Skip Not-Started tasks that are already assigned to someone

A task carrying an "Assigned to" avatar in the Planner grid has been picked
up by a person, so its attachments must not be downloaded and its status must
not be moved to In Progress. Until now the only filter was percentComplete
== 0, so an assigned task with an attachment was downloaded like any other —
on the Balboa Nursing and Rehab plan that was one of the two "Facesheets
Needed (GG)" tasks assigned to Dr. Mark Lukasik and Adina Smets.

- _not_started_tasks() now selects `assignments` as well.
- _task_assignee_ids()/_resolve_user_names(): the assignee check runs before
  the task-details call, so an assigned task costs no extra Graph request;
  display names are resolved once per user and cached for the run, falling
  back to the raw ID when the lookup fails.
- Summary workbook gains a "Tasks Skipped (Assigned)" column and a second
  "Tasks" sheet with one row per task — facility, task name, assigned to,
  attachment count, Downloaded Yes/No and the reason — so a skip can be
  explained without reading the log.
- New "All assigned — skipped" facility status, distinct from "No attachments".

scripts/probe_assignments.py is the read-only probe used to confirm the rule
against the live plan before changing behaviour.

Verified on Balboa Nursing and Rehab: 4 Not Started -> 8.24.26 and 8.21.26
downloaded (unassigned), both Facesheets Needed (GG) skipped. 110 tests pass
(tests/test_uipath_pilot.py and tests/test_config.py fail on this checkout for
pre-existing, unrelated reasons).

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

tests/test_assigned_task_skip.py

@@ -0,0 +1,121 @@
+"""Only unassigned Not-Started tasks may be downloaded.
+
+A task carrying an 'Assigned to' avatar in the Planner grid has been picked up
+by a person; the bot must leave it — and its status — alone.
+"""
+from __future__ import annotations
+
+import sys
+from pathlib import Path
+
+import pandas as pd
+
+ROOT = Path(__file__).resolve().parent.parent
+SRC = ROOT / "src"
+for _p in (str(ROOT), str(SRC)):
+    if _p not in sys.path:
+        sys.path.insert(0, _p)
+
+from onedrive_planner_http import _task_assignee_ids, _write_summary  # noqa: E402
+
+
+class TestTaskAssigneeIds:
+    def test_unassigned_task_has_no_assignees(self):
+        assert _task_assignee_ids({"title": "8.24.26"}) == []
+        assert _task_assignee_ids({"title": "8.21.26", "assignments": {}}) == []
+        assert _task_assignee_ids({"assignments": None}) == []
+
+    def test_assigned_task_returns_every_user_id(self):
+        task = {
+            "title": "Facesheets Needed (GG)",
+            "assignments": {
+                "aaaaaaaa-1111-2222-3333-444444444444": {"orderHint": "8"},
+                "bbbbbbbb-5555-6666-7777-888888888888": {"orderHint": "9"},
+            },
+        }
+        assert sorted(_task_assignee_ids(task)) == [
+            "aaaaaaaa-1111-2222-3333-444444444444",
+            "bbbbbbbb-5555-6666-7777-888888888888",
+        ]
+
+    def test_blank_user_id_is_ignored(self):
+        assert _task_assignee_ids({"assignments": {"": {}}}) == []
+
+
+class TestSummaryWorkbook:
+    def _write(self, tmp_path: Path, payload: dict) -> Path:
+        out = tmp_path / "Download_Summary_HTTP_test.xlsx"
+        _write_summary({"Balboa Nursing and Rehab": payload}, [], out, "2026-08-24 18:00")
+        return out
+
+    def _balboa_payload(self) -> dict:
+        """The real Balboa plan: 2 unassigned with files, 2 assigned."""
+        return {
+            "original": [Path("a.pdf"), Path("b.pdf")],
+            "tasks_total": 4,
+            "tasks_with_attachments": 2,
+            "tasks_with_downloads": 2,
+            "tasks_skipped_no_attachment": 0,
+            "tasks_skipped_assigned": 2,
+            "status_changed": 2,
+            "task_rows": [
+                {"Facility Name": "Balboa Nursing and Rehab", "Task Name": "8.24.26",
+                 "Assigned To": "", "Attachments": 1, "Downloaded": "Yes",
+                 "Reason": "Downloaded 1/1 attachment(s)"},
+                {"Facility Name": "Balboa Nursing and Rehab", "Task Name": "8.21.26",
+                 "Assigned To": "", "Attachments": 1, "Downloaded": "Yes",
+                 "Reason": "Downloaded 1/1 attachment(s)"},
+                {"Facility Name": "Balboa Nursing and Rehab",
+                 "Task Name": "Facesheets Needed (GG)", "Assigned To": "A S",
+                 "Attachments": "", "Downloaded": "No",
+                 "Reason": "Skipped — assigned to A S"},
+                {"Facility Name": "Balboa Nursing and Rehab",
+                 "Task Name": "Facesheets Needed (GG)", "Assigned To": "D M",
+                 "Attachments": "", "Downloaded": "No",
+                 "Reason": "Skipped — assigned to D M"},
+            ],
+        }
+
+    def test_download_sheet_reports_assigned_skips(self, tmp_path):
+        out = self._write(tmp_path, self._balboa_payload())
+        df = pd.read_excel(out, sheet_name="Download")
+        assert df.loc[0, "Tasks Skipped (Assigned)"] == 2
+        assert df.loc[0, "Files Downloaded"] == 2
+        assert df.loc[0, "Status"] == "Success"
+
+    def test_tasks_sheet_has_a_yes_no_row_per_task(self, tmp_path):
+        out = self._write(tmp_path, self._balboa_payload())
+        df = pd.read_excel(out, sheet_name="Tasks")
+        assert list(df.columns) == [
+            "Facility Name", "Task Name", "Assigned To",
+            "Attachments", "Downloaded", "Reason",
+        ]
+        assert list(df["Downloaded"]) == ["Yes", "Yes", "No", "No"]
+        facesheets = df[df["Task Name"] == "Facesheets Needed (GG)"]
+        assert len(facesheets) == 2
+        assert all(facesheets["Downloaded"] == "No")
+        assert all("assigned to" in r for r in facesheets["Reason"])
+
+    def test_facility_whose_tasks_are_all_assigned(self, tmp_path):
+        out = self._write(tmp_path, {
+            "original": [],
+            "tasks_total": 3,
+            "tasks_with_attachments": 0,
+            "tasks_with_downloads": 0,
+            "tasks_skipped_no_attachment": 0,
+            "tasks_skipped_assigned": 3,
+            "status_changed": 0,
+            "task_rows": [],
+        })
+        df = pd.read_excel(out, sheet_name="Download")
+        assert df.loc[0, "Status"] == "All assigned — skipped"
+
+    def test_tasks_sheet_exists_even_with_no_task_rows(self, tmp_path):
+        out = self._write(tmp_path, {
+            "original": [], "tasks_total": 0, "tasks_with_attachments": 0,
+            "tasks_with_downloads": 0, "tasks_skipped_no_attachment": 0,
+            "tasks_skipped_assigned": 0, "status_changed": 0, "task_rows": [],
+        })
+        df = pd.read_excel(out, sheet_name="Tasks")
+        assert df.empty
+        assert "Downloaded" in df.columns