v2.5.0-task-model: lightweight task model services + tasks table

Task model infrastructure for cheap LLM calls (auto-naming, search
rewrite, tags, summaries) via a dedicated llama-server instance at
TASK_MODEL_URL, falling back to LLAMA_SWAP_URL with FAST_MODEL when
unset. Replaces the inline fetch in auto_name.ts with taskModelCompletion.

Adds search query rewriting: on step 0 when web tools are enabled, the
user's message is summarized into a search intent hint appended to the
system prompt, improving web_search relevance.

Schema: tasks table for provider dispatch and arena, sessions.tags column.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

apps/server/src/services/auto_name.ts

@@ -1,9 +1,10 @@
 import type { InferenceContext } from './inference/index.js';
+import { taskModelCompletion } from './task-model.js';
 
 const NAMING_SYSTEM_PROMPT =
-  'You name chat sessions based on what the assistant did. Summarize the topic or outcome — do NOT copy the first few words verbatim. Reply directly with no thinking, reasoning, or explanation. Output ONLY the title, 4 words max, no quotes, no punctuation, no prefix like "Title:".';
+  'You name chat sessions. Reply with ONLY the title. 4 to 6 words. No quotes, no punctuation, no prefix.';
 
-const MAX_TITLE_CHARS = 60;
+const MAX_TITLE_CHARS = 80;
 
 function cleanTitle(raw: string): string {
   let name = raw.trim();
@@ -18,27 +19,7 @@ function cleanTitle(raw: string): string {
   return name;
 }
 
-interface NamingResponse {
-  choices?: Array<{
-    message?: {
-      content?: string;
-      reasoning_content?: string;
-    };
-  }>;
-}
-
-function pickTitleSource(data: NamingResponse): string {
-  const choice = data.choices?.[0]?.message;
-  if (!choice) return '';
-  if (choice.content && choice.content.trim().length > 0) return choice.content;
-  const reasoning = choice.reasoning_content ?? '';
-  if (reasoning.length === 0) return '';
-  const lines = reasoning
-    .split('\n')
-    .map((l) => l.trim())
-    .filter((l) => l.length > 0);
-  return lines[lines.length - 1] ?? '';
-}
+// TODO: wire suggestTags after task model validation
 
 export async function maybeAutoNameChat(
   ctx: InferenceContext,
@@ -64,13 +45,6 @@ export async function maybeAutoNameChat(
   if (!chat) return;
   if (chat.name !== null && chat.name !== '') return;
 
-  const sessionRows = await ctx.sql<{ model: string }[]>`
-    SELECT model FROM sessions WHERE id = ${sessionId}
-  `;
-  // v2.0.5: prefer FAST_MODEL for cheap LLM calls (titles, summaries).
-  const model = ctx.config.FAST_MODEL ?? sessionRows[0]?.model;
-  if (!model) return;
-
   const assistantMsg = await ctx.sql<{ content: string }[]>`
     SELECT content FROM messages
     WHERE chat_id = ${chatId}
@@ -84,32 +58,12 @@ export async function maybeAutoNameChat(
 
   const assistantText = assistantMsg[0].content.slice(0, 2000);
 
-  const body = {
-    model,
-    messages: [
-      { role: 'system', content: NAMING_SYSTEM_PROMPT },
-      {
-        role: 'user',
-        content: assistantText,
-      },
-    ],
-    max_tokens: 30,
+  const raw = await taskModelCompletion({
+    system: NAMING_SYSTEM_PROMPT,
+    user: assistantText,
+    maxTokens: 30,
     temperature: 0.3,
-    stream: false,
-    chat_template_kwargs: { enable_thinking: false },
-  };
-
-  const res = await fetch(`${ctx.config.LLAMA_SWAP_URL}/v1/chat/completions`, {
-    method: 'POST',
-    headers: { 'Content-Type': 'application/json' },
-    body: JSON.stringify(body),
   });
-  if (!res.ok) {
-    const text = await res.text().catch(() => '');
-    throw new Error(`naming request failed: ${res.status} ${text.slice(0, 200)}`);
-  }
-  const data = (await res.json()) as NamingResponse;
-  const raw = pickTitleSource(data);
   const name = cleanTitle(raw);
   if (!name) {
     ctx.log.warn({ chatId, raw }, 'auto-name: empty title from model');