worker: auto-resume falls back to fresh goal for never-started tasks

2026-10-29 live test: task A was interrupted 61s after claim but had
never written anything (no host-state entry, no worktree, no trace,
no checkpoint); auto-resume ran goal-resume, which host rightly
refused, and the task failed. Refusal only makes sense for tasks
with progress but a lost breakpoint (those go through the P0
checkpoint rebuild); for a never-started task the only meaningful
outcome is a fresh run, identical to its first attempt (same task
id, same goal, worktree does not exist yet so nothing is
overwritten).

Change (claim path only, via _resolve_breakpoint_resume):
- rebuild from checkpoint first, unchanged;
- if rebuild fails AND _task_never_started holds (state entry,
  checkpoint row, worktree dir, trace file all absent), clear the
  breakpoint flag so the run takes the fresh goal path, with a
  worker_fresh_restart audit event;
- any evidence of prior execution, or any check failure,
  keeps the old refusal path (fail-closed); the auto-resume
  count cap and no-progress stall guard still bound retries.

Tests: cloud/tests/test_neverstart_fresh.py, 10 items (fallback
holds, entry/checkpoint/worktree/trace presence each block it,
corrupt state blocks it, rebuild still wins when a checkpoint
exists); test_goal_resume's breakpoint-routing test now sets up
a task with a checkpoint so it still exercises goal-resume.
Full suites: cloud 1178 (skipped 4), host 104, deploy 39, green.

cloud/worker.py

@@ -1245,6 +1245,78 @@ class Worker:
         except Exception:
             return False
 
+    def _task_never_started(self, task):
+        """从未启动判据（2026-10-29 实测任务 A 的教训）。
+
+        四者全无才算从未启动：host-state 里无本任务条目、
+        库里无断点行、state 同目录 host-worktrees 下无本任务
+        worktree、host-traces 下无本任务链路文件。任一存在
+        都说明执行过，绝不能当成全新任务重跑（会撞已存在
+        或丢进度），调用方保持原有拒绝/重建路径。
+        任何一步查不出来（state 路径不明、文件损坏、库异常）
+        都返回 False——fail-closed，宁可拒绝也不误重跑。
+        """
+        try:
+            state_path = self._host_state_path_for_task(task)
+            if not state_path:
+                return False
+            host_name = _worker_host_name(self.executor, task)
+            if os.path.exists(state_path):
+                with open(state_path, encoding="utf-8") as fh:
+                    state = json.load(fh)
+                if host_name in (state.get("tasks") or {}):
+                    return False
+            conn = _db_connect(self.db_path)
+            try:
+                self._ensure_result_columns(conn)
+                row = conn.execute(
+                    "SELECT 1 FROM task_checkpoints WHERE task_id = ?"
+                    " LIMIT 1", (task["id"],)).fetchone()
+                if row:
+                    return False
+            finally:
+                conn.close()
+            state_dir = os.path.dirname(os.path.abspath(state_path))
+            if os.path.isdir(os.path.join(state_dir, "host-worktrees",
+                                          host_name)):
+                return False
+            if os.path.exists(os.path.join(state_dir, "host-traces",
+                                            host_name + ".jsonl")):
+                return False
+            return True
+        except Exception:
+            return False
+
+    def _resolve_breakpoint_resume(self, task):
+        """断点续跑解析：先重建，建不出且从未启动时回落全新跑。
+
+        正常情况（条目在，或库里有断点可重建）原样返回，
+        调用方走 goal-resume 断点续跑。只有重建失败（库里
+        也无断点）且 _task_never_started 全无执行痕迹时，
+        才清掉断点标记、改走全新 goal：此时 goal-resume 注定
+        被 host 拒绝 → failed，而全新跑与首轮完全相同
+        （worktree 本就不存在，不会覆盖任何进度）。
+        次数上限与无进展熔断仍在上游卡着，不会空转。
+        手动 /goal-resume 同样走这里：对从未跑过的任务，
+        全新跑也是唯一有意义的结果。
+        """
+        try:
+            rebuilt = self._rebuild_host_state_from_checkpoint(task)
+        except Exception:
+            rebuilt = False
+        if not rebuilt and self._task_never_started(task):
+            task = dict(task)
+            task["_goal_resume_breakpoint"] = False
+            self._audit(task.get("project_id"),
+                        "worker_fresh_restart", task["id"])
+            try:
+                log_event(self.logger, "worker_fresh_restart",
+                          task_id=task["id"],
+                          project_id=task.get("project_id"))
+            except Exception:
+                pass
+        return task
+
     def _ensure_result_columns(self, conn):
         """worker 侧的迁移兜底：与 server.migrate_tasks_table 同列。
 
@@ -4236,12 +4308,12 @@ class Worker:
             task["_goal_resume_breakpoint"] = True
             # P0 三件其一：断点丢失重建——host-state 里条目已丢
             # 但库里有断点时，先按库断点拼最小条目写回，再由
-            # 正常 goal-resume 续跑；库里也没有断点时返回 False，
-            # 正常路径继续（host 拒绝 → failed，不造假断点）。
-            try:
-                self._rebuild_host_state_from_checkpoint(task)
-            except Exception:
-                pass
+            # 正常 goal-resume 续跑；库里也没有断点时返回 False。
+            # 从未启动回落（2026-10-29 实测任务 A）：重建失败
+            # 且四者全无执行痕迹时，goal-resume 注定被 host
+            # 拒绝 → failed，改走全新 goal（见
+            # _resolve_breakpoint_resume），不造假断点。
+            task = self._resolve_breakpoint_resume(task)
             try:
                 conn = _db_connect(self.db_path)
                 try: