fix(FN-7210): stop merge-retry loop starving executor remediation pass

recoverCompletedTask refused workflow-graph re-entry when the live task
has incomplete plan steps or a remediation bounce is already scheduled.
A pre-merge optional/advisory REVISE reopens steps and schedules a
sendTaskBackForFix bounce; a competing graph re-entry re-passed the
advisory step (budget exhausted), advanced to merge, and looped forever
on the 'task has incomplete steps' gate. Defer to the bounce /
stale-incomplete-review recovery so the executor finishes the steps.

Fusion-Task-Id: FN-7210

.changeset/fn-7210-recover-completed-incomplete-steps-guard.md

@@ -0,0 +1,7 @@
+---
+"@runfusion/fusion": patch
+---
+
+summary: Fix tasks stuck in review after a code-review revision by stopping the merge-retry loop from starving the executor's fix pass.
+category: fix
+dev: recoverCompletedTask now refuses workflow-graph re-entry when the live task has incomplete steps or a remediation bounce (sendTaskBackForFix → scheduleWorkflowRerun) is already scheduled, so a pre-merge optional/advisory REVISE that reopens plan steps lets the executor finish them instead of re-passing the advisory step (budget exhausted) and looping on the "task has incomplete steps" merge gate. Regression: restart.integration.test.ts.

packages/engine/src/__tests__/restart.integration.test.ts

@@ -869,6 +869,68 @@ describe("In-progress task resume after restart", () => {
     // Must NOT silently finalize to review as a success.
     expect(store.moveTask).not.toHaveBeenCalledWith("FN-963", "in-review");
   });
+
+  // FNXC:WorkflowOptionalStepFix 2026-06-28-12:00 (FN-7210 regression):
+  // A pre-merge optional/advisory REVISE reopens plan steps to `pending` and
+  // schedules a remediation bounce so the executor can finish them. recoverCompletedTask
+  // must NOT re-enter the workflow graph (and thus the merge node) while the live task
+  // still has incomplete steps — otherwise the re-run re-passes the advisory step
+  // (fix budget exhausted), advances to merge, and the merge gate refuses forever with
+  // "task has incomplete steps".
+  it("recoverCompletedTask() refuses graph re-entry when the live task has incomplete steps", async () => {
+    const store = createMockStore({
+      getTask: vi.fn().mockResolvedValue(makeTaskDetail("FN-7210", "in-review", {
+        worktree: "/tmp/wt/FN-7210",
+        steps: makeSteps("done", "done", "pending"),
+      })),
+    });
+    const task = makeTask("FN-7210", "in-review", {
+      worktree: "/tmp/wt/FN-7210",
+      steps: makeSteps("done", "done", "pending"),
+    });
+
+    const executor = new TaskExecutor(store, "/tmp/test");
+    vi.spyOn(executor as any, "captureModifiedFiles").mockResolvedValue([]);
+    const graphEntry = vi
+      .spyOn(executor as any, "maybeExecuteWorkflowGraph")
+      .mockResolvedValue(true);
+
+    const recovered = await executor.recoverCompletedTask(task);
+
+    expect(recovered).toBe(false);
+    expect(graphEntry).not.toHaveBeenCalled();
+  });
+
+  // FNXC:WorkflowOptionalStepFix 2026-06-28-12:00 (FN-7210 regression):
+  // While a remediation bounce is scheduled (scheduleWorkflowRerun registers a
+  // rerun watchdog synchronously), recoverCompletedTask must yield to it rather than
+  // racing a competing graph re-entry to the merge node.
+  it("recoverCompletedTask() yields to a scheduled workflow remediation bounce", async () => {
+    const store = createMockStore({
+      getTask: vi.fn().mockResolvedValue(makeTaskDetail("FN-7211", "in-review", {
+        worktree: "/tmp/wt/FN-7211",
+        steps: makeSteps("done"),
+      })),
+    });
+    const task = makeTask("FN-7211", "in-review", {
+      worktree: "/tmp/wt/FN-7211",
+      steps: makeSteps("done"),
+    });
+
+    const executor = new TaskExecutor(store, "/tmp/test");
+    vi.spyOn(executor as any, "captureModifiedFiles").mockResolvedValue([]);
+    const graphEntry = vi
+      .spyOn(executor as any, "maybeExecuteWorkflowGraph")
+      .mockResolvedValue(true);
+    // Simulate a bounce already scheduled for this task.
+    (executor as any).workflowRerunWatchdogs.set("FN-7211", setTimeout(() => {}, 0));
+
+    const recovered = await executor.recoverCompletedTask(task);
+
+    expect(recovered).toBe(false);
+    expect(graphEntry).not.toHaveBeenCalled();
+    clearTimeout((executor as any).workflowRerunWatchdogs.get("FN-7211"));
+  });
 });
 
 // ── Step 3: In-review merge re-queue tests ────────────────────────────────

packages/engine/src/executor.ts

@@ -3720,6 +3720,36 @@ export class TaskExecutor {
         executorLog.log(`${task.id}: skipping recoverCompletedTask — task has active execution in flight`);
         return false;
       }
+
+      /*
+      FNXC:WorkflowOptionalStepFix 2026-06-28-12:00:
+      A pre-merge optional/advisory step REVISE (Code Review / Browser Verification) reopens
+      plan steps to `pending` and schedules a remediation bounce (sendTaskBackForFix →
+      scheduleWorkflowRerun) that moves the task in-review → todo → in-progress so the executor
+      can finish the reopened steps. Re-entering the workflow graph here while that bounce is
+      still scheduled — or while the live task already carries incomplete plan steps — preempts
+      the executor's single fix cycle: the re-run re-passes the advisory step (its fix budget is
+      now exhausted), advances to the `merge` node, and the merge gate refuses with
+      "task has incomplete steps" forever (observed on FN-7210; the FN-7122 bounce fix handled
+      the column race but not this competing graph re-entry). recoverCompletedTask only owns
+      tasks whose work is genuinely COMPLETE, so refuse re-entry when a remediation bounce is in
+      flight or the live task has non-terminal steps, and let the bounce / stale-incomplete-review
+      recovery re-launch execution instead.
+      */
+      if (this.workflowRerunWatchdogs.has(task.id) || this.workflowRerunPending.has(task.id)) {
+        executorLog.log(`${task.id}: skipping recoverCompletedTask — workflow remediation bounce already scheduled`);
+        return false;
+      }
+      const liveForCompletenessCheck = await this.store.getTask(task.id).catch(() => task);
+      if (
+        liveForCompletenessCheck
+        && (liveForCompletenessCheck.steps?.length ?? 0) > 0
+        && !this.isTaskWorkComplete(liveForCompletenessCheck)
+      ) {
+        executorLog.log(`${task.id}: skipping recoverCompletedTask — task has incomplete steps awaiting executor remediation`);
+        return false;
+      }
+
       const settings = await this.store.getSettings();
       if (settings.globalPause || settings.enginePaused) {
         executorLog.log(