diff --git a/apps/server/src/schema.sql b/apps/server/src/schema.sql
index f8274d2..68ee801 100644
--- a/apps/server/src/schema.sql
+++ b/apps/server/src/schema.sql
@@ -120,6 +120,19 @@ BEGIN
   END IF;
 END $$;
 
+-- v1.12.1: drop stale inline CHECK constraints that were superseded by the
+-- named *_chk variants above. messages_status_check missed 'cancelled' and
+-- messages_role_check missed 'system' — both narrower than what's in use.
+DO $$
+BEGIN
+  IF EXISTS (SELECT 1 FROM pg_constraint WHERE conname = 'messages_status_check') THEN
+    ALTER TABLE messages DROP CONSTRAINT messages_status_check;
+  END IF;
+  IF EXISTS (SELECT 1 FROM pg_constraint WHERE conname = 'messages_role_check') THEN
+    ALTER TABLE messages DROP CONSTRAINT messages_role_check;
+  END IF;
+END $$;
+
 -- v1.2-project-ux: projects.status + projects.gitea_remote
 -- KEEP IN SYNC: apps/server/src/types/api.ts PROJECT_STATUSES
 ALTER TABLE projects ADD COLUMN IF NOT EXISTS status TEXT NOT NULL DEFAULT 'open';
diff --git a/apps/server/src/services/inference.ts b/apps/server/src/services/inference.ts
index 9e57595..27dd24b 100644
--- a/apps/server/src/services/inference.ts
+++ b/apps/server/src/services/inference.ts
@@ -62,7 +62,6 @@ const CAP_HIT_SUMMARY_NOTE = (limit: number) =>
 // session/processor.ts). Threshold of 3 is the smallest value that doesn't
 // false-positive on a model that retries once after a transient error.
 export const DOOM_LOOP_THRESHOLD = 3;
-export const DOOM_LOOP_SAME_NAME_THRESHOLD = 5;
 
 const DOOM_LOOP_NOTE = (name: string) =>
   `You called ${name} with the same arguments ${DOOM_LOOP_THRESHOLD} times in a row. Stop calling it. Produce the best answer you can with what you have.`;
@@ -86,18 +85,6 @@ export function detectDoomLoop(
   return { name: ref.name, args: ref.args };
 }
 
-export function detectSameNameLoop(
-  recentToolCalls: ToolCall[],
-): { name: string } | null {
-  if (recentToolCalls.length < DOOM_LOOP_SAME_NAME_THRESHOLD) return null;
-  const last = recentToolCalls.slice(-DOOM_LOOP_SAME_NAME_THRESHOLD);
-  const name = last[0]!.name;
-  for (let i = 1; i < last.length; i++) {
-    if (last[i]!.name !== name) return null;
-  }
-  return { name };
-}
-
 function isCapHitSentinel(m: Message): boolean {
   return (
     m.role === 'system' &&
@@ -814,6 +801,17 @@ async function handleAbortOrError(
   // genuine errors flip the dot red. v1.8.2: error path also carries a
   // machine-readable `reason` so the UI can render specifics inline.
   if (isAbort) {
+    // v1.12.1: defensive cancellation write. The status=${finalStatus} UPDATE
+    // above already sets 'cancelled' for the AbortError case, but a row can
+    // leak as 'streaming' when the abort fires between the post-tool-phase
+    // INSERT (executeToolPhase) and the next runAssistantTurn's stream setup,
+    // bypassing the try/catch around executeStreamPhase. The status guard
+    // makes this a no-op when the earlier write already landed.
+    await ctx.sql`
+      UPDATE messages
+      SET status = 'cancelled', content = ${accumulated}, finished_at = clock_timestamp()
+      WHERE id = ${args.assistantMessageId} AND status = 'streaming'
+    `;
     ctx.publishUser({ type: 'chat_status', chat_id: chatId, status: 'idle', at: new Date().toISOString() });
     ctx.publish(sessionId, {
       type: 'message_complete',
@@ -907,6 +905,7 @@ async function executeToolPhase(
   // 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.
+  ctx.publishUser({ type: 'chat_status', chat_id: chatId, status: 'tool_running', at: new Date().toISOString() });
   let pausingForUserInput = false;
   await Promise.all(
     toolCalls.map(async (tc) => {
@@ -951,13 +950,10 @@ 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',
+      status: 'waiting_for_input',
       at: new Date().toISOString(),
     });
     ctx.log.info(
@@ -972,7 +968,6 @@ async function executeToolPhase(
     VALUES (${sessionId}, ${chatId}, 'assistant', '', 'streaming', clock_timestamp())
     RETURNING id
   `;
-  ctx.log.info({ chatId, nextToolsUsed: toolsUsed + result.toolCalls.length, phase: 'executeToolPhase:before_recurse' }, 'recursing into next turn');
   await runAssistantTurn(ctx, {
     sessionId,
     chatId,
@@ -1057,7 +1052,6 @@ async function runAssistantTurn(
   args: TurnArgs,
 ): Promise<void> {
   const { sessionId, chatId } = args;
-  ctx.log.info({ chatId, sessionId, toolsUsed: args.toolsUsed, recentToolCallsLen: args.recentToolCalls?.length ?? 0, phase: 'runAssistantTurn:enter' }, 'turn enter');
 
   // v1.11: if the prior turn flagged this chat for compaction, run it first
   // so loadContext below reads the post-compaction history. We swallow
@@ -1088,7 +1082,6 @@ async function runAssistantTurn(
     return;
   }
   const { session, project, history } = loaded;
-  ctx.log.info({ chatId, historyLen: history.length, phase: 'runAssistantTurn:loaded' }, 'context loaded');
   const projectRoot = await resolveProjectRoot(project.path);
   // Agent resolution is per-turn so PATCH agent_id mid-conversation takes
   // effect on the next message. Unknown agent_id returns null silently —
@@ -1106,7 +1099,6 @@ async function runAssistantTurn(
     await runCapHitSummary(ctx, args, session, project, history, agent, budget);
     return;
   }
-  ctx.log.info({ chatId, budget, toolsUsed: args.toolsUsed, phase: 'runAssistantTurn:budget_ok' }, 'budget ok');
 
   // v1.11.6: doom-loop guard. Detected BEFORE the budget cap (the model can
   // burn through 3 identical calls long before the 15-call budget fires).
@@ -1118,17 +1110,6 @@ async function runAssistantTurn(
     await runDoomLoopSummary(ctx, args, session, project, history, agent, loop);
     return;
   }
-  ctx.log.info({ chatId, phase: 'runAssistantTurn:no_doom_loop' }, 'no doom loop');
-
-  const sameNameLoop = detectSameNameLoop(args.recentToolCalls);
-  if (sameNameLoop) {
-    await runDoomLoopSummary(ctx, args, session, project, history, agent, {
-      name: sameNameLoop.name,
-      args: {},
-    });
-    return;
-  }
-  ctx.log.info({ chatId, phase: 'runAssistantTurn:no_same_name_loop' }, 'no same-name loop');
 
   const messages = await buildMessagesPayload(session, project, history, agent);
 
@@ -1142,24 +1123,17 @@ async function runAssistantTurn(
   const webToolsEnabled =
     session.web_search_enabled ?? project.default_web_search_enabled ?? false;
 
-  ctx.log.info({ chatId, msgCount: messages.length, phase: 'runAssistantTurn:payload_built' }, 'payload built');
-
   const state: StreamPhaseState = { accumulated: '', startedAt: null };
   let result: StreamResult;
   try {
-    ctx.log.info({ chatId, model: session.model, phase: 'runAssistantTurn:before_stream' }, 'calling upstream');
     result = await executeStreamPhase(ctx, args, session, messages, state, agent, webToolsEnabled);
   } catch (err) {
     await handleAbortOrError(ctx, args, state.accumulated, err);
     return;
   }
 
-  ctx.log.info({ chatId, toolCallsLen: result.toolCalls.length, finishReason: result.finishReason, contentLen: result.content?.length ?? 0, phase: 'runAssistantTurn:after_stream' }, 'upstream returned');
-
   if (result.toolCalls.length > 0) {
-    ctx.log.info({ chatId, toolNames: result.toolCalls.map(tc => tc.name), phase: 'runAssistantTurn:before_tools' }, 'executing tools');
     await executeToolPhase(ctx, args, result, state.startedAt, session, projectRoot);
-    ctx.log.info({ chatId, phase: 'runAssistantTurn:after_tools' }, 'tools complete, returning');
     return;
   }
 
@@ -1465,7 +1439,6 @@ async function runDoomLoopSummary(
   loop: { name: string; args: Record<string, unknown> },
 ): Promise<void> {
   const { sessionId, chatId, assistantMessageId, signal } = args;
-  ctx.log.info({ chatId, loopName: loop.name, phase: 'runDoomLoopSummary:enter' }, 'doom-loop summary firing');
 
   const messages = await buildMessagesPayload(session, project, history, agent);
   messages.push({ role: 'system', content: DOOM_LOOP_NOTE(loop.name) });
@@ -1713,7 +1686,7 @@ export function createInferenceRunner(
       };
       // v1.8 mobile-tabs: announce working before the async loop starts so
       // every device subscribed to the user channel sees the amber dot.
-      callCtx.publishUser({ type: 'chat_status', chat_id: chatId, status: 'working', at: new Date().toISOString() });
+      callCtx.publishUser({ type: 'chat_status', chat_id: chatId, status: 'streaming', at: new Date().toISOString() });
       const controller = new AbortController();
       let resolveCompleted!: () => void;
       const completed = new Promise<void>((res) => { resolveCompleted = res; });