Restore archived tasks, fix dashboard scroll-reset, bind ornith, add Playwright smoke test

- **Restore 82 completed tasks** from tasks/complete/ back to tasks/ top
  level (all <7 days old per the cleanup policy; premature bulk archive
  was fixed).
- **Dashboard: fix scroll-reset on auto-refresh** — renderBoard rebuilds
  the board via innerHTML every 2s, destroying each column-body's
  scrollTop. Now snapshots column-body scrollTop + board.scrollLeft +
  view.scrollTop before rebuild and restores after (matched by
  PHASE_GROUPS index).
- **Dashboard UI additions** (pre-existing unstaged work): approval
  section cards, transition buttons, inline artifact editor (textarea for
  writing missing SPEC/VERDICT/etc from the detail modal).
- **Bind ornith as Implement model** — config.md: Model explicit to
  omlx/Ornith-1.0-35B-4bit-mlx, context window 32768. Interactive
  autopilot already used ornith via opencode default; now explicit.
- **Fix cleanup stub** — automaton-cleanup.sh had a stale --project arg
  pointing at a pytest temp dir (test isolation leak). Rewired to point
  at ~/.automaton.
- **Fix plist-isolation test** — test asserted host plist doesn't exist,
  but a real install creates it. Now snapshots mtime before run, asserts
  unchanged after (only a write during the test counts as bleed).
- **New Playwright smoke test** (tests/test_dashboard_ui.py) — 2 tests:
  board renders tasks, column scroll survives auto-refresh tick.
  Verified the test fails without the scroll fix (scrollTop resets to 0).
  Skipped via importorskip when playwright is absent (main CI stays
  green).
- **Clarify SI loop scope in README** — new-project onboarding section
  documents the framework-scoped self-improvement loop and options
  (leave/pause/create project loop).
- **CHANGELOG** documents all changes including the known model-divergence
  gap (mde tasks marked complete but per-role model binding was never
  implemented).

tests/test_dashboard_ui.py

@@ -0,0 +1,115 @@
+"""Playwright browser smoke test for the dashboard UI.
+
+Verifies the dashboard boots, renders tasks, and preserves column scroll
+position across auto-refresh (the fix for the scroll-reset bug). Skipped
+when playwright/chromium is not installed so the main suite stays green
+in CI environments without a browser.
+
+Run standalone:
+    python3 -m pytest tests/test_dashboard_ui.py -v
+"""
+
+from __future__ import annotations
+
+import socket
+import threading
+import time
+from http.server import HTTPServer
+from pathlib import Path
+
+import pytest
+
+from automaton.dashboard.ui.app import DashboardHandler, DashboardApp
+
+playwright = pytest.importorskip("playwright")
+from playwright.sync_api import sync_playwright  # noqa: E402
+
+
+def _free_port() -> int:
+    s = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
+    s.bind(("127.0.0.1", 0))
+    port = s.getsockname()[1]
+    s.close()
+    return port
+
+
+def _make_done_tasks(project_root: Path, count: int) -> None:
+    tasks = project_root / ".automaton" / "tasks"
+    tasks.mkdir(parents=True, exist_ok=True)
+    for i in range(count):
+        d = tasks / f"done-task-{i:02d}"
+        d.mkdir(exist_ok=True)
+        (d / "SPEC.md").write_text(f"# Done task {i}\n" + ("line. " * 40) + "\n")
+        (d / ".state").write_text("complete\n")
+
+
+@pytest.fixture
+def dashboard_server(tmp_path: Path):
+    _make_done_tasks(tmp_path, 40)
+    app = DashboardApp(start_path=tmp_path)
+    assert app.initialize(), "dashboard failed to init"
+    port = _free_port()
+    DashboardHandler.config = app.config
+    DashboardHandler.project_root = app.project_root
+    DashboardHandler.scope = app.scope
+    server = HTTPServer(("127.0.0.1", port), DashboardHandler)
+    thread = threading.Thread(target=server.serve_forever, daemon=True)
+    thread.start()
+    base_url = f"http://127.0.0.1:{port}"
+    try:
+        yield base_url
+    finally:
+        server.shutdown()
+        server.server_close()
+        thread.join(timeout=5)
+
+
+def _wait_for(predicate, timeout=10.0, interval=0.2):
+    deadline = time.time() + timeout
+    while time.time() < deadline:
+        if predicate():
+            return True
+        time.sleep(interval)
+    return False
+
+
+class TestDashboardUI:
+    def test_board_renders_done_tasks(self, dashboard_server):
+        with sync_playwright() as p:
+            browser = p.chromium.launch()
+            page = browser.new_page()
+            page.goto(dashboard_server)
+            # Footer refresh counter increments once auto-refresh fires.
+            assert _wait_for(lambda: page.text_content("#refresh-count") not in (None, "0", ""))
+            # The Done (resolution) group contains our tasks.
+            cards = page.locator(".task-card")
+            assert _wait_for(lambda: cards.count() >= 40, timeout=15)
+            assert page.text_content("#done-tasks") == "40"
+            browser.close()
+
+    def test_refresh_preserves_column_scroll(self, dashboard_server):
+        """The scroll-reset bug: re-render snapped a column's scrollTop to 0,
+        so users couldn't scroll the Done group down to review older tasks.
+        Verify the column-body scrollTop survives an auto-refresh tick."""
+        with sync_playwright() as p:
+            browser = p.chromium.launch()
+            # Wide viewport so columns fit and the Done column-body overflows
+            # vertically (rather than the board squeezing them horizontally).
+            page = browser.new_page(viewport={"width": 1400, "height": 500})
+            page.goto(dashboard_server)
+            assert _wait_for(lambda: page.locator(".task-card").count() >= 40, timeout=15)
+            # The Done (resolution) group is the last column.
+            bodies = page.locator(".column-body")
+            n = bodies.count()
+            done_body = bodies.nth(n - 1)
+            assert done_body.evaluate("el => el.scrollHeight > el.clientHeight")
+            # Scroll the column-body well past the top.
+            target = 400
+            done_body.evaluate("(el, y) => { el.scrollTop = y; }", target)
+            assert done_body.evaluate("el => el.scrollTop") == target
+            # Wait for an auto-refresh re-render (default interval 2s).
+            before = int(page.text_content("#refresh-count") or "0")
+            assert _wait_for(lambda: int(page.text_content("#refresh-count") or "0") > before, timeout=10)
+            # Scroll must survive the re-render (the bug fix).
+            assert done_body.evaluate("el => el.scrollTop") == target
+            browser.close()