {
         tool_call_id: tc.id,
@@ -697,6 +713,23 @@ async function executeToolPhase(
     })
   );
 
+  if (pausingForUserInput) {
+    // Drop the dot back to idle — the card is the actionable surface now.
+    // The next inference turn fires from POST /api/chats/:id/answer_user_input
+    // once the user submits their answers.
+    ctx.publishUser({
+      type: 'chat_status',
+      chat_id: chatId,
+      status: 'idle',
+      at: new Date().toISOString(),
+    });
+    ctx.log.info(
+      { sessionId, chatId, assistantMessageId },
+      'inference paused awaiting user input',
+    );
+    return;
+  }
+
   const [nextAssistant] = await ctx.sql<{ id: string }[]>`
     INSERT INTO messages (session_id, chat_id, role, content, status, created_at)
     VALUES (${sessionId}, ${chatId}, 'assistant', '', 'streaming', clock_timestamp())