Add rpa-bot-handoff skill: ops guide for bot handover

The skill scans a bot with subagents (never runs it) and writes
HANDOVER.md, an 11-section ops guide in the operator's order: where is
what, how to run, how long to wait, after the run, how to see an
exception, exception list, how you will know, how to tell us, don'ts,
issues seen before. Dev material (check against 12 failure classes from
Tebra Auto Post, gaps with file:line, incidents) goes to HANDOVER-DEV.md.

Also adds ADR-0004, glossary terms, five fixture bots and a handoff eval
runner (--resume, --regrade) with results from 47 runs of the earlier
13-section version. The grader still expects that structure and must be
updated before the next eval run.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

tests/fixtures/h-packaged-desktop/app.py

@@ -0,0 +1,101 @@
+"""Ledger reconciler: compares the bank statement Excel with LedgerPro and posts missing entries.
+Shipped as LedgerBot.exe (PyInstaller) and started through launcher.exe."""
+import json
+import logging
+import os
+import subprocess
+import sys
+import time
+from datetime import datetime
+from pathlib import Path
+
+import pyautogui
+from openpyxl import Workbook, load_workbook
+
+VERSION = "3.2.0"  # shown in the window title "LedgerBot v3.2.0"
+FROZEN = getattr(sys, "frozen", False)
+BASE = Path(sys.executable).parent if FROZEN else Path(__file__).parent
+ASSETS = Path(getattr(sys, "_MEIPASS", BASE)) / "assets"
+SHORTCUT_DIRS = [Path(os.environ.get("USERPROFILE", "")) / "Desktop", Path(r"C:\Users\Public\Desktop")]
+CONFIDENCE = 0.9  # image match; buttons were captured at 100% display scaling, 1920x1080
+FIND_WAIT_S = 20
+TOLERANCE = 0.01
+
+logging.basicConfig(filename=BASE / "logs" / f"ledger_{datetime.now():%Y%m%d}.log", level=logging.INFO,
+                    format="%(asctime)s %(levelname)s %(message)s")
+log = logging.getLogger("ledger")
+
+
+def find_shortcut():
+    for d in SHORTCUT_DIRS:
+        lnk = d / "LedgerPro.lnk"
+        if lnk.exists():
+            return lnk
+    raise RuntimeError("LedgerPro shortcut not found on the Desktop")
+
+
+def click(image):
+    end = time.time() + FIND_WAIT_S
+    while time.time() < end:
+        hit = pyautogui.locateCenterOnScreen(str(ASSETS / image), confidence=CONFIDENCE)
+        if hit:
+            pyautogui.click(hit)
+            return
+        time.sleep(1)
+    pyautogui.screenshot(str(BASE / "logs" / f"notfound_{image}"))
+    raise RuntimeError(f"Button {image} not found on screen")
+
+
+def ledger_entries(path):
+    """LedgerPro export (File > Export > Excel) saved by the bot to data/ledger_export.xlsx."""
+    ws = load_workbook(path, data_only=True).active
+    return {str(r[0]): float(r[1]) for r in ws.iter_rows(min_row=2, values_only=True) if r[0]}
+
+
+def reconcile(statement, ledger):
+    for ref, amount in statement.items():
+        if ref not in ledger:
+            yield ref, amount, "Missing", "Not in LedgerPro - posting"
+        elif abs(ledger[ref] - amount) > TOLERANCE:
+            yield ref, amount, "Mismatch", f"LedgerPro has {ledger[ref]:.2f}"
+        else:
+            yield ref, amount, "OK", ""
+
+
+def post_entry(ref, amount):
+    click("new_entry.png")
+    pyautogui.write(ref)
+    pyautogui.press("tab")
+    pyautogui.write(f"{amount:.2f}")
+    click("save_btn.png")
+
+
+def main(statement_path):
+    (BASE / "BOT_WAS_RUN").write_text("1")
+    subprocess.Popen(["cmd", "/c", "start", "", str(find_shortcut())])
+    time.sleep(15)
+    click("export_excel.png")
+    time.sleep(10)
+    ledger = ledger_entries(BASE / "data" / "ledger_export.xlsx")
+    ws_in = load_workbook(statement_path, data_only=True).active
+    statement = {str(r[0]): float(r[1]) for r in ws_in.iter_rows(min_row=2, values_only=True) if r[0]}
+    out = Workbook()
+    ws = out.active
+    ws.append(["Reference", "Amount", "Result", "Note"])
+    counts = {"OK": 0, "Missing": 0, "Mismatch": 0, "Error": 0}
+    for ref, amount, result, note in reconcile(statement, ledger):
+        if result == "Missing":
+            try:
+                post_entry(ref, amount)
+                note = "Posted"
+            except RuntimeError as e:
+                result, note = "Error", str(e)
+        counts[result] += 1
+        ws.append([ref, amount, result, note])
+        log.info("%s %s %s", ref, result, note)
+    out.save(BASE / "data" / "output" / f"Reconcile_{datetime.now():%Y%m%d_%H%M}.xlsx")
+    print(json.dumps({"summary": counts}))
+
+
+if __name__ == "__main__":
+    main(sys.argv[1])