fix: check_plugin.py honors per-plugin test/harness.json (#365)

check_one() always compares the render against committed golden images, but
the CLI never loaded the plugin's test/harness.json — so the deterministic
settings the goldens were generated with (config, mock data, frozen time,
sizes) weren't applied. For any time/data-dependent plugin this means the CLI
(and the plugins-repo CI workflow that calls it) renders live data and the
golden drifts on every run, even with no real regression. The pytest matrix
path already reads harness.json via load_harness_spec; the CLI now does too.

- check_one loads load_harness_spec(plugin_dir) and layers it under explicit
  CLI flags: config = schema defaults < harness.json < --config; sizes =
  --sizes > LEDMATRIX_TEST_SIZES env > harness.json > default sample;
  mock_data/freeze_time/skip_update fall back to harness.json when not given
  on the CLI.
- parse_sizes returns None (not DEFAULT_TEST_SIZES) when --sizes is omitted,
  so the env/harness.json/default fallback chain in resolve_test_sizes applies.
- Regression tests: harness.json supplies render settings, and CLI flags
  override it. Use a temp fixture plugin so they run in core CI (no plugins).

Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>

.github/workflows/test.yml

@@ -0,0 +1,33 @@
+name: Tests
+
+on:
+  pull_request:
+  push:
+    branches: [main]
+
+jobs:
+  plugin-safety:
+    name: Plugin safety harness + unit tests
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # v4.2.2
+        with:
+          persist-credentials: false
+
+      - uses: actions/setup-python@0b93645e9fea7318ecaed2b359559ac225c90a2b # v5.3.0
+        with:
+          python-version: "3.12"
+          cache: pip
+
+      - name: Install dependencies
+        run: |
+          python -m pip install --upgrade pip
+          pip install -r requirements.txt -r requirements-test.txt
+          pip install RGBMatrixEmulator
+
+      - name: Run harness + visual rendering tests
+        run: |
+          pytest --no-cov \
+            test/plugins/test_harness.py \
+            test/plugins/test_visual_rendering.py \
+            test/plugins/test_plugin_matrix.py

docs/plugin-safety-harness.md

@@ -0,0 +1,136 @@
+# Plugin Safety Harness
+
+Renders a plugin across **every declared screen (mode)** and **a spread of
+matrix sizes**, and fails if any combination crashes, draws past the panel edge,
+or — for plugins that ship golden images — drifts visually. The goal: change a
+plugin without breaking a size or screen you didn't think to test.
+
+## Sizes: a sample, not a fixed list
+
+There is **no fixed set of supported panel sizes** — an RGB matrix build can be
+any width/height and configuration (square, rectangle, 2×2, 4×4, 8×2, long
+strips, tall stacks). Plugins are expected to read dimensions dynamically
+(`self.display_manager.matrix.width/height`) and lay themselves out
+accordingly, so a hardcoded coordinate or unscaled font shows up as a failure
+here.
+
+The harness therefore renders against a **representative sample** that spans the
+axes of variation (`DEFAULT_TEST_SIZES` in `src/plugin_system/testing/sizes.py`),
+not an authoritative list:
+
+Each module is 64×32; entries are real panel-grid arrangements (cols × rows):
+
+| Size    | Grid | Why it's in the sample                     |
+|---------|------|--------------------------------------------|
+| 64×32   | 1×1  | single panel — tightest common rectangle   |
+| 128×32  | 2×1  | the baseline most plugins are tuned for    |
+| 64×64   | 1×2  | stacked — tall-narrow centering            |
+| 128×64  | 2×2  | block — icon scaling / vertical centering  |
+| 256×32  | 4×1  | long strip — wide horizontal layout        |
+| 128×96  | 2×3  | tall — vertical overflow                   |
+| 256×128 | 4×4  | large block — both dimensions big at once  |
+
+**Override the sizes entirely** to test your actual hardware (or any shape):
+
+```bash
+# CLI — one-off:
+python scripts/check_plugin.py --plugin clock-simple --sizes 8x16,64x64,256x32
+
+# pytest — force every plugin onto your panel(s):
+LEDMATRIX_TEST_SIZES="8x16,128x128" pytest test/plugins/test_plugin_matrix.py
+
+# Per-plugin — declare the shapes a plugin targets in its test/harness.json:
+#   { "sizes": [[8, 16], [64, 64]] }
+```
+
+Precedence: `LEDMATRIX_TEST_SIZES` env (global) → per-plugin `harness.json`
+`sizes` → the default sample. Bounds checking adapts to whatever sizes a run
+uses — the backing canvas is padded out to the **largest** panel in the run, so
+a coordinate meant for a big build is still caught when rendering a small one.
+
+## Quick start
+
+```bash
+# Functional + bounds check across all sizes/screens:
+python scripts/check_plugin.py --plugin clock-simple
+
+# Every discovered plugin:
+python scripts/check_plugin.py --all
+
+# Dump PNGs to eyeball each size/screen:
+python scripts/check_plugin.py --plugin ledmatrix-weather --out-dir /tmp/preview
+```
+
+Exit code is non-zero if any `(plugin, size, screen)` fails. Plugins are
+discovered in `plugin-repos/` and `plugins/` (override with `--plugin-dir`).
+
+## What it checks (Phase 1 — always on)
+
+1. **Loads** and builds its mode list.
+2. **Renders every screen** at every size without raising. `update()` may fail
+   (no network in CI) and is tolerated; a crash in `display()` is a failure —
+   `display()` must handle the no-data state.
+3. **Bounds**: nothing is drawn past the right/bottom edge. Implemented by
+   `BoundsCheckingDisplayManager`, which backs the declared panel with an
+   oversized canvas and flags any pixels that land in the margin. (Left/top
+   overflow at negative coordinates and BDF text are not flagged — golden images
+   cover those.)
+
+## Golden images (Phase 2 — opt-in per plugin)
+
+A plugin opts in by committing reference PNGs and (usually) a small harness spec:
+
+```
+plugins/<id>/test/harness.json          # how to render deterministically
+plugins/<id>/test/fixtures/mock.json     # optional cached data
+plugins/<id>/test/golden/<WxH>/<mode>.png
+```
+
+`test/harness.json` keys (all optional):
+
+```json
+{
+  "config":      { "timezone": "UTC" },
+  "mock_data":   "fixtures/mock.json",
+  "freeze_time": "2025-08-01 15:25:00",
+  "skip_update": false,
+  "sizes":       [[128, 32], [128, 64]]
+}
+```
+
+Generate / refresh goldens after an intentional visual change, then review the
+diff before committing:
+
+```bash
+python scripts/check_plugin.py --plugin clock-simple --update-golden \
+  --config '{"timezone":"UTC"}' --freeze-time "2025-08-01 15:25:00"
+```
+
+Comparison is exact by default (`compare_images` in `harness.py` accepts a
+tolerance for known anti-aliasing noise). Determinism requires a pinned Pillow
+and the bundled fonts — keep both stable when regenerating goldens.
+
+## Tests & CI
+
+- `test/plugins/test_harness.py` — unit tests for bounds detection, image
+  comparison, and mode enumeration (run anywhere).
+- `test/plugins/test_plugin_matrix.py` — parametrized over discovered plugins ×
+  sizes × screens; honors each plugin's `test/harness.json` and goldens. Skips
+  when no plugins are present (e.g. a fresh core checkout); set
+  `LEDMATRIX_REQUIRE_PLUGINS=1` in a pipeline where plugins must be present to
+  turn an empty discovery into a hard failure instead. Point it at the monorepo
+  with `LEDMATRIX_PLUGINS_DIR=/path/to/ledmatrix-plugins/plugins`.
+- `.github/workflows/test.yml` — runs the harness + visual tests on every PR.
+
+The plugin monorepo has its own `Plugin Safety` workflow that runs this harness
+against changed plugins on every PR.
+
+## Developer workflow
+
+1. Change the plugin on a branch.
+2. `python scripts/check_plugin.py --plugin <id> --out-dir /tmp/preview` and
+   eyeball the PNGs.
+3. Intentional visual change? `--update-golden`, review diffs, commit goldens.
+4. (Monorepo) bump `manifest.json` version and let the pre-commit hook sync
+   `plugins.json`.
+5. Push — CI re-runs the harness across all sizes and gates the PR.

requirements-test.txt

@@ -0,0 +1,9 @@
+# Test-only dependencies for the plugin safety harness and pytest suite.
+# Install alongside requirements.txt:  pip install -r requirements.txt -r requirements-test.txt
+#
+# pytest, pytest-cov, pytest-mock, and jsonschema are already pinned (with
+# major-version caps) in requirements.txt, so they are intentionally NOT
+# repeated here — re-pinning pytest to <9 collided with requirements.txt's
+# pytest>=9.0.3,<10 and made the two files impossible to install together.
+# Only declare what requirements.txt doesn't already provide.
+freezegun>=1.2,<2        # deterministic time for golden-image tests

scripts/check_plugin.py

@@ -0,0 +1,232 @@
+#!/usr/bin/env python3
+"""
+Plugin safety checker.
+
+Renders a plugin across every declared screen (mode) and every supported matrix
+size, and fails if any screen crashes, overflows the panel, or (for plugins with
+committed golden images) drifts visually.
+
+Usage:
+    # Functional + bounds check across all sizes/modes:
+    python scripts/check_plugin.py --plugin clock-simple
+
+    # Every discovered plugin:
+    python scripts/check_plugin.py --all
+
+    # Dump PNGs for each size/mode so you can eyeball them:
+    python scripts/check_plugin.py --plugin ledmatrix-weather --out-dir /tmp/preview
+
+    # Refresh committed golden images after an intentional visual change:
+    python scripts/check_plugin.py --plugin clock-simple --update-golden \
+        --mock-data plugins/clock-simple/test/fixtures/mock.json
+
+Exit code is non-zero if any (plugin, size, mode) fails.
+"""
+
+import argparse
+import json
+import os
+import sys
+from pathlib import Path
+from typing import Dict, List, Optional
+
+PROJECT_ROOT = Path(__file__).resolve().parent.parent
+sys.path.insert(0, str(PROJECT_ROOT))
+
+os.environ['EMULATOR'] = 'true'
+
+from src.logging_config import get_logger  # noqa: E402
+from src.plugin_system.testing.loading import (  # noqa: E402
+    find_plugin_dir, load_config_defaults, load_harness_spec,
+)
+from src.plugin_system.testing.harness import (  # noqa: E402
+    RenderResult, render_plugin_matrix, compare_to_goldens, write_goldens,
+)
+from src.plugin_system.testing.sizes import (  # noqa: E402
+    parse_size_token, resolve_test_sizes, safe_mode_filename, size_label,
+)
+
+logger = get_logger("[Check Plugin]")
+
+DEFAULT_SEARCH_DIRS = [
+    str(PROJECT_ROOT / 'plugins'),
+    str(PROJECT_ROOT / 'plugin-repos'),
+]
+
+
+def discover_plugins(search_dirs: List[str]) -> List[str]:
+    """All plugin ids found across the search dirs (dirs containing manifest.json)."""
+    found = []
+    for d in search_dirs:
+        base = Path(d)
+        if not base.exists():
+            continue
+        for child in sorted(base.iterdir()):
+            if (child / 'manifest.json').exists() and child.name not in found:
+                found.append(child.name)
+    return found
+
+
+def parse_sizes(spec: Optional[str]):
+    if not spec:
+        return None
+    sizes = []
+    for token in spec.split(','):
+        if not token.strip():
+            continue
+        try:
+            sizes.append(parse_size_token(token))
+        except ValueError as exc:
+            raise SystemExit(str(exc)) from exc
+    return sizes
+
+
+def check_one(plugin_id: str, search_dirs: List[str], sizes, mock_data: Dict,
+              config: Dict, run_update: bool, out_dir: Optional[Path],
+              update_golden: bool, golden_dir_override: Optional[Path],
+              freeze_time: Optional[str]) -> List[RenderResult]:
+    plugin_dir = find_plugin_dir(plugin_id, search_dirs)
+    if not plugin_dir:
+        logger.error("Plugin '%s' not found in: %s", plugin_id, search_dirs)
+        return [RenderResult(plugin_id, 0, 0, "<not-found>", error="plugin directory not found")]
+
+    # Per-plugin test/harness.json holds the deterministic settings the committed
+    # goldens were generated with (config, mock data, frozen time, sizes). Load
+    # them so the CLI/CI render reproduces the golden the same way the pytest
+    # matrix path does; explicit CLI flags still override the file.
+    spec = load_harness_spec(plugin_dir)
+
+    # config_schema defaults (real-install behavior), then harness.json config,
+    # then CLI --config — most specific wins.
+    full_config = {"enabled": True}
+    full_config.update(load_config_defaults(plugin_dir))
+    full_config.update(spec.get("config", {}))
+    full_config.update(config)
+
+    # Precedence: CLI flag > LEDMATRIX_TEST_SIZES env > harness.json > default.
+    effective_sizes = sizes if sizes else resolve_test_sizes(spec.get("sizes"))
+    # CLI value wins when provided, else fall back to the harness.json setting.
+    effective_mock_data = mock_data or spec.get("mock_data_contents", {})
+    effective_freeze = freeze_time or spec.get("freeze_time")
+    effective_run_update = run_update and not spec.get("skip_update", False)
+
+    results = render_plugin_matrix(
+        plugin_id=plugin_id, plugin_dir=plugin_dir, config=full_config,
+        mock_data=effective_mock_data, sizes=effective_sizes,
+        run_update=effective_run_update, freeze_time=effective_freeze,
+    )
+
+    golden_dir = golden_dir_override or (plugin_dir / 'test' / 'golden')
+    if update_golden:
+        written = write_goldens(results, golden_dir)
+        logger.info("Wrote %d golden image(s) for %s to %s", written, plugin_id, golden_dir)
+    else:
+        compare_to_goldens(results, golden_dir)
+
+    if out_dir:
+        for r in results:
+            if r.image is None:
+                continue
+            dest = out_dir / plugin_id / size_label(r.width, r.height)
+            dest.mkdir(parents=True, exist_ok=True)
+            r.image.save(dest / f"{safe_mode_filename(r.mode)}.png", format="PNG")
+
+    return results
+
+
+def print_report(all_results: Dict[str, List[RenderResult]]) -> bool:
+    """Print a per-plugin grid. Returns True if everything passed."""
+    everything_ok = True
+    for plugin_id, results in all_results.items():
+        print(f"\n=== {plugin_id} ===")
+        for r in results:
+            if r.ok:
+                status = "PASS"
+                detail = ""
+                if r.golden_checked:
+                    detail = " (golden ✓)"
+                if r.update_error is not None:
+                    detail += f" (update warn: {r.update_error})"
+            else:
+                everything_ok = False
+                if r.error is not None:
+                    status, detail = "FAIL", f" error={r.error}"
+                elif r.overflow is not None:
+                    status, detail = "FAIL", f" overflow bbox={r.overflow}"
+                elif r.golden_ok is False:
+                    status = "FAIL"
+                    detail = f" golden drift: {r.golden_diff_pixels}px (max Δ={r.golden_max_delta})"
+                else:
+                    status, detail = "FAIL", ""
+            print(f"  [{status}] {r.size_label:>7}  {r.mode}{detail}")
+    print()
+    return everything_ok
+
+
+def main() -> int:
+    parser = argparse.ArgumentParser(description="Check a plugin renders safely across sizes & screens")
+    group = parser.add_mutually_exclusive_group(required=True)
+    group.add_argument('--plugin', '-p', help='Plugin id to check')
+    group.add_argument('--all', action='store_true', help='Check every discovered plugin')
+    parser.add_argument('--plugin-dir', '-d', default=None, help='Directory to search for plugins')
+    parser.add_argument('--sizes', default=None, help='Comma-separated WxH list (default: all supported)')
+    parser.add_argument('--config', '-c', default='{}', help='Plugin config overrides as JSON')
+    parser.add_argument('--mock-data', '-m', default=None, help='Path to JSON file with mock cache data')
+    parser.add_argument('--out-dir', '-o', default=None, help='Also dump rendered PNGs here')
+    parser.add_argument('--skip-update', action='store_true', help='Skip calling update()')
+    parser.add_argument('--update-golden', action='store_true', help='Write/refresh golden images')
+    parser.add_argument('--golden-dir', default=None, help='Override golden dir (default: <plugin>/test/golden)')
+    parser.add_argument('--freeze-time', default=None,
+                        help='Freeze wall clock, e.g. "2025-08-01 15:25:00" (for time-dependent plugins)')
+    args = parser.parse_args()
+
+    search_dirs = [args.plugin_dir] if args.plugin_dir else DEFAULT_SEARCH_DIRS
+    sizes = parse_sizes(args.sizes)
+
+    try:
+        config = json.loads(args.config)
+    except json.JSONDecodeError as e:
+        logger.error("Invalid --config JSON: %s", e)
+        return 2
+    if not isinstance(config, dict):
+        logger.error("--config must be a JSON object, got %s", type(config).__name__)
+        return 2
+
+    mock_data = {}
+    if args.mock_data:
+        mock_path = Path(args.mock_data)
+        if not mock_path.exists():
+            logger.error("Mock data file not found: %s", args.mock_data)
+            return 2
+        with open(mock_path) as f:
+            mock_data = json.load(f)
+        if not isinstance(mock_data, dict):
+            logger.error("--mock-data must be a JSON object (key -> cache value), got %s",
+                         type(mock_data).__name__)
+            return 2
+
+    plugin_ids = discover_plugins(search_dirs) if args.all else [args.plugin]
+    if not plugin_ids:
+        logger.error("No plugins found in: %s", search_dirs)
+        return 2
+
+    out_dir = Path(args.out_dir) if args.out_dir else None
+    golden_dir_override = Path(args.golden_dir) if args.golden_dir else None
+
+    all_results: Dict[str, List[RenderResult]] = {}
+    for plugin_id in plugin_ids:
+        all_results[plugin_id] = check_one(
+            plugin_id=plugin_id, search_dirs=search_dirs, sizes=sizes,
+            mock_data=mock_data, config=config, run_update=not args.skip_update,
+            out_dir=out_dir, update_golden=args.update_golden,
+            golden_dir_override=golden_dir_override, freeze_time=args.freeze_time,
+        )
+
+    # When refreshing goldens we skip drift comparison, but a crash or overflow
+    # still means the plugin is broken — never let --update-golden mask that.
+    ok = print_report(all_results)
+    return 0 if ok else 1
+
+
+if __name__ == '__main__':
+    sys.exit(main())

scripts/render_plugin.py

@@ -17,7 +17,6 @@ import os
 import json
 import argparse
 from pathlib import Path
-from typing import Any, Dict, Optional, Sequence, Union
 
 # Add project root to path
 PROJECT_ROOT = Path(__file__).resolve().parent.parent
@@ -28,49 +27,15 @@ os.environ['EMULATOR'] = 'true'
 
 # Import logger after path setup so src.logging_config is importable
 from src.logging_config import get_logger  # noqa: E402
+from src.plugin_system.testing.loading import (  # noqa: E402
+    find_plugin_dir, load_manifest, load_config_defaults,
+)
 logger = get_logger("[Render Plugin]")
 
 MIN_DIMENSION = 1
 MAX_DIMENSION = 512
 
 
-def find_plugin_dir(plugin_id: str, search_dirs: Sequence[Union[str, Path]]) -> Optional[Path]:
-    """Find a plugin directory by searching multiple paths."""
-    from src.plugin_system.plugin_loader import PluginLoader
-    loader = PluginLoader()
-    for search_dir in search_dirs:
-        search_path = Path(search_dir)
-        if not search_path.exists():
-            continue
-        result = loader.find_plugin_directory(plugin_id, search_path)
-        if result:
-            return Path(result)
-    return None
-
-
-def load_manifest(plugin_dir: Path) -> Dict[str, Any]:
-    """Load and return manifest.json from plugin directory."""
-    manifest_path = plugin_dir / 'manifest.json'
-    if not manifest_path.exists():
-        raise FileNotFoundError(f"No manifest.json in {plugin_dir}")
-    with open(manifest_path, 'r') as f:
-        return json.load(f)
-
-
-def load_config_defaults(plugin_dir: Path) -> Dict[str, Any]:
-    """Extract default values from config_schema.json."""
-    schema_path = plugin_dir / 'config_schema.json'
-    if not schema_path.exists():
-        return {}
-    with open(schema_path, 'r') as f:
-        schema = json.load(f)
-    defaults: Dict[str, Any] = {}
-    for key, prop in schema.get('properties', {}).items():
-        if 'default' in prop:
-            defaults[key] = prop['default']
-    return defaults
-
-
 def main() -> int:
     """Load a plugin, call update() + display(), and save the result as a PNG image."""
     parser = argparse.ArgumentParser(description='Render a plugin display to a PNG image')

src/cache_manager.py

@@ -1,3 +1,28 @@
+"""
+Cache Manager — multi-tier response cache for the LEDMatrix application.
+
+:class:`CacheManager` provides a unified caching layer used by all plugins
+to reduce external API calls and survive network outages gracefully.
+
+Two storage tiers
+-----------------
+* **Memory tier** (:class:`~src.cache.memory_cache.MemoryCache`): fast LRU
+  cache (up to 1 000 entries by default).  Hit on this tier before touching
+  disk.
+* **Disk tier** (:class:`~src.cache.disk_cache.DiskCache`): filesystem-backed
+  persistent store that survives process restarts.
+
+Data written to cache is serialised as JSON.  :class:`DateTimeEncoder` handles
+``datetime`` objects transparently so callers don't have to pre-serialise them.
+
+Typical plugin usage::
+
+    data = self.cache_manager.get_cached_data('my_key', max_age=300)
+    if data is None:
+        data = fetch_from_api()
+        self.cache_manager.save_cache('my_key', data)
+"""
+
 import json
 import os
 import time
@@ -15,7 +40,10 @@ from src.cache.cache_metrics import CacheMetrics
 from src.logging_config import get_logger
 
 class DateTimeEncoder(json.JSONEncoder):
+    """JSON encoder that serialises ``datetime`` objects as ISO-8601 strings."""
+
     def default(self, obj):
+        """Return ISO-8601 string for datetime; delegate all other types to the base encoder."""
         if isinstance(obj, datetime):
             return obj.isoformat()
         return super().default(obj)

src/common/scroll_helper.py

@@ -347,34 +347,40 @@ class ScrollHelper:
         return self._get_visible_portion_integer(start_x_int, end_x_int)
     
     def _get_visible_portion_integer(self, start_x: int, end_x: int) -> Image.Image:
-        """Fast integer pixel extraction (no interpolation)."""
+        """Fast integer pixel extraction (no interpolation).
-        # Fast numpy array slicing for normal case (no wrap-around)
+
-        if end_x <= self.cached_image.width:
+        Uses Image.frombytes instead of Image.fromarray: frombytes skips
-            # Normal case: single slice - fastest path
+        numpy's array-protocol overhead and is ~50% faster for the display-sized
-            frame_array = self.cached_array[:, start_x:end_x]
+        slices (128×32 = 12 KB) used here.
-            # Convert to PIL Image (minimal overhead)
+        """
-            return Image.fromarray(frame_array)
+        _size = (self.display_width, self.display_height)
+        img_w = self.cached_image.width
+
+        if end_x <= img_w:
+            # Normal case: single contiguous slice (fastest path)
+            frame_array = np.ascontiguousarray(self.cached_array[:, start_x:end_x])
+            return Image.frombytes('RGB', _size, frame_array.tobytes())
         else:
-            # Wrap-around case: combine two slices using numpy
+            # Ensure frame buffer is allocated for all non-simple paths
-            width1 = self.cached_image.width - start_x
+            if self._frame_buffer is None or self._frame_buffer.shape != (self.display_height, self.display_width, 3):
+                self._frame_buffer = np.zeros((self.display_height, self.display_width, 3), dtype=np.uint8)
+
+            width1 = img_w - start_x
             if width1 > 0:
-                # Use pre-allocated buffer for output
+                # Wrap-around: tail of image + head of image
-                if self._frame_buffer is None or self._frame_buffer.shape != (self.display_height, self.display_width, 3):
-                    self._frame_buffer = np.zeros((self.display_height, self.display_width, 3), dtype=np.uint8)
-                
-                # First part from end of image (fast numpy slice)
                 self._frame_buffer[:, :width1] = self.cached_array[:, start_x:]
-                
-                # Second part from beginning of image
                 remaining_width = self.display_width - width1
                 self._frame_buffer[:, width1:] = self.cached_array[:, :remaining_width]
-                
-                # Convert combined buffer to PIL Image
-                return Image.fromarray(self._frame_buffer)
             else:
-                # Edge case: start_x >= image width, wrap to beginning
+                # Edge case: start_x at or past image end — show from beginning,
-                frame_array = self.cached_array[:, :self.display_width]
+                # clamped to available width (scroll_position should wrap before
-                return Image.fromarray(frame_array)
+                # reaching this state in normal operation).
+                available = min(self.display_width, img_w)
+                self._frame_buffer[:, :available] = self.cached_array[:, :available]
+                if available < self.display_width:
+                    self._frame_buffer[:, available:] = 0
+
+            return Image.frombytes('RGB', _size, self._frame_buffer.tobytes())
     
     def _get_visible_portion_subpixel(self, start_x_int: int, fractional: float) -> Image.Image:
         """

src/config_manager.py

@@ -1,3 +1,29 @@
+"""
+Config Manager — reads, writes, and validates ``config/config.json``.
+
+:class:`ConfigManager` is the single owner of the on-disk configuration
+files:
+
+* ``config/config.json``         — main user-editable configuration.
+* ``config/config_secrets.json`` — sensitive values (API keys, tokens).
+
+All writes go through :class:`~src.config_manager_atomic.AtomicConfigManager`
+which performs a backup before overwriting, validates the result, and rolls
+back on error.  This makes config corruption essentially impossible.
+
+Plugin configuration
+--------------------
+Plugin configs are stored inside ``config.json`` under the plugin's ID key
+and survive plugin reinstalls.  Use :meth:`ConfigManager.update_plugin_config`
+to write plugin settings; never write directly to the plugin directory.
+
+Hot-reload
+----------
+:class:`~src.config_service.ConfigService` wraps ``ConfigManager`` and
+detects file changes, broadcasting the new config to registered listeners
+without requiring a restart.
+"""
+
 import json
 import os
 import logging
@@ -17,6 +43,13 @@ from src.common.permission_utils import (
 )
 
 class ConfigManager:
+    """
+    Reads and writes the main application configuration files.
+
+    Wraps :class:`~src.config_manager_atomic.AtomicConfigManager` for safe
+    atomic writes with automatic backup and rollback.  Also exposes helpers
+    for plugin configuration persistence and secret-field masking.
+    """
     def __init__(self, config_path: Optional[str] = None, secrets_path: Optional[str] = None) -> None:
         # Use current working directory as base
         self.config_path: str = config_path or "config/config.json"
@@ -29,9 +62,11 @@ class ConfigManager:
         self._atomic_manager: Optional[AtomicConfigManager] = None
 
     def get_config_path(self) -> str:
+        """Return the path to the main config file (``config/config.json``)."""
         return self.config_path
 
     def get_secrets_path(self) -> str:
+        """Return the path to the secrets file (``config/config_secrets.json``)."""
         return self.secrets_path
     
     def _get_atomic_manager(self) -> AtomicConfigManager:

src/display_controller.py

@@ -1,3 +1,25 @@
+"""
+Display Controller — top-level orchestration for the LEDMatrix application.
+
+This module owns the main run loop that drives the LED display.  It ties
+together every major subsystem:
+
+  - ConfigManager / ConfigService  — loads config.json, hot-reloads on change
+  - DisplayManager                 — hardware (or emulator) output interface
+  - FontManager                    — TTF/BDF font loading and caching
+  - CacheManager                   — multi-tier API response cache
+  - PluginManager                  — plugin lifecycle (load, update, display)
+  - DisplaySyncManager             — optional leader/follower multi-Pi sync
+  - VegasModeCoordinator           — optional continuous Vegas scroll mode
+
+The main loop inside :meth:`DisplayController.run` rotates through enabled
+plugin display modes, respecting schedule windows, brightness dim schedules,
+on-demand overrides, and live-priority interrupts.
+
+Entry point: :func:`main` — instantiates :class:`DisplayController` and calls
+:meth:`~DisplayController.run`.
+"""
+
 import time
 import os
 import json
@@ -28,6 +50,24 @@ DEFAULT_DYNAMIC_DURATION_CAP = 180.0
 WIFI_STATUS_FILE = None  # Will be initialized in __init__
 
 class DisplayController:
+    """
+    Top-level controller that owns the LED display run loop.
+
+    Responsibilities
+    ----------------
+    * Initialise and wire together all subsystems at startup.
+    * Rotate through plugin display modes in :meth:`run`.
+    * Honour schedule windows (active/inactive hours) and dim schedules.
+    * Handle on-demand override requests (external callers can pin a
+      specific plugin/mode for a fixed duration via the cache bus).
+    * Coordinate with a follower Pi when multi-display sync is configured.
+    * Delegate all actual content to the plugin system — this class contains
+      no display logic of its own.
+
+    There is exactly one instance per process; call :func:`main` to create
+    it and start the run loop.
+    """
+
     def __init__(self):
         start_time = time.time()
         logger.info("Starting DisplayController initialization")
@@ -138,7 +178,11 @@ class DisplayController:
         self.on_demand_last_event: Optional[str] = None
         self.on_demand_schedule_override = False
         self.rotation_resume_index: Optional[int] = None
-        
+        # Saved rotation position when a live-priority plugin preempts the
+        # rotation, so it resumes where it left off (not after the live plugin)
+        # once live priority ends.
+        self._live_resume_index: Optional[int] = None
+
         # WiFi status message tracking
         global WIFI_STATUS_FILE
         if WIFI_STATUS_FILE is None:
@@ -148,7 +192,11 @@ class DisplayController:
         self.wifi_status_file = WIFI_STATUS_FILE
         self.wifi_status_active = False
         self.wifi_status_expires_at: Optional[float] = None
-        
+
+        # Plugin display() signature cache — must be initialised before the plugin
+        # loading loop below so the .pop() invalidation at load time is always safe.
+        self._plugin_accepts_display_mode: Dict[str, bool] = {}
+
         try:
             logger.info("Attempting to import plugin system...")
             from src.plugin_system import PluginManager
@@ -321,6 +369,8 @@ class DisplayController:
                             self.plugin_modes[mode] = plugin_instance
                             self.mode_to_plugin_id[mode] = plugin_id
                             logger.debug("  Added mode: %s", mode)
+                        # Invalidate signature cache so the new instance is re-inspected
+                        self._plugin_accepts_display_mode.pop(plugin_id, None)
                         
                         # Show progress
                         progress_pct = int((loaded_count / enabled_count) * 100)
@@ -367,11 +417,39 @@ class DisplayController:
         self.is_display_active = True
         self._was_display_active = True  # Track previous state for schedule change detection
 
+        # --- Opt #2: cached config values ---
+        # Avoids chained dict.get() with temporary {} defaults on every hot path call.
+        # Refreshed via _refresh_config_cache() on every hot-reload.
+        self._normal_brightness: int = (
+            self.config.get('display', {}).get('hardware', {}).get('brightness', 90)
+        )
+        self._scroll_speed: float = (
+            self.config.get('display', {}).get('vegas_scroll', {}).get('scroll_speed', 75)
+        )
+
         # Brightness state tracking for dim schedule
-        self.current_brightness = self.config.get('display', {}).get('hardware', {}).get('brightness', 90)
+        self.current_brightness = self._normal_brightness
         self.is_dimmed = False
         self._was_dimmed = False
 
+        # --- Opt #3: schedule minute-gate ---
+        # Both _check_schedule and _check_dim_schedule re-evaluated at most once per
+        # clock minute.  Storing the (hour, minute) tuple that was last evaluated lets
+        # the methods skip all timezone / strptime work within the same minute.
+        # Reset to None on config change so the next call re-evaluates immediately.
+        self._tz = None                              # pytz timezone, lazily built from config
+        self._schedule_checked_minute: Optional[tuple] = None
+        self._dim_checked_minute: Optional[tuple] = None
+        self._cached_target_brightness: int = self._normal_brightness
+
+        # Register controller-level hot-reload callback so cached config values
+        # (_normal_brightness, _scroll_speed, _tz, minute-gates) stay in sync
+        # when the user saves settings via the web UI.
+        def _controller_config_change(old_config: Dict[str, Any], new_config: Dict[str, Any]) -> None:
+            self._refresh_config_cache(new_config)
+
+        self.config_service.subscribe(_controller_config_change)
+
         # Publish initial on-demand state
         try:
             self._publish_on_demand_state()
@@ -533,17 +611,24 @@ class DisplayController:
             logger.debug("Schedule is disabled - display always active")
             return
 
-        # Get configured timezone, default to UTC
+        # Lazily build the timezone object once; reuse on every subsequent call.
-        timezone_str = self.config.get('timezone', 'UTC')
+        if self._tz is None:
-        try:
+            timezone_str = self.config.get('timezone', 'UTC')
-            tz = pytz.timezone(timezone_str)
+            try:
-        except pytz.UnknownTimeZoneError:
+                self._tz = pytz.timezone(timezone_str)
-            logger.warning(f"Unknown timezone '{timezone_str}', using UTC")
+            except pytz.UnknownTimeZoneError:
-            tz = pytz.UTC
+                logger.warning("Unknown timezone '%s', using UTC", timezone_str)
+                self._tz = pytz.UTC
 
-        # Use timezone-aware current time
+        current_time = datetime.now(self._tz)
-        current_time = datetime.now(tz)
+        # Gate: schedule state can only change on a minute boundary, so skip
-        current_day = current_time.strftime('%A').lower()  # Get day name (monday, tuesday, etc.)
+        # all the strptime / comparison work if we already evaluated this minute.
+        current_minute_key = (current_time.hour, current_time.minute)
+        if current_minute_key == self._schedule_checked_minute:
+            return
+        self._schedule_checked_minute = current_minute_key
+
+        current_day = current_time.strftime('%A').lower()  # e.g. 'monday'
         current_time_only = current_time.time()
         
         # Check if per-day schedule is configured
@@ -632,8 +717,8 @@ class DisplayController:
             Target brightness level (dim_brightness if in dim period,
             normal brightness otherwise)
         """
-        # Get normal brightness from config
+        # Opt #2: use cached brightness rather than re-traversing config dict
-        normal_brightness = self.config.get('display', {}).get('hardware', {}).get('brightness', 90)
+        normal_brightness = self._normal_brightness
 
         # If display is OFF via schedule, don't process dim schedule
         if not self.is_display_active:
@@ -647,15 +732,21 @@ class DisplayController:
             self.is_dimmed = False
             return normal_brightness
 
-        # Get configured timezone
+        # Opt #3: lazily build timezone; gate full re-parse to once per clock minute
-        timezone_str = self.config.get('timezone', 'UTC')
+        if self._tz is None:
-        try:
+            timezone_str = self.config.get('timezone', 'UTC')
-            tz = pytz.timezone(timezone_str)
+            try:
-        except pytz.UnknownTimeZoneError:
+                self._tz = pytz.timezone(timezone_str)
-            logger.warning(f"Unknown timezone '{timezone_str}' in dim schedule, using UTC")
+            except pytz.UnknownTimeZoneError:
-            tz = pytz.UTC
+                logger.warning("Unknown timezone '%s' in dim schedule, using UTC", timezone_str)
+                self._tz = pytz.UTC
+
+        current_time = datetime.now(self._tz)
+        current_minute_key = (current_time.hour, current_time.minute)
+        if current_minute_key == self._dim_checked_minute:
+            return self._cached_target_brightness
+        self._dim_checked_minute = current_minute_key
 
-        current_time = datetime.now(tz)
         current_day = current_time.strftime('%A').lower()
         current_time_only = current_time.time()
 
@@ -703,10 +794,12 @@ class DisplayController:
                 logger.info(f"Dim schedule deactivated: brightness restored to {target_brightness}%")
 
             self._was_dimmed = self.is_dimmed
+            self._cached_target_brightness = target_brightness  # persist for minute-gate
             return target_brightness
 
         except ValueError as e:
-            logger.warning(f"Invalid dim schedule time format: {e}")
+            logger.warning("Invalid dim schedule time format: %s", e)
+            self._cached_target_brightness = normal_brightness  # persist for minute-gate
             return normal_brightness
 
     def _update_modules(self):
@@ -1382,6 +1475,36 @@ class DisplayController:
         except Exception as e:
             logger.debug(f"Error logging memory stats: {e}")
 
+    def _apply_live_priority(self, live_priority_mode):
+        """Switch to a live-priority mode, or resume rotation when it ends.
+
+        When a live-priority plugin preempts the rotation, the position the
+        rotation had reached is saved so that, once live priority ends, the
+        rotation resumes from there instead of continuing after the live
+        plugin's mode (which would skip every mode between the two). The save
+        happens only on the initial switch, not on each re-check while the
+        live hold continues.
+        """
+        if live_priority_mode:
+            if self.current_display_mode != live_priority_mode:
+                logger.info("Live content detected - switching immediately to %s", live_priority_mode)
+                if self._live_resume_index is None:
+                    self._live_resume_index = self.current_mode_index
+                self.current_display_mode = live_priority_mode
+                self.force_change = True
+                # Update mode index to match the new mode
+                try:
+                    self.current_mode_index = self.available_modes.index(live_priority_mode)
+                except ValueError:
+                    pass
+        elif self._live_resume_index is not None and self.available_modes:
+            # Live priority ended — resume rotation where it was interrupted.
+            self.current_mode_index = self._live_resume_index % len(self.available_modes)
+            self.current_display_mode = self.available_modes[self.current_mode_index]
+            self.force_change = True
+            logger.info("Live priority ended - resuming rotation at %s", self.current_display_mode)
+            self._live_resume_index = None
+
     def _check_live_priority(self):
         """
         Check all plugins for live priority content.
@@ -1483,12 +1606,8 @@ class DisplayController:
                     rp = vc.render_pipeline if (vc and vc.render_pipeline) else None
                     width = self.display_manager.width
 
-                    # Advance local position at Vegas scroll speed (px/s → px/tick)
+                    # Opt #2: use pre-cached scroll speed (constant for the run)
-                    vegas_speed = (
+                    vegas_speed = self._scroll_speed
-                        self.config.get('display', {})
-                            .get('vegas_scroll', {})
-                            .get('scroll_speed', 75)
-                    )
                     local_x = getattr(self, '_follower_local_x', None)
                     if local_x is None:
                         local_x = float(width)  # safe start (past pre-roll guard)
@@ -1573,15 +1692,7 @@ class DisplayController:
                 # Check for live priority content and switch to it immediately
                 if not self.on_demand_active and not wifi_status_data:
                     live_priority_mode = self._check_live_priority()
-                    if live_priority_mode and self.current_display_mode != live_priority_mode:
+                    self._apply_live_priority(live_priority_mode)
-                        logger.info("Live content detected - switching immediately to %s", live_priority_mode)
-                        self.current_display_mode = live_priority_mode
-                        self.force_change = True
-                        # Update mode index to match the new mode
-                        try:
-                            self.current_mode_index = self.available_modes.index(live_priority_mode)
-                        except ValueError:
-                            pass
 
                 # Vegas scroll mode - continuous ticker across all plugins
                 # Priority: on-demand > wifi-status > live-priority > vegas > normal rotation
@@ -1628,7 +1739,8 @@ class DisplayController:
 
                 manager_to_display = None
                 
-                logger.info(f"Processing mode: {active_mode}, available_modes: {len(self.available_modes)}, plugin_modes: {list(self.plugin_modes.keys())}")
+                logger.info("Processing mode: %s (%d available)", active_mode, len(self.available_modes))
+                logger.debug("Loaded plugin modes: %s", list(self.plugin_modes.keys()))
                 
                 # Handle plugin-based display modes
                 if active_mode in self.plugin_modes:
@@ -1664,17 +1776,22 @@ class DisplayController:
                     try:
                         logger.debug(f"Calling display() for {active_mode} with force_clear={self.force_change}")
                         if hasattr(manager_to_display, 'display'):
-                            # Check if plugin accepts display_mode parameter
+                            # Opt #1: look up (or compute once) whether display() accepts display_mode
-                            import inspect
+                            _cache_key = plugin_id
-                            sig = inspect.signature(manager_to_display.display)
+                            if _cache_key not in self._plugin_accepts_display_mode:
-                            
+                                import inspect as _inspect
+                                self._plugin_accepts_display_mode[_cache_key] = (
+                                    'display_mode' in _inspect.signature(manager_to_display.display).parameters
+                                )
+                            _accepts_display_mode = self._plugin_accepts_display_mode[_cache_key]
+
                             # Use PluginExecutor for safe execution with timeout
                             if self.plugin_manager and hasattr(self.plugin_manager, 'plugin_executor'):
                                 result = self.plugin_manager.plugin_executor.execute_display(
                                     manager_to_display,
                                     plugin_id,
                                     force_clear=self.force_change,
-                                    display_mode=active_mode if 'display_mode' in sig.parameters else None
+                                    display_mode=active_mode if _accepts_display_mode else None
                                 )
                                 # execute_display returns bool, convert to expected format
                                 if result:
@@ -1683,7 +1800,7 @@ class DisplayController:
                                     result = False  # Failed
                             else:
                                 # Fallback to direct call if executor not available
-                                if 'display_mode' in sig.parameters:
+                                if _accepts_display_mode:
                                     result = manager_to_display.display(display_mode=active_mode, force_clear=self.force_change)
                                 else:
                                     result = manager_to_display.display(force_clear=self.force_change)
@@ -1820,9 +1937,9 @@ class DisplayController:
                     min_duration = base_duration
                     if dynamic_enabled:
                         # Try to get plugin-calculated cycle duration first
-                        logger.info("Attempting to get cycle duration for mode %s", active_mode)
+                        logger.debug("Attempting to get cycle duration for mode %s", active_mode)
                         plugin_cycle_duration = self._plugin_cycle_duration(manager_to_display, active_mode)
-                        logger.info("Got cycle duration: %s", plugin_cycle_duration)
+                        logger.debug("Got cycle duration: %s", plugin_cycle_duration)
                         
                         # Get caps for validation
                         plugin_cap = self._plugin_dynamic_cap(manager_to_display)
@@ -1962,7 +2079,7 @@ class DisplayController:
                         if needs_high_fps:
                             # Ultra-smooth FPS for scrolling plugins (8ms = 125 FPS)
                             display_interval = 0.008
-                            logger.info(
+                            logger.debug(
                                 "Entering high-FPS loop for %s with display_interval=%.3fs (%.1f FPS)",
                                 active_mode,
                                 display_interval,
@@ -1972,7 +2089,7 @@ class DisplayController:
                             while True:
                                 try:
                                     # Pass display_mode to maintain sticky manager state
-                                    if 'display_mode' in sig.parameters:
+                                    if _accepts_display_mode:
                                         result = manager_to_display.display(display_mode=active_mode, force_clear=False)
                                     else:
                                         result = manager_to_display.display(force_clear=False)
@@ -2014,7 +2131,7 @@ class DisplayController:
                         else:
                             # Normal FPS for other plugins (1 second)
                             display_interval = 1.0
-                            logger.info(
+                            logger.debug(
                                 "Entering normal FPS loop for %s with display_interval=%.3fs",
                                 active_mode,
                                 display_interval
@@ -2036,7 +2153,7 @@ class DisplayController:
 
                                 try:
                                     # Pass display_mode to maintain sticky manager state
-                                    if 'display_mode' in sig.parameters:
+                                    if _accepts_display_mode:
                                         result = manager_to_display.display(display_mode=active_mode, force_clear=False)
                                     else:
                                         result = manager_to_display.display(force_clear=False)
@@ -2333,6 +2450,30 @@ class DisplayController:
             self.wifi_status_active = False
             self.wifi_status_expires_at = None
 
+    def _refresh_config_cache(self, new_config: Dict[str, Any]) -> None:
+        """Refresh all config-derived caches when a hot-reload fires.
+
+        Called by the controller-level ConfigService subscriber.  Keeps
+        ``_normal_brightness``, ``_scroll_speed``, the cached timezone, and the
+        schedule minute-gates consistent with the live config so callers never
+        read stale values after the user saves settings via the web UI.
+        """
+        self.config = new_config
+        self._normal_brightness = (
+            self.config.get('display', {}).get('hardware', {}).get('brightness', 90)
+        )
+        self._scroll_speed = (
+            self.config.get('display', {}).get('vegas_scroll', {}).get('scroll_speed', 75)
+        )
+        # Force the timezone to be re-derived from the new config on next schedule check
+        self._tz = None
+        # Invalidate minute-gates so the new schedule/dim times take effect immediately
+        self._schedule_checked_minute = None
+        self._dim_checked_minute = None
+        self._cached_target_brightness = self._normal_brightness
+        logger.debug("Config cache refreshed (brightness=%s, scroll_speed=%s)",
+                     self._normal_brightness, self._scroll_speed)
+
     def cleanup(self):
         """Clean up resources."""
         # Shutdown config service if it exists
@@ -2347,6 +2488,7 @@ class DisplayController:
         logger.info("Cleanup complete.")
 
 def main():
+    """Application entry point — create a DisplayController and run until interrupted."""
     controller = DisplayController()
     controller.run()
 

src/display_manager.py

@@ -1,3 +1,28 @@
+"""
+Display Manager — hardware abstraction layer for the RGB LED matrix.
+
+This module provides :class:`DisplayManager`, the single interface between
+application code and the physical (or emulated) LED panel.
+
+Key responsibilities
+--------------------
+* Initialise the ``RGBMatrix`` (hardware) or ``RGBMatrixEmulator`` depending
+  on the ``EMULATOR`` environment variable.
+* Expose a PIL ``Image``/``ImageDraw`` canvas that plugins draw into, then
+  flush it to the matrix via double-buffering (:meth:`DisplayManager.update_display`).
+* Load and cache TTF/BDF fonts; expose ``draw_text`` for consistent text rendering.
+* Provide ``width`` / ``height`` properties — always use these instead of
+  hard-coding display dimensions.
+* Write periodic PNG snapshots to ``/tmp/led_matrix_preview.png`` for the
+  web-interface live preview.
+* Track scrolling state and gate deferred updates so plugins don't race with
+  an in-progress scroll.
+
+Singleton: only one ``DisplayManager`` instance exists per process.  The
+first call to ``DisplayManager(config)`` creates it; subsequent calls return
+the same object.
+"""
+
 import json
 import os
 import tempfile
@@ -18,6 +43,24 @@ logger = logging.getLogger(__name__)
 logger.setLevel(logging.INFO)  # Set to INFO level
 
 class DisplayManager:
+    """
+    Singleton hardware abstraction layer for the RGB LED matrix.
+
+    Plugins should never interact with ``RGBMatrix`` directly; they use this
+    class to draw content and call :meth:`update_display` to push frames to
+    the panel.
+
+    Typical plugin usage::
+
+        canvas = Image.new('RGB', (self.display_manager.width,
+                                   self.display_manager.height), (0, 0, 0))
+        draw = ImageDraw.Draw(canvas)
+        # ... draw content ...
+        self.display_manager.image = canvas
+        self.display_manager.draw = ImageDraw.Draw(self.display_manager.image)
+        self.display_manager.update_display()
+    """
+
     _instance = None
     _initialized = False
 
@@ -33,6 +76,10 @@ class DisplayManager:
         self._suppress_test_pattern = suppress_test_pattern
         # When True, update_display() and clear() skip hardware writes (used during off-screen content capture)
         self._capture_mode_active = False
+        # Text-width measurement cache: (text, id(font)) -> pixel_width
+        # Avoids re-measuring the same string+font on every display() call.
+        # Cleared on _load_fonts() so stale entries don't survive a font reload.
+        self._text_width_cache: Dict[tuple, int] = {}
         # Snapshot settings for web preview integration (service writes, web reads)
         self._snapshot_path = "/tmp/led_matrix_preview.png"  # nosec B108 - fixed path intentional; web UI reads same path
         self._snapshot_min_interval_sec = 0.2  # max ~5 fps
@@ -437,6 +484,9 @@ class DisplayManager:
 
     def _load_fonts(self):
         """Load fonts with proper error handling."""
+        # Font objects get new id()s after reload, so the text-width cache would
+        # return stale measurements keyed on the old ids.  Clear it here.
+        self._text_width_cache.clear()
         try:
             # Load Press Start 2P font
             self.regular_font = ImageFont.truetype("assets/fonts/PressStart2P-Regular.ttf", 8)
@@ -497,22 +547,32 @@ class DisplayManager:
 
 
     def get_text_width(self, text, font):
-        """Get the width of text when rendered with the given font."""
+        """Get the width of text when rendered with the given font.
+
+        Results are cached by (text, font identity) so plugins that measure
+        the same string every frame (e.g. to centre a score) pay only one
+        measurement per unique (text, font) pair.
+        """
+        cache_key = (text, id(font))
+        cached = self._text_width_cache.get(cache_key)
+        if cached is not None:
+            return cached
+
         try:
             if isinstance(font, freetype.Face):
-                # For FreeType faces, calculate width using freetype
                 width = 0
                 for char in text:
                     font.load_char(char)
                     width += font.glyph.advance.x >> 6
-                return width
             else:
-                # For PIL fonts, use textbbox
                 bbox = self.draw.textbbox((0, 0), text, font=font)
-                return bbox[2] - bbox[0]
+                width = bbox[2] - bbox[0]
-        except Exception as e:
+        except (AttributeError, TypeError, ValueError, OSError) as e:
-            logger.error(f"Error getting text width: {e}")
+            logger.error("Error getting text width: %s", e)
-            return 0  # Return 0 as fallback
+            return 0
+
+        self._text_width_cache[cache_key] = width
+        return width
 
     def get_font_height(self, font):
         """Get the height of the given font for line spacing purposes."""

src/font_manager.py

@@ -1,3 +1,30 @@
+"""
+Font Manager — TTF/BDF font loading, caching, and dynamic registration.
+
+:class:`FontManager` serves two purposes:
+
+1. **System fonts** — loads the configured small/medium/large TTF fonts (and
+   their BDF bitmap equivalents) at startup, caches metrics, and exposes them
+   via ``DisplayManager`` attributes (``small_font``, ``medium_font``, etc.).
+
+2. **Plugin fonts** — lets plugins register their own fonts at runtime via
+   :meth:`FontManager.register_manager_font` and resolve them later via
+   :meth:`FontManager.resolve_font`.  Registered fonts are namespaced by
+   plugin ID so they cannot collide.
+
+Font sources
+------------
+* Local paths relative to the project root.
+* Remote URLs — downloaded once, cached to disk, and never re-fetched while
+  the cached copy is fresh.
+
+BDF fallback
+------------
+Pixel-accurate LED fonts are stored as ``.bdf`` (Bitmap Distribution Format)
+files.  When PIL cannot measure BDF glyphs natively, ``freetype-py`` is used
+for accurate width/height calculations.
+"""
+
 import os
 import logging
 import freetype

src/plugin_system/plugin_manager.py

@@ -15,7 +15,7 @@ import threading
 from pathlib import Path
 from typing import Dict, List, Optional, Any
 import logging
-from src.exceptions import PluginError
+from src.exceptions import PluginError, ConfigError
 from src.logging_config import get_logger
 from src.plugin_system.plugin_loader import PluginLoader
 from src.plugin_system.plugin_executor import PluginExecutor
@@ -81,7 +81,13 @@ class PluginManager:
         self.plugin_manifests: Dict[str, Dict[str, Any]] = {}
         self.plugin_modules: Dict[str, Any] = {}
         self.plugin_last_update: Dict[str, float] = {}
-        
+
+        # Cached data-fetch intervals per plugin_id.
+        # _get_plugin_update_interval falls back to config_manager.get_config()
+        # (a full dict copy) when the manifest lacks an interval — caching avoids
+        # that copy on every 30-fps tick.  Cleared on load/unload.
+        self._update_interval_cache: Dict[str, Optional[float]] = {}
+
         # Health tracking (optional, set by display_controller if available)
         self.health_tracker = None
         self.resource_monitor = None
@@ -388,6 +394,8 @@ class PluginManager:
             # Store plugin instance
             self.plugins[plugin_id] = plugin_instance
             self.plugin_last_update[plugin_id] = 0.0
+            # Invalidate cached interval so next tick re-derives it for this plugin
+            self._update_interval_cache.pop(plugin_id, None)
             
             # Update state based on enabled status
             if config.get('enabled', True):
@@ -444,8 +452,8 @@ class PluginManager:
             
             # Remove from active plugins
             del self.plugins[plugin_id]
-            if plugin_id in self.plugin_last_update:
+            self.plugin_last_update.pop(plugin_id, None)
-                del self.plugin_last_update[plugin_id]
+            self._update_interval_cache.pop(plugin_id, None)
             
             # Remove main module from sys.modules if present
             module_name = f"plugin_{plugin_id.replace('-', '_')}"
@@ -639,41 +647,46 @@ class PluginManager:
 
     def _get_plugin_update_interval(self, plugin_id: str, plugin_instance: Any) -> Optional[float]:
         """
-        Get the update interval for a plugin.
+        Get the data-fetch interval for a plugin (seconds between update() calls).
-        
+
-        Args:
+        Result is cached per plugin_id after the first lookup to avoid calling
-            plugin_id: Plugin identifier
+        config_manager.get_config() — which returns a full dict copy — on every
-            plugin_instance: Plugin instance
+        tick of the 30-fps display loop.  The cache is invalidated when a plugin
-            
+        is loaded or unloaded.
-        Returns:
-            Update interval in seconds or None if not configured
         """
-        # Check manifest first
+        if plugin_id in self._update_interval_cache:
+            return self._update_interval_cache[plugin_id]
+
+        interval: Optional[float] = None
+
+        # 1. Manifest (immutable after load — preferred source)
         manifest = self.plugin_manifests.get(plugin_id, {})
-        update_interval = manifest.get('update_interval')
+        raw = manifest.get('update_interval')
-        
+        if raw is not None:
-        if update_interval:
             try:
-                return float(update_interval)
+                interval = float(raw)
             except (ValueError, TypeError):
                 pass
-        
+
-        # Check plugin config
+        # 2. Plugin config (mutable; only read once and then cached)
-        if self.config_manager:
+        if interval is None and self.config_manager:
             try:
                 config = self.config_manager.get_config()
-                plugin_config = config.get(plugin_id, {})
+                raw = config.get(plugin_id, {}).get('update_interval')
-                update_interval = plugin_config.get('update_interval')
+                if raw is not None:
-                if update_interval:
                     try:
-                        return float(update_interval)
+                        interval = float(raw)
                     except (ValueError, TypeError):
                         pass
-            except Exception as e:
+            except (ConfigError, OSError, ValueError, TypeError) as e:
                 self.logger.debug("Could not get update interval from config: %s", e)
-        
+
-        # Default: 60 seconds
+        # 3. Default
-        return 60.0
+        if interval is None:
+            interval = 60.0
+
+        self._update_interval_cache[plugin_id] = interval
+        return interval
 
     def _record_update_failure(
         self,

src/plugin_system/testing/__init__.py

@@ -7,13 +7,22 @@ Provides base classes and utilities for testing LEDMatrix plugins.
 from .plugin_test_base import PluginTestCase
 from .mocks import MockDisplayManager, MockCacheManager, MockConfigManager, MockPluginManager
 from .visual_display_manager import VisualTestDisplayManager
+from .bounds_display_manager import BoundsCheckingDisplayManager
+from .sizes import (
+    DEFAULT_TEST_SIZES, SUPPORTED_SIZES, resolve_test_sizes, size_label,
+)
 
 __all__ = [
     'PluginTestCase',
     'VisualTestDisplayManager',
+    'BoundsCheckingDisplayManager',
     'MockDisplayManager',
     'MockCacheManager',
     'MockConfigManager',
     'MockPluginManager',
+    'DEFAULT_TEST_SIZES',
+    'SUPPORTED_SIZES',
+    'resolve_test_sizes',
+    'size_label',
 ]
 

src/plugin_system/testing/bounds_display_manager.py

@@ -0,0 +1,129 @@
+"""
+Bounds-checking display manager.
+
+A VisualTestDisplayManager that draws onto an oversized canvas (the declared
+panel size plus a right/bottom margin) while still reporting the declared size
+to the plugin. Content that a plugin draws past the right or bottom edge lands
+in the margin instead of being silently clipped by PIL, so the harness can
+detect overflow — the classic symptom of hardcoded coordinates or fonts/icons
+that don't scale down to a smaller panel.
+
+Limitations (documented on purpose):
+- Overflow past the LEFT or TOP edge (negative coordinates) is still clipped by
+  PIL and not detected here. The dominant real-world breakage is content that is
+  too wide/tall for a smaller panel, which this catches.
+- BDF text is clipped to the declared bounds by the parent's bitmap drawer, so
+  BDF overflow is not flagged. Golden-image regression covers those plugins.
+- If a plugin replaces the canvas with its own image (display_manager.image = ...),
+  the margin can't be measured and overflow is reported as undetermined (None).
+"""
+
+from typing import Optional, Tuple
+
+from .sizes import DEFAULT_TEST_SIZES
+from .visual_display_manager import VisualTestDisplayManager, _MatrixProxy
+
+# Smallest extra band kept on the right/bottom so a few pixels of overflow are
+# still visible even on the largest panel in a run.
+_BASE_MARGIN = 16
+# Fallback overflow reference when a caller doesn't pass one: the largest shape
+# in the default sample. We extend every (smaller) canvas out to at least this
+# size so content drawn at a coordinate meant for a bigger build — e.g. x=200 on
+# a 64-wide panel — lands in the padded region and is flagged, instead of being
+# clipped off-canvas and read as a false pass.
+_DEFAULT_EXTENT_WIDTH = max(w for w, _ in DEFAULT_TEST_SIZES)
+_DEFAULT_EXTENT_HEIGHT = max(h for _, h in DEFAULT_TEST_SIZES)
+
+
+class BoundsCheckingDisplayManager(VisualTestDisplayManager):
+    """Detects drawing that overflows the declared panel size."""
+
+    # Kept for backwards compatibility; real padding is computed per-axis below.
+    MARGIN = _BASE_MARGIN
+
+    def __init__(self, width: int = 128, height: int = 32,
+                 overflow_extent: Optional[Tuple[int, int]] = None):
+        self._declared_width = int(width)
+        self._declared_height = int(height)
+        # Pad the canvas out to at least `overflow_extent` (the largest panel
+        # this run cares about) plus a base margin, so coordinates meant for a
+        # bigger build are caught — not clipped — when rendering a smaller panel.
+        # Defaults to the largest shape in the sample when no run is known.
+        ext_w, ext_h = overflow_extent or (_DEFAULT_EXTENT_WIDTH, _DEFAULT_EXTENT_HEIGHT)
+        self._canvas_width = max(self._declared_width, int(ext_w)) + _BASE_MARGIN
+        self._canvas_height = max(self._declared_height, int(ext_h)) + _BASE_MARGIN
+        # Parent builds the (oversized) backing canvas + fonts.
+        super().__init__(self._canvas_width, self._canvas_height)
+        # Plugins must see the DECLARED size, not the padded canvas size.
+        self.matrix = _MatrixProxy(self._declared_width, self._declared_height)
+
+    # -- declared dimensions (override parent's image-derived properties) --
+
+    @property
+    def width(self) -> int:
+        return self._declared_width
+
+    @property
+    def height(self) -> int:
+        return self._declared_height
+
+    @property
+    def display_width(self) -> int:
+        return self._declared_width
+
+    @property
+    def display_height(self) -> int:
+        return self._declared_height
+
+    # -- overflow detection --
+
+    def _canvas_is_padded(self) -> bool:
+        return self.image.size == (self._canvas_width, self._canvas_height)
+
+    def check_overflow(self) -> Optional[Tuple[int, int, int, int]]:
+        """Bounding box (in full-canvas coords) of any drawing beyond the
+        declared panel, or None if nothing overflowed / undetermined."""
+        if not self._canvas_is_padded():
+            return None
+
+        exp_w = self._canvas_width
+        exp_h = self._canvas_height
+        boxes = []
+
+        right = self.image.crop((self._declared_width, 0, exp_w, exp_h)).getbbox()
+        if right:
+            boxes.append((right[0] + self._declared_width, right[1],
+                          right[2] + self._declared_width, right[3]))
+
+        bottom = self.image.crop((0, self._declared_height, exp_w, exp_h)).getbbox()
+        if bottom:
+            boxes.append((bottom[0], bottom[1] + self._declared_height,
+                          bottom[2], bottom[3] + self._declared_height))
+
+        if not boxes:
+            return None
+        return (
+            min(b[0] for b in boxes), min(b[1] for b in boxes),
+            max(b[2] for b in boxes), max(b[3] for b in boxes),
+        )
+
+    # -- snapshot/image accessors return the cropped, true-panel image --
+
+    def declared_image(self):
+        """The visible panel: the canvas cropped to the declared size."""
+        if self._canvas_is_padded():
+            return self.image.crop((0, 0, self._declared_width, self._declared_height))
+        return self.image
+
+    def save_snapshot(self, path: str) -> None:
+        self.declared_image().save(path, format='PNG')
+
+    def get_image(self):
+        return self.declared_image()
+
+    def get_image_base64(self) -> str:
+        import base64
+        import io
+        buffer = io.BytesIO()
+        self.declared_image().save(buffer, format='PNG')
+        return base64.b64encode(buffer.getvalue()).decode('utf-8')

src/plugin_system/testing/harness.py

@@ -0,0 +1,314 @@
+"""
+Plugin safety harness.
+
+Renders a plugin across every declared screen (mode) and every supported matrix
+size, capturing crashes and overflow. Used by scripts/check_plugin.py and the
+pytest matrix test to guarantee a plugin change doesn't break a screen at a size
+the author didn't try.
+
+The render flow mirrors scripts/render_plugin.py (same PluginLoader call), but
+this module adds: multi-size iteration, per-mode rendering, overflow detection
+via BoundsCheckingDisplayManager, and golden-image comparison.
+"""
+
+import contextlib
+import http.client
+import inspect
+import socket
+import ssl
+import urllib.error
+from dataclasses import dataclass
+from pathlib import Path
+from typing import Any, Dict, List, Optional, Tuple
+
+from PIL import Image, ImageChops
+
+from src.logging_config import get_logger
+from .bounds_display_manager import BoundsCheckingDisplayManager
+from .loading import load_config_defaults, load_manifest
+from .sizes import DEFAULT_TEST_SIZES, safe_mode_filename, size_label
+
+logger = get_logger("[Plugin Harness]")
+
+
+def _tolerated_update_errors() -> Tuple[type, ...]:
+    """Exception types from update() we treat as a tolerated no-connectivity
+    failure (expected in CI / headless dev) rather than a real plugin bug.
+
+    Anything NOT in this set is a genuine regression — a plugin that lets a
+    non-network exception escape update() should fail the harness, not pass
+    green because display() happened to survive.
+    """
+    types: List[type] = [
+        ConnectionError, TimeoutError,        # builtins
+        socket.gaierror, socket.timeout,      # DNS / socket timeouts
+        ssl.SSLError,
+        urllib.error.URLError,
+        http.client.HTTPException,
+    ]
+    try:  # requests is optional; cover its whole error tree when present
+        import requests
+        types.append(requests.exceptions.RequestException)
+    except ImportError:  # pragma: no cover - requests not installed
+        logger.debug("requests not installed; its connectivity errors won't be specifically tolerated")
+    return tuple(types)
+
+
+_TOLERATED_UPDATE_ERRORS = _tolerated_update_errors()
+
+
+@dataclass
+class RenderResult:
+    """Outcome of rendering one (size, mode) of a plugin."""
+    plugin_id: str
+    width: int
+    height: int
+    mode: str
+    image: Optional[Image.Image] = None
+    error: Optional[str] = None          # fatal: load/display crash, or a non-network update() error
+    update_error: Optional[str] = None   # tolerated: connectivity error from update() (no network in CI)
+    overflow: Optional[Tuple[int, int, int, int]] = None  # bbox past the panel
+    # golden comparison (populated only when a golden was provided)
+    golden_checked: bool = False
+    golden_ok: Optional[bool] = None
+    golden_diff_pixels: int = 0
+    golden_max_delta: int = 0
+
+    @property
+    def size_label(self) -> str:
+        return size_label(self.width, self.height)
+
+    @property
+    def ok(self) -> bool:
+        """Phase-1 pass: rendered without crashing and without overflow, and if a
+        golden was checked it matched."""
+        if self.error is not None or self.overflow is not None:
+            return False
+        if self.golden_checked and self.golden_ok is False:
+            return False
+        return True
+
+
+def list_modes(plugin_instance: Any, manifest: Dict[str, Any], plugin_id: str) -> List[str]:
+    """Enumerate a plugin's screens: instance.modes wins, then manifest
+    display_modes, then the plugin id as a single mode."""
+    modes = getattr(plugin_instance, "modes", None)
+    if modes:
+        return [str(m) for m in modes]
+    declared = manifest.get("display_modes")
+    if declared:
+        return [str(m) for m in declared]
+    return [plugin_id]
+
+
+def _instantiate(plugin_id: str, manifest: Dict[str, Any], plugin_dir: Path,
+                 config: Dict[str, Any], mock_data: Dict[str, Any],
+                 display_manager: Any) -> Any:
+    """Load and construct a plugin instance with mocked managers."""
+    from src.plugin_system.plugin_loader import PluginLoader
+    from src.plugin_system.testing import MockCacheManager, MockPluginManager
+
+    cache_manager = MockCacheManager()
+    for key, value in (mock_data or {}).items():
+        cache_manager.set(key, value)
+
+    loader = PluginLoader()
+    plugin_instance, _module = loader.load_plugin(
+        plugin_id=plugin_id,
+        manifest=manifest,
+        plugin_dir=plugin_dir,
+        config=config,
+        display_manager=display_manager,
+        cache_manager=cache_manager,
+        plugin_manager=MockPluginManager(),
+        install_deps=False,
+    )
+    return plugin_instance
+
+
+def _render_mode(plugin_instance: Any, mode: str) -> None:
+    """Render a specific screen. Prefer an explicit display_mode kwarg; otherwise
+    drive the plugin's internal mode state machine (first display() call renders
+    modes[current_mode_index] when current_display_mode is None)."""
+    sig = inspect.signature(plugin_instance.display)
+    if "display_mode" in sig.parameters:
+        plugin_instance.display(force_clear=True, display_mode=mode)
+        return
+
+    modes = getattr(plugin_instance, "modes", None)
+    if modes and mode in modes:
+        plugin_instance.current_mode_index = list(modes).index(mode)
+    if hasattr(plugin_instance, "current_display_mode"):
+        plugin_instance.current_display_mode = None
+    plugin_instance.display(force_clear=False)
+
+
+def _freeze(freeze_time: Optional[str]):
+    """Context manager that freezes wall-clock time when freeze_time is given,
+    so time-dependent plugins (clocks, countdowns) render deterministic goldens."""
+    if not freeze_time:
+        return contextlib.nullcontext()
+    try:
+        from freezegun import freeze_time as _ft
+    except ImportError as e:  # pragma: no cover - only hit without the dep
+        raise RuntimeError(
+            "freeze_time requires the 'freezegun' package (pip install freezegun)"
+        ) from e
+    return _ft(freeze_time)
+
+
+def render_plugin_matrix(
+    plugin_id: str,
+    plugin_dir: Path,
+    config: Optional[Dict[str, Any]] = None,
+    mock_data: Optional[Dict[str, Any]] = None,
+    sizes: Optional[List[Tuple[int, int]]] = None,
+    run_update: bool = True,
+    freeze_time: Optional[str] = None,
+) -> List[RenderResult]:
+    """Render every (size, mode) combination for a plugin.
+
+    Returns a flat list of RenderResult. A fresh plugin instance is built per
+    (size, mode) so state never leaks between screens. Pass freeze_time (e.g.
+    "2025-08-01 15:25:00") to make time-dependent plugins reproducible.
+    """
+    plugin_dir = Path(plugin_dir)
+    manifest = load_manifest(plugin_dir)
+    # Start from config_schema.json defaults so the plugin behaves like a real
+    # install; explicit caller config still wins over a schema default.
+    config = {"enabled": True, **load_config_defaults(plugin_dir), **(config or {})}
+    sizes = sizes or DEFAULT_TEST_SIZES
+    results: List[RenderResult] = []
+
+    # The largest panel in this run. Every (smaller) canvas is padded out to it
+    # so a coordinate meant for the biggest configuration is still caught when
+    # rendering a smaller one, instead of being clipped into a false pass.
+    extent = (max(w for w, _ in sizes), max(h for _, h in sizes))
+
+    with _freeze(freeze_time):
+        for width, height in sizes:
+            results.extend(_render_size(
+                plugin_id, manifest, plugin_dir, config, mock_data or {},
+                width, height, run_update, extent,
+            ))
+
+    return results
+
+
+def _render_size(plugin_id, manifest, plugin_dir, config, mock_data,
+                 width, height, run_update, extent) -> List[RenderResult]:
+    """Render every mode at one size. A fresh instance per mode avoids state leaks."""
+    results: List[RenderResult] = []
+
+    # Discover modes once per size (instance build can depend on config).
+    try:
+        probe_dm = BoundsCheckingDisplayManager(width=width, height=height, overflow_extent=extent)
+        probe = _instantiate(plugin_id, manifest, plugin_dir, config, mock_data, probe_dm)
+        modes = list_modes(probe, manifest, plugin_id)
+    except Exception as e:  # noqa: BLE001 — surface any load failure as a result
+        return [RenderResult(plugin_id, width, height, "<load>", error=repr(e))]
+
+    for mode in modes:
+        result = RenderResult(plugin_id, width, height, mode)
+        dm = BoundsCheckingDisplayManager(width=width, height=height, overflow_extent=extent)
+        try:
+            inst = _instantiate(plugin_id, manifest, plugin_dir, config, mock_data, dm)
+            if run_update:
+                try:
+                    inst.update()
+                except _TOLERATED_UPDATE_ERRORS as e:
+                    # Expected when CI / headless dev has no network: record it
+                    # (surfaced in the report) but don't fail the run.
+                    result.update_error = repr(e)
+                    logger.debug("update() connectivity error for %s [%s]: %s", plugin_id, mode, e)
+                except Exception as e:  # noqa: BLE001 — a non-network update() failure is a real bug
+                    # A regression in update() must not pass green just because
+                    # display() survives, so treat it as a failure of this render.
+                    result.error = repr(e)
+                    logger.warning("update() raised a non-connectivity error for %s [%s]: %s",
+                                   plugin_id, mode, e)
+            if result.error is None:
+                _render_mode(inst, mode)
+                result.image = dm.get_image()
+                result.overflow = dm.check_overflow()
+        except Exception as e:  # noqa: BLE001 — a display crash is a real failure
+            result.error = repr(e)
+        results.append(result)
+
+    return results
+
+
+# ---------------------------------------------------------------------------
+# Golden-image comparison
+# ---------------------------------------------------------------------------
+
+def compare_images(rendered: Image.Image, golden: Image.Image,
+                   max_delta: int = 0, max_diff_pixels: int = 0) -> Tuple[bool, int, int]:
+    """Compare two images. Returns (ok, diff_pixel_count, max_per_channel_delta).
+
+    Tolerances default to exact match; bump them only to absorb known platform
+    anti-aliasing noise (requires a pinned Pillow + bundled fonts for stability).
+    """
+    if rendered.size != golden.size:
+        return False, rendered.size[0] * rendered.size[1], 255
+    a = rendered.convert("RGB")
+    b = golden.convert("RGB")
+    diff = ImageChops.difference(a, b)
+    bbox = diff.getbbox()
+    if bbox is None:
+        return True, 0, 0
+    # Count pixels whose largest per-channel delta exceeds the allowed tolerance,
+    # and track the worst delta seen (for reporting).
+    diff_pixels = 0
+    observed_max = 0
+    for px in diff.crop(bbox).getdata():
+        m = max(px) if isinstance(px, tuple) else px
+        if m > observed_max:
+            observed_max = m
+        if m > max_delta:
+            diff_pixels += 1
+    # Pass when the number of out-of-tolerance pixels is within budget.
+    ok = diff_pixels <= max_diff_pixels
+    return ok, diff_pixels, observed_max
+
+
+def golden_path(golden_dir: Path, width: int, height: int, mode: str) -> Path:
+    """Location of a golden image: <golden_dir>/<WxH>/<mode>.png.
+
+    The mode is sanitized to a safe basename so a mode name with '/' or '..'
+    can't read or write outside the golden directory.
+    """
+    return Path(golden_dir) / size_label(width, height) / f"{safe_mode_filename(mode)}.png"
+
+
+def compare_to_goldens(results: List[RenderResult], golden_dir: Path,
+                       max_delta: int = 0, max_diff_pixels: int = 0) -> List[RenderResult]:
+    """Compare rendered results against committed goldens, mutating each result's
+    golden_* fields. Results with no golden file on disk are left unchecked."""
+    for r in results:
+        if r.image is None:
+            continue
+        gp = golden_path(golden_dir, r.width, r.height, r.mode)
+        if not gp.exists():
+            continue
+        r.golden_checked = True
+        with Image.open(gp) as g:
+            ok, diff_pixels, observed_max = compare_images(
+                r.image, g, max_delta=max_delta, max_diff_pixels=max_diff_pixels)
+        r.golden_ok = ok
+        r.golden_diff_pixels = diff_pixels
+        r.golden_max_delta = observed_max
+    return results
+
+
+def write_goldens(results: List[RenderResult], golden_dir: Path) -> int:
+    """Write each successfully-rendered result to its golden path. Returns count."""
+    written = 0
+    for r in results:
+        if r.image is None or r.error is not None:
+            continue
+        gp = golden_path(golden_dir, r.width, r.height, r.mode)
+        gp.parent.mkdir(parents=True, exist_ok=True)
+        r.image.save(gp, format="PNG")
+        written += 1
+    return written

src/plugin_system/testing/loading.py

@@ -0,0 +1,82 @@
+"""
+Shared helpers for loading a plugin headlessly.
+
+Used by scripts/render_plugin.py, scripts/check_plugin.py, and the harness so
+plugin discovery / manifest / config-default logic lives in exactly one place.
+"""
+
+import json
+from pathlib import Path
+from typing import Any, Dict, Optional, Sequence, Union
+
+
+def find_plugin_dir(plugin_id: str, search_dirs: Sequence[Union[str, Path]]) -> Optional[Path]:
+    """Find a plugin directory by searching multiple paths."""
+    from src.plugin_system.plugin_loader import PluginLoader
+    loader = PluginLoader()
+    for search_dir in search_dirs:
+        search_path = Path(search_dir)
+        if not search_path.exists():
+            continue
+        result = loader.find_plugin_directory(plugin_id, search_path)
+        if result:
+            return Path(result)
+    return None
+
+
+def load_manifest(plugin_dir: Union[str, Path]) -> Dict[str, Any]:
+    """Load and return manifest.json from a plugin directory."""
+    manifest_path = Path(plugin_dir) / 'manifest.json'
+    if not manifest_path.exists():
+        raise FileNotFoundError(f"No manifest.json in {plugin_dir}")
+    with open(manifest_path, 'r') as f:
+        return json.load(f)
+
+
+def load_config_defaults(plugin_dir: Union[str, Path]) -> Dict[str, Any]:
+    """Extract default values from a plugin's config_schema.json (empty if none)."""
+    schema_path = Path(plugin_dir) / 'config_schema.json'
+    if not schema_path.exists():
+        return {}
+    with open(schema_path, 'r') as f:
+        schema = json.load(f)
+    defaults: Dict[str, Any] = {}
+    for key, prop in schema.get('properties', {}).items():
+        if isinstance(prop, dict) and 'default' in prop:
+            defaults[key] = prop['default']
+    return defaults
+
+
+def load_harness_spec(plugin_dir: Union[str, Path]) -> Dict[str, Any]:
+    """Optional per-plugin harness settings from <plugin>/test/harness.json.
+
+    Lets a plugin opt into golden-image testing by declaring how to render it
+    deterministically. All keys optional:
+        {
+          "config":     {...},            # config overrides
+          "mock_data":  "fixtures/mock.json",  # path (relative to plugin dir) to cache fixtures
+          "freeze_time": "2025-08-01 15:25:00",
+          "skip_update": false
+        }
+    Returns {} when no harness.json exists.
+    """
+    spec_path = Path(plugin_dir) / 'test' / 'harness.json'
+    if not spec_path.exists():
+        return {}
+    with open(spec_path, 'r') as f:
+        spec = json.load(f)
+
+    # Resolve mock_data path and inline its contents for convenience.
+    mock_rel = spec.get('mock_data')
+    if mock_rel:
+        mock_path = Path(plugin_dir) / mock_rel
+        if not mock_path.exists():
+            # A declared-but-missing fixture is a harness config error: failing
+            # loudly beats silently rendering the plugin with no mock data.
+            raise FileNotFoundError(
+                f"harness.json references mock_data '{mock_rel}' but "
+                f"{mock_path} does not exist"
+            )
+        with open(mock_path, 'r') as mf:
+            spec['mock_data_contents'] = json.load(mf)
+    return spec

src/plugin_system/testing/mocks.py

@@ -63,11 +63,23 @@ class MockCacheManager:
     """Mock cache manager for testing."""
     
     def __init__(self):
+        import shutil
+        import tempfile
+        import weakref
         self._cache: Dict[str, Any] = {}
         self._cache_timestamps: Dict[str, float] = {}
         self.get_calls = []
         self.set_calls = []
         self.delete_calls = []
+        # Real temp dir for plugins that write/read files under cache_dir.
+        # Registered for cleanup so each mock instance doesn't leak a tmp dir.
+        self.cache_dir = tempfile.mkdtemp(prefix="ledmatrix-mock-cache-")
+        self._finalizer = weakref.finalize(
+            self, shutil.rmtree, self.cache_dir, ignore_errors=True)
+
+    def cleanup(self) -> None:
+        """Remove the temp cache directory created for this instance."""
+        self._finalizer()
     
     def get(self, key: str, max_age: Optional[float] = None) -> Optional[Any]:
         """Get a value from cache."""

src/plugin_system/testing/sizes.py

@@ -0,0 +1,120 @@
+"""
+LED matrix sizes the plugin safety harness renders against.
+
+There is no fixed set of "supported" panel sizes — an RGB matrix build can be
+any width/height and configuration (square, rectangle, 2x2, 4x4, 8x2, long
+strips, tall stacks, ...). Plugins are expected to read width/height
+dynamically and lay themselves out accordingly, so the harness's job is to
+prove a plugin survives a *spread* of shapes, not a canonical list.
+
+`DEFAULT_TEST_SIZES` is therefore a representative SAMPLE chosen to span the
+axes of variation (narrow, wide, square, tall, small, long), not an
+exhaustive or authoritative list. Callers can override it entirely:
+
+  - CLI:        scripts/check_plugin.py --sizes 8x16,64x64,256x32
+  - pytest:     LEDMATRIX_TEST_SIZES="8x16,64x64" env var (all plugins), or
+                per-plugin test/harness.json {"sizes": [[8, 16], [64, 64]]}
+
+so anyone can point the harness at the exact panel(s) their build uses.
+"""
+
+import os
+from typing import Iterable, List, Optional, Sequence, Tuple, Union
+
+# A spread of real panel-grid arrangements (each module is 64x32), not a list of
+# "blessed" sizes. Each entry exercises a different layout assumption a plugin
+# might accidentally bake in. Annotations are the panel grid (cols x rows).
+DEFAULT_TEST_SIZES: List[Tuple[int, int]] = [
+    (64, 32),    # 1x1 — single panel, the tightest common rectangle
+    (128, 32),   # 2x1 — the baseline most plugins are tuned for
+    (64, 64),    # 1x2 — stacked, exercises tall-narrow centering
+    (128, 64),   # 2x2 — block, icon scaling / vertical centering
+    (256, 32),   # 4x1 — long strip, wide horizontal layout
+    (128, 96),   # 2x3 — tall, exercises vertical overflow
+    (256, 128),  # 4x4 — large block, both dimensions big at once
+]
+
+# Backwards-compatible alias. Prefer DEFAULT_TEST_SIZES in new code — the old
+# name implied these were the only valid panel sizes, which they are not.
+SUPPORTED_SIZES = DEFAULT_TEST_SIZES
+
+
+def size_label(width: int, height: int) -> str:
+    """Human/path-friendly label for a size, e.g. '128x32'."""
+    return f"{width}x{height}"
+
+
+def parse_size_token(token: str) -> Tuple[int, int]:
+    """Parse a single 'WxH' token into an (int, int) pair.
+
+    Raises ValueError (with a user-friendly message) on malformed input so
+    callers can surface it however they like.
+    """
+    cleaned = token.strip().lower()
+    if "x" not in cleaned:
+        raise ValueError(f"Invalid size '{token}' (expected WxH, e.g. 128x32)")
+    w, h = cleaned.split("x", 1)
+    try:
+        width, height = int(w), int(h)
+    except ValueError as exc:
+        raise ValueError(
+            f"Invalid size '{token}' (expected numeric WxH, e.g. 128x32)"
+        ) from exc
+    if width <= 0 or height <= 0:
+        raise ValueError(
+            f"Invalid size '{token}' (width and height must be positive, e.g. 128x32)"
+        )
+    return (width, height)
+
+
+def coerce_sizes(
+    value: Union[str, Iterable[Sequence[int]], None]
+) -> Optional[List[Tuple[int, int]]]:
+    """Normalize a size spec into a list of (w, h) tuples, or None if empty.
+
+    Accepts a comma-separated 'WxH,WxH' string (CLI / env var) or an iterable
+    of [w, h] / (w, h) pairs (harness.json). Returns None when value is falsy
+    so callers can fall back to the default sample.
+    """
+    if not value:
+        return None
+    if isinstance(value, str):
+        return [parse_size_token(tok) for tok in value.split(",") if tok.strip()]
+    sizes: List[Tuple[int, int]] = []
+    for pair in value:
+        w, h = pair  # raises if not a 2-element sequence
+        width, height = int(w), int(h)
+        if width <= 0 or height <= 0:
+            raise ValueError(f"Invalid size pair {pair!r} (width and height must be positive)")
+        sizes.append((width, height))
+    return sizes or None
+
+
+def resolve_test_sizes(
+    spec_sizes: Union[str, Iterable[Sequence[int]], None] = None,
+) -> List[Tuple[int, int]]:
+    """Decide which sizes to render, by precedence:
+
+    1. LEDMATRIX_TEST_SIZES env var — a global "test on my hardware" override
+       that wins for every plugin.
+    2. spec_sizes — e.g. a per-plugin harness.json "sizes" list.
+    3. DEFAULT_TEST_SIZES — the representative sample.
+    """
+    env = coerce_sizes(os.environ.get("LEDMATRIX_TEST_SIZES"))
+    if env:
+        return env
+    spec = coerce_sizes(spec_sizes)
+    if spec:
+        return spec
+    return list(DEFAULT_TEST_SIZES)
+
+
+def safe_mode_filename(mode: str) -> str:
+    """A filesystem-safe basename for a plugin mode.
+
+    Mode names come from plugin metadata/render state, so a value containing
+    '/' or '..' could otherwise escape the intended output directory. Collapse
+    anything that isn't alphanumeric / dash / underscore to '_'.
+    """
+    cleaned = "".join(ch if ch.isalnum() or ch in ("-", "_") else "_" for ch in mode)
+    return cleaned or "mode"

test/plugins/test_basketball_scoreboard.py

@@ -49,9 +49,10 @@ class TestBasketballScoreboardPlugin(PluginTestBase):
         """Test that plugin has display modes."""
         manifest = self.load_plugin_manifest(plugin_id)
         assert 'display_modes' in manifest
-        assert 'basketball_live' in manifest['display_modes']
+        # Manifest uses league-prefixed modes (nba_, wnba_, ncaam_, ncaaw_)
-        assert 'basketball_recent' in manifest['display_modes']
+        assert 'nba_live' in manifest['display_modes']
-        assert 'basketball_upcoming' in manifest['display_modes']
+        assert 'nba_recent' in manifest['display_modes']
+        assert 'nba_upcoming' in manifest['display_modes']
     
     def test_plugin_has_get_display_modes(self, plugin_id):
         """Test that plugin can return display modes."""

test/plugins/test_harness.py

@@ -0,0 +1,255 @@
+"""
+Unit tests for the plugin safety harness primitives:
+bounds detection, image comparison, and mode enumeration.
+
+These don't load real plugins, so they run anywhere (including core CI where
+plugin-repos is empty).
+"""
+
+import importlib.util
+import json
+from pathlib import Path
+
+import pytest
+from PIL import Image
+
+from src.plugin_system.testing.bounds_display_manager import BoundsCheckingDisplayManager
+from src.plugin_system.testing.harness import (
+    _TOLERATED_UPDATE_ERRORS, compare_images, list_modes,
+)
+from src.plugin_system.testing.sizes import (
+    DEFAULT_TEST_SIZES, coerce_sizes, parse_size_token, resolve_test_sizes,
+)
+
+
+class TestBoundsDetection:
+    def test_reports_declared_size_not_canvas_size(self):
+        dm = BoundsCheckingDisplayManager(width=64, height=32)
+        assert dm.width == 64 and dm.height == 32
+        assert dm.matrix.width == 64 and dm.matrix.height == 32
+        # Backing canvas is padded out past the declared panel so far-overshoot
+        # coordinates land on-canvas and get flagged instead of clipped.
+        canvas_w, canvas_h = dm.image.size
+        assert canvas_w > 64 and canvas_h > 32
+
+    def test_far_overshoot_on_small_panel_is_detected(self):
+        # A coordinate meant for a wide build (x past 64) must still be caught
+        # when the declared panel is only 64 wide.
+        dm = BoundsCheckingDisplayManager(width=64, height=32)
+        dm.draw.rectangle([200, 5, 210, 10], fill=(255, 0, 0))
+        bbox = dm.check_overflow()
+        assert bbox is not None
+        assert bbox[0] >= 64
+
+    def test_in_bounds_drawing_has_no_overflow(self):
+        dm = BoundsCheckingDisplayManager(width=64, height=32)
+        dm.draw.rectangle([0, 0, 63, 31], fill=(255, 255, 255))
+        assert dm.check_overflow() is None
+
+    def test_right_overflow_is_detected(self):
+        dm = BoundsCheckingDisplayManager(width=64, height=32)
+        # Draw a few pixels past the right edge.
+        dm.draw.rectangle([60, 5, 70, 10], fill=(255, 0, 0))
+        bbox = dm.check_overflow()
+        assert bbox is not None
+        assert bbox[0] >= 64  # overflow starts at or past the declared width
+
+    def test_bottom_overflow_is_detected(self):
+        dm = BoundsCheckingDisplayManager(width=64, height=32)
+        dm.draw.rectangle([5, 30, 10, 40], fill=(0, 255, 0))
+        bbox = dm.check_overflow()
+        assert bbox is not None
+        assert bbox[3] > 32  # overflow extends past the declared height
+
+    def test_declared_image_is_cropped_to_panel(self):
+        dm = BoundsCheckingDisplayManager(width=64, height=32)
+        assert dm.get_image().size == (64, 32)
+
+    def test_snapshot_saves_cropped_panel(self, tmp_path):
+        dm = BoundsCheckingDisplayManager(width=128, height=32)
+        out = tmp_path / "snap.png"
+        dm.save_snapshot(str(out))
+        with Image.open(out) as img:
+            assert img.size == (128, 32)
+
+
+class TestArbitraryPanelSizes:
+    """The harness must handle any panel shape, not a fixed supported list."""
+
+    def test_overflow_extent_pads_to_largest_in_run(self):
+        # A wide run (extent 256) means content at x=200 on a 64-wide panel is
+        # caught; the same draw with a small extent would be clipped (false pass).
+        wide = BoundsCheckingDisplayManager(width=64, height=32, overflow_extent=(256, 32))
+        wide.draw.rectangle([200, 5, 210, 10], fill=(255, 0, 0))
+        assert wide.check_overflow() is not None
+
+        tight = BoundsCheckingDisplayManager(width=64, height=32, overflow_extent=(64, 32))
+        tight.draw.rectangle([200, 5, 210, 10], fill=(255, 0, 0))
+        assert tight.check_overflow() is None  # clipped beyond the small canvas
+
+    def test_unusual_shapes_report_their_declared_size(self):
+        for w, h in [(8, 2), (6, 6), (200, 8), (64, 96)]:
+            dm = BoundsCheckingDisplayManager(width=w, height=h)
+            assert dm.width == w and dm.height == h
+            assert dm.matrix.width == w and dm.matrix.height == h
+
+
+class TestUpdateErrorClassification:
+    """update() may fail for lack of network (tolerated) but a logic bug must
+    not pass green just because display() survives."""
+
+    def test_connectivity_errors_are_tolerated(self):
+        import socket
+        import urllib.error
+        for exc in (ConnectionError("x"), TimeoutError("x"), socket.gaierror("x"),
+                    urllib.error.URLError("x")):
+            assert isinstance(exc, _TOLERATED_UPDATE_ERRORS)
+
+    def test_logic_errors_are_not_tolerated(self):
+        for exc in (ValueError("x"), KeyError("x"), AttributeError("x"), TypeError("x")):
+            assert not isinstance(exc, _TOLERATED_UPDATE_ERRORS)
+
+
+class TestSizeParsing:
+    def test_parse_size_token_ok(self):
+        assert parse_size_token(" 128X32 ") == (128, 32)
+
+    def test_parse_size_token_rejects_garbage(self):
+        with pytest.raises(ValueError):
+            parse_size_token("128xabc")
+        with pytest.raises(ValueError):
+            parse_size_token("128-32")
+
+    def test_rejects_non_positive_dimensions(self):
+        for bad in ("0x32", "-64x32", "64x0", "64x-1"):
+            with pytest.raises(ValueError):
+                parse_size_token(bad)
+        with pytest.raises(ValueError):
+            coerce_sizes([[0, 32]])
+        with pytest.raises(ValueError):
+            coerce_sizes("64x-1")
+
+    def test_coerce_sizes_from_string_and_pairs(self):
+        assert coerce_sizes("8x16,64x64") == [(8, 16), (64, 64)]
+        assert coerce_sizes([[8, 16], (64, 64)]) == [(8, 16), (64, 64)]
+        assert coerce_sizes(None) is None
+        assert coerce_sizes("") is None
+
+    def test_resolve_precedence_env_then_spec_then_default(self, monkeypatch):
+        monkeypatch.delenv("LEDMATRIX_TEST_SIZES", raising=False)
+        assert resolve_test_sizes(None) == list(DEFAULT_TEST_SIZES)
+        assert resolve_test_sizes([[8, 16]]) == [(8, 16)]
+        monkeypatch.setenv("LEDMATRIX_TEST_SIZES", "5x5")
+        # env wins over a per-plugin spec
+        assert resolve_test_sizes([[8, 16]]) == [(5, 5)]
+
+
+class TestCompareImages:
+    def test_identical_images_match(self):
+        a = Image.new("RGB", (16, 16), (10, 20, 30))
+        b = a.copy()
+        ok, diff_pixels, max_delta = compare_images(a, b)
+        assert ok and diff_pixels == 0 and max_delta == 0
+
+    def test_different_images_fail_at_zero_tolerance(self):
+        a = Image.new("RGB", (16, 16), (0, 0, 0))
+        b = a.copy()
+        b.putpixel((1, 1), (255, 255, 255))
+        ok, diff_pixels, max_delta = compare_images(a, b)
+        assert not ok and diff_pixels == 1 and max_delta == 255
+
+    def test_tolerance_absorbs_small_noise(self):
+        a = Image.new("RGB", (16, 16), (100, 100, 100))
+        b = a.copy()
+        b.putpixel((2, 2), (103, 100, 100))  # delta 3
+        ok, _, max_delta = compare_images(a, b, max_delta=5, max_diff_pixels=0)
+        assert ok and max_delta == 3
+
+    def test_size_mismatch_fails(self):
+        a = Image.new("RGB", (16, 16))
+        b = Image.new("RGB", (32, 16))
+        ok, _, _ = compare_images(a, b)
+        assert not ok
+
+
+class TestListModes:
+    def test_instance_modes_take_precedence(self):
+        inst = type("P", (), {"modes": ["a", "b"]})()
+        assert list_modes(inst, {"display_modes": ["x"]}, "pid") == ["a", "b"]
+
+    def test_falls_back_to_manifest_display_modes(self):
+        inst = type("P", (), {})()
+        assert list_modes(inst, {"display_modes": ["x", "y"]}, "pid") == ["x", "y"]
+
+    def test_falls_back_to_plugin_id(self):
+        inst = type("P", (), {})()
+        assert list_modes(inst, {}, "pid") == ["pid"]
+
+
+def _load_check_plugin_cli():
+    """Load scripts/check_plugin.py by path (it isn't an importable package)."""
+    root = Path(__file__).resolve().parents[2]
+    path = root / "scripts" / "check_plugin.py"
+    spec = importlib.util.spec_from_file_location("check_plugin_cli", path)
+    mod = importlib.util.module_from_spec(spec)
+    spec.loader.exec_module(mod)
+    return mod
+
+
+def _make_fixture_plugin(tmp_path, harness):
+    """Create a minimal plugin dir with a test/harness.json; return its parent
+    (the search dir)."""
+    pdir = tmp_path / "plugins" / "demo-clock"
+    (pdir / "test").mkdir(parents=True)
+    (pdir / "manifest.json").write_text(json.dumps({
+        "id": "demo-clock", "name": "Demo Clock", "version": "1.0.0",
+        "author": "test", "entry_point": "manager.py", "class_name": "DemoClock",
+        "display_modes": ["demo-clock"], "compatible_versions": ["*"],
+    }))
+    (pdir / "test" / "harness.json").write_text(json.dumps(harness))
+    return pdir.parent
+
+
+class TestCheckPluginHonorsHarnessJson:
+    """Regression: check_plugin.py (the CI tool) must apply test/harness.json so
+    its render reproduces the committed goldens — otherwise time/data-dependent
+    plugins drift on every CI run."""
+
+    def test_harness_json_supplies_render_settings(self, tmp_path, monkeypatch):
+        mod = _load_check_plugin_cli()
+        search = _make_fixture_plugin(tmp_path, {
+            "config": {"timezone": "UTC"},
+            "freeze_time": "2025-08-01 15:25:00",
+            "sizes": [[128, 32]],
+        })
+        captured = {}
+        monkeypatch.setattr(mod, "render_plugin_matrix",
+                            lambda **kw: captured.update(kw) or [])
+        monkeypatch.setattr(mod, "compare_to_goldens", lambda *a, **k: [])
+        mod.check_one(
+            plugin_id="demo-clock", search_dirs=[str(search)], sizes=None,
+            mock_data={}, config={}, run_update=True, out_dir=None,
+            update_golden=False, golden_dir_override=None, freeze_time=None,
+        )
+        assert captured["freeze_time"] == "2025-08-01 15:25:00"
+        assert captured["config"]["timezone"] == "UTC"
+        assert captured["sizes"] == [(128, 32)]
+
+    def test_cli_flags_override_harness_json(self, tmp_path, monkeypatch):
+        mod = _load_check_plugin_cli()
+        search = _make_fixture_plugin(tmp_path, {
+            "config": {"timezone": "UTC"},
+            "freeze_time": "2025-08-01 15:25:00",
+        })
+        captured = {}
+        monkeypatch.setattr(mod, "render_plugin_matrix",
+                            lambda **kw: captured.update(kw) or [])
+        monkeypatch.setattr(mod, "compare_to_goldens", lambda *a, **k: [])
+        mod.check_one(
+            plugin_id="demo-clock", search_dirs=[str(search)], sizes=None,
+            mock_data={}, config={"timezone": "America/New_York"},
+            run_update=True, out_dir=None, update_golden=False,
+            golden_dir_override=None, freeze_time="2030-01-01 00:00:00",
+        )
+        assert captured["freeze_time"] == "2030-01-01 00:00:00"
+        assert captured["config"]["timezone"] == "America/New_York"

test/plugins/test_plugin_matrix.py

@@ -0,0 +1,115 @@
+"""
+Cross-size / cross-screen plugin safety test.
+
+For every discovered plugin, render every declared screen at every supported
+matrix size and assert it: loads, renders without crashing, stays within the
+panel bounds, and — for plugins that ship golden images — matches them.
+
+Plugin discovery (first match wins):
+  - $LEDMATRIX_PLUGINS_DIR  (os.pathsep-separated list of dirs), else
+  - <project_root>/plugin-repos and <project_root>/plugins
+
+A plugin opts into golden-image checks by adding test/golden/<WxH>/<mode>.png
+(and usually test/harness.json for deterministic config / mock data / time).
+"""
+
+import os
+from pathlib import Path
+from typing import Dict, List
+
+import pytest
+
+from src.plugin_system.testing.harness import (
+    render_plugin_matrix, compare_to_goldens,
+)
+from src.plugin_system.testing.loading import load_config_defaults, load_harness_spec
+from src.plugin_system.testing.sizes import resolve_test_sizes
+
+PROJECT_ROOT = Path(__file__).resolve().parents[2]
+
+# Set LEDMATRIX_REQUIRE_PLUGINS=1 in any CI/hardware pipeline where plugins are
+# expected to be present, so a discovery drift (empty search path) fails loudly
+# instead of silently skipping and losing this safety signal.
+_REQUIRE_PLUGINS = os.environ.get("LEDMATRIX_REQUIRE_PLUGINS") == "1"
+
+
+def _plugin_search_dirs() -> List[Path]:
+    env = os.environ.get("LEDMATRIX_PLUGINS_DIR")
+    if env:
+        return [Path(p) for p in env.split(os.pathsep) if p]
+    return [PROJECT_ROOT / "plugin-repos", PROJECT_ROOT / "plugins"]
+
+
+def _discover() -> Dict[str, Path]:
+    """Map plugin_id -> plugin_dir for all plugins on the search path."""
+    found: Dict[str, Path] = {}
+    for base in _plugin_search_dirs():
+        if not base.exists():
+            continue
+        for child in sorted(base.iterdir()):
+            if (child / "manifest.json").exists() and child.name not in found:
+                found[child.name] = child
+    return found
+
+
+_PLUGINS = _discover()
+
+
+@pytest.mark.plugin
+def test_plugins_were_discovered() -> None:
+    """Guard against silently skipping the whole matrix when discovery drifts.
+
+    Local dev and the plugin-less core CI legitimately have no plugins, so we
+    skip there; but when LEDMATRIX_REQUIRE_PLUGINS=1 an empty search path is a
+    hard failure rather than a green no-op.
+    """
+    if _PLUGINS:
+        return
+    search = [str(p) for p in _plugin_search_dirs()]
+    if _REQUIRE_PLUGINS:
+        pytest.fail(
+            "LEDMATRIX_REQUIRE_PLUGINS=1 but no plugins were discovered on the "
+            f"search path: {search}"
+        )
+    pytest.skip(f"no plugins found on the search path: {search}")
+
+
+@pytest.mark.plugin
+@pytest.mark.skipif(not _PLUGINS, reason="no plugins found on the search path")
+@pytest.mark.parametrize("plugin_id", sorted(_PLUGINS))
+def test_plugin_renders_across_sizes_and_screens(plugin_id: str) -> None:
+    plugin_dir = _PLUGINS[plugin_id]
+    spec = load_harness_spec(plugin_dir)
+
+    config = {"enabled": True}
+    config.update(load_config_defaults(plugin_dir))
+    config.update(spec.get("config", {}))
+
+    # Sizes: LEDMATRIX_TEST_SIZES env (test on real hardware) wins, then the
+    # plugin's own harness.json "sizes", else the default representative sample.
+    sizes = resolve_test_sizes(spec.get("sizes"))
+
+    results = render_plugin_matrix(
+        plugin_id=plugin_id,
+        plugin_dir=plugin_dir,
+        config=config,
+        mock_data=spec.get("mock_data_contents", {}),
+        sizes=sizes,
+        run_update=not spec.get("skip_update", False),
+        freeze_time=spec.get("freeze_time"),
+    )
+    compare_to_goldens(results, plugin_dir / "test" / "golden")
+
+    failures = []
+    for r in results:
+        if r.error is not None:
+            failures.append(f"{r.size_label} {r.mode}: crashed: {r.error}")
+        elif r.overflow is not None:
+            failures.append(f"{r.size_label} {r.mode}: overflow past panel bbox={r.overflow}")
+        elif r.golden_checked and r.golden_ok is False:
+            failures.append(
+                f"{r.size_label} {r.mode}: golden drift {r.golden_diff_pixels}px "
+                f"(max Δ={r.golden_max_delta})"
+            )
+
+    assert not failures, f"{plugin_id} failed:\n  " + "\n  ".join(failures)

test/test_display_controller.py

@@ -167,6 +167,53 @@ class TestDisplayControllerLivePriority:
         assert controller.current_display_mode == "test_plugin_live"
         assert controller.force_change is True
 
+    def test_live_priority_resume_continues_rotation(self, test_display_controller):
+        """Regression: when live priority ends, rotation resumes where it was
+        interrupted, not after the live plugin's mode.
+
+        Without the fix, _apply_live_priority left current_mode_index pointing at
+        the live plugin's slot, so the next rotation step skipped every mode
+        between the interrupted position and the live plugin (e.g. elections,
+        which sits just before a flights plugin in the order)."""
+        controller = test_display_controller
+        controller.available_modes = [
+            "weather", "forecast", "almanac", "election_ticker", "flight_live"
+        ]
+        # Rotation is about to show the 3rd mode (index 2).
+        controller.current_mode_index = 2
+        controller.current_display_mode = "almanac"
+        controller._live_resume_index = None
+
+        # Live priority (e.g. planes overhead) preempts -> flight_live (index 4).
+        controller._apply_live_priority("flight_live")
+        assert controller.current_display_mode == "flight_live"
+        assert controller.current_mode_index == 4
+        assert controller._live_resume_index == 2  # saved rotation position
+
+        # Re-checks while the hold continues must not move the saved position.
+        controller._apply_live_priority("flight_live")
+        assert controller._live_resume_index == 2
+
+        # Live priority ends -> resume at the saved index (almanac), so the next
+        # rotation step lands on election_ticker (index 3) rather than skipping it.
+        controller._apply_live_priority(None)
+        assert controller.current_mode_index == 2
+        assert controller.current_display_mode == "almanac"
+        assert controller._live_resume_index is None
+
+    def test_live_priority_no_resume_when_idle(self, test_display_controller):
+        """No saved position + no live content is a no-op (normal rotation)."""
+        controller = test_display_controller
+        controller.available_modes = ["a", "b", "c"]
+        controller.current_mode_index = 1
+        controller.current_display_mode = "b"
+        controller._live_resume_index = None
+
+        controller._apply_live_priority(None)
+
+        assert controller.current_mode_index == 1
+        assert controller.current_display_mode == "b"
+
 
 class TestDisplayControllerDynamicDuration:
     """Test dynamic duration handling."""
@@ -229,18 +276,20 @@ class TestDisplayControllerSchedule:
     def test_inactive_hours(self, test_display_controller):
         """Test inactive hours check."""
         controller = test_display_controller
+        # Inject schedule directly into self.config (what _check_schedule actually reads)
+        # and reset the minute gate so the cached result from any prior call is cleared.
+        controller.config['schedule'] = {
+            "enabled": True,
+            "start_time": "09:00",
+            "end_time": "17:00",
+        }
+        controller._schedule_checked_minute = None
+        controller._tz = None
+
         with patch('src.display_controller.datetime') as mock_datetime:
             mock_datetime.now.return_value.strftime.return_value.lower.return_value = "monday"
             mock_datetime.now.return_value.time.return_value = datetime.strptime("20:00", "%H:%M").time()
             mock_datetime.strptime = datetime.strptime
 
-            schedule_config = {
+            controller._check_schedule()
-                "schedule": {
+            assert controller.is_display_active is False
-                    "enabled": True,
-                    "start_time": "09:00",
-                    "end_time": "17:00"
-                }
-            }
-            with patch.object(controller.config_service, 'get_config', return_value=schedule_config):
-                controller._check_schedule()
-                assert controller.is_display_active is False

test/test_display_controller_optimizations.py

@@ -0,0 +1,322 @@
+"""
+Tests for the three display_controller.py optimizations:
+
+  Opt #1 — inspect.signature() caching per plugin_id
+  Opt #2 — pre-cached config values (_normal_brightness, _scroll_speed)
+  Opt #3 — schedule minute-gate (_check_schedule, _check_dim_schedule)
+"""
+
+import pytest
+import time
+from datetime import datetime
+from unittest.mock import MagicMock, patch, call
+
+
+# ---------------------------------------------------------------------------
+# Shared fixture
+# ---------------------------------------------------------------------------
+
+@pytest.fixture
+def controller(test_display_controller):
+    """Return a ready DisplayController from the existing suite fixture."""
+    return test_display_controller
+
+
+# ---------------------------------------------------------------------------
+# Opt #1 — signature cache
+# ---------------------------------------------------------------------------
+
+class TestSignatureCache:
+    """inspect.signature() should be called at most once per plugin_id."""
+
+    class _PluginWithMode:
+        """Real class whose display() accepts display_mode — inspectable by signature."""
+        plugin_id = "mode_plugin"
+        def display(self, display_mode=None, force_clear=False):
+            return True
+
+    class _PluginNoMode:
+        """Real class whose display() does NOT accept display_mode."""
+        plugin_id = "no_mode_plugin"
+        def display(self, force_clear=False):
+            return True
+
+    def test_cache_starts_empty(self, controller):
+        assert controller._plugin_accepts_display_mode == {}
+
+    def test_signature_computed_and_cached(self, controller):
+        """After the first cache population, the dict holds a bool and stays unchanged
+        if queried again without explicitly deleting the key."""
+        import inspect as _inspect
+        plugin = self._PluginNoMode()
+        key = "sig_test"
+        if key not in controller._plugin_accepts_display_mode:
+            controller._plugin_accepts_display_mode[key] = (
+                "display_mode" in _inspect.signature(plugin.display).parameters
+            )
+        original = controller._plugin_accepts_display_mode[key]
+
+        # Accessing cache again should not change the value
+        second = controller._plugin_accepts_display_mode[key]
+        assert second == original
+
+    def test_cache_stores_false_for_no_display_mode(self, controller):
+        """Plugin whose display() doesn't accept display_mode → cached False."""
+        import inspect as _inspect
+        plugin = self._PluginNoMode()
+        controller._plugin_accepts_display_mode["no_mode_plugin"] = (
+            "display_mode" in _inspect.signature(plugin.display).parameters
+        )
+        assert controller._plugin_accepts_display_mode["no_mode_plugin"] is False
+
+    def test_cache_stores_true_for_display_mode(self, controller):
+        """Plugin whose display() accepts display_mode → cached True."""
+        import inspect as _inspect
+        plugin = self._PluginWithMode()
+        controller._plugin_accepts_display_mode["mode_plugin"] = (
+            "display_mode" in _inspect.signature(plugin.display).parameters
+        )
+        assert controller._plugin_accepts_display_mode["mode_plugin"] is True
+
+    def test_cache_cleared_on_plugin_reload(self, controller):
+        """Populating plugin_modes for an id that's already cached must clear the entry."""
+        plugin = MagicMock()
+        controller._plugin_accepts_display_mode["reload_plugin"] = False
+
+        # Simulate the plugin_modes population code path (as in __init__)
+        plugin_id = "reload_plugin"
+        controller.plugin_modes["reload_plugin"] = plugin
+        if hasattr(controller, "_plugin_accepts_display_mode"):
+            controller._plugin_accepts_display_mode.pop(plugin_id, None)
+
+        assert "reload_plugin" not in controller._plugin_accepts_display_mode
+
+
+# ---------------------------------------------------------------------------
+# Opt #2 — cached config values
+# ---------------------------------------------------------------------------
+
+class TestCachedConfigValues:
+    """_normal_brightness and _scroll_speed are populated from config at init."""
+
+    def test_normal_brightness_cached(self, controller):
+        """_normal_brightness must equal what the config says."""
+        expected = (
+            controller.config
+            .get("display", {})
+            .get("hardware", {})
+            .get("brightness", 90)
+        )
+        assert controller._normal_brightness == expected
+
+    def test_scroll_speed_cached(self, controller):
+        """_scroll_speed must equal what the config says."""
+        expected = (
+            controller.config
+            .get("display", {})
+            .get("vegas_scroll", {})
+            .get("scroll_speed", 75)
+        )
+        assert controller._scroll_speed == expected
+
+    def test_current_brightness_uses_cached_value(self, controller):
+        """current_brightness is initialised from _normal_brightness."""
+        assert controller.current_brightness == controller._normal_brightness
+
+    def test_cached_target_brightness_init(self, controller):
+        """_cached_target_brightness starts equal to _normal_brightness."""
+        assert controller._cached_target_brightness == controller._normal_brightness
+
+    def test_normal_brightness_default_is_90(self, controller):
+        """If config has no brightness key the default is 90."""
+        controller.config = {}
+        controller._normal_brightness = (
+            controller.config.get("display", {})
+            .get("hardware", {})
+            .get("brightness", 90)
+        )
+        assert controller._normal_brightness == 90
+
+
+# ---------------------------------------------------------------------------
+# Opt #3 — schedule minute-gate
+# ---------------------------------------------------------------------------
+
+class TestScheduleMinuteGate:
+    """_check_schedule and _check_dim_schedule skip re-evaluation within the same minute."""
+
+    # ── _check_schedule ──────────────────────────────────────────────────────
+
+    def test_schedule_checked_minute_starts_none(self, controller):
+        assert controller._schedule_checked_minute is None
+
+    def test_first_call_sets_checked_minute(self, controller):
+        """After the first real evaluation the minute key is stored."""
+        controller.config["schedule"] = {
+            "enabled": True,
+            "start_time": "00:00",
+            "end_time": "23:59",
+        }
+        controller._schedule_checked_minute = None
+        controller._tz = None
+
+        controller._check_schedule()
+        assert controller._schedule_checked_minute is not None
+
+    def test_second_call_same_minute_does_not_re_evaluate(self, controller):
+        """A second call with the same (hour, minute) returns without changing state."""
+        controller.config["schedule"] = {
+            "enabled": True,
+            "start_time": "00:00",
+            "end_time": "23:59",
+        }
+        controller._tz = None
+        controller._schedule_checked_minute = None
+
+        # First call — evaluates and marks as active (whole-day window)
+        controller._check_schedule()
+        assert controller.is_display_active is True
+        first_minute_key = controller._schedule_checked_minute
+
+        # Force is_display_active to False so we can tell if it gets re-evaluated
+        controller.is_display_active = False
+
+        # Second call within the same minute — gate fires, is_display_active unchanged
+        controller._schedule_checked_minute = first_minute_key  # same minute
+        controller._check_schedule()
+        assert controller.is_display_active is False, (
+            "Second call in same minute should return immediately without re-evaluation"
+        )
+
+    def test_new_minute_forces_re_evaluation(self, controller):
+        """A different (hour, minute) key causes a full re-evaluation."""
+        controller.config["schedule"] = {
+            "enabled": True,
+            "start_time": "00:00",
+            "end_time": "23:59",
+        }
+        controller._tz = None
+
+        # Plant a stale minute key from yesterday
+        controller._schedule_checked_minute = (-1, -1)
+        controller.is_display_active = False  # wrong value to be corrected
+
+        controller._check_schedule()
+        assert controller.is_display_active is True, (
+            "A new minute key should trigger re-evaluation and correct is_display_active"
+        )
+
+    def test_gate_skipped_when_schedule_disabled(self, controller):
+        """When schedule.enabled=False the method returns before reaching the gate."""
+        controller.config["schedule"] = {"enabled": False}
+        controller._schedule_checked_minute = None
+        controller._tz = None
+
+        controller._check_schedule()
+        # The early-return path doesn't set the minute key
+        assert controller._schedule_checked_minute is None
+
+    # ── _check_dim_schedule ──────────────────────────────────────────────────
+
+    def test_dim_checked_minute_starts_none(self, controller):
+        assert controller._dim_checked_minute is None
+
+    def test_first_dim_call_sets_checked_minute(self, controller):
+        """First call with dim schedule enabled stores the minute key."""
+        controller.config["dim_schedule"] = {
+            "enabled": True,
+            "start_time": "22:00",
+            "end_time": "06:00",
+        }
+        controller.is_display_active = True
+        controller._dim_checked_minute = None
+        controller._tz = None
+
+        controller._check_dim_schedule()
+        assert controller._dim_checked_minute is not None
+
+    def test_dim_second_call_returns_cached_brightness(self, controller):
+        """Second call with same minute returns _cached_target_brightness immediately."""
+        controller.config["dim_schedule"] = {
+            "enabled": True,
+            "start_time": "22:00",
+            "end_time": "06:00",
+        }
+        controller.is_display_active = True
+        controller._dim_checked_minute = None
+        controller._tz = None
+
+        # First call stores the result
+        first_result = controller._check_dim_schedule()
+        assert controller._cached_target_brightness == first_result
+        minute_key = controller._dim_checked_minute
+
+        # Corrupt cached value to something recognisable
+        controller._cached_target_brightness = 42
+
+        # Second call in same minute — must return the cached 42
+        controller._dim_checked_minute = minute_key
+        second_result = controller._check_dim_schedule()
+        assert second_result == 42, (
+            "Same-minute call must return cached brightness, not re-compute"
+        )
+
+    def test_dim_gate_skipped_when_display_off(self, controller):
+        """When display is off the method exits before the minute gate."""
+        controller.config["dim_schedule"] = {"enabled": True, "start_time": "22:00", "end_time": "06:00"}
+        controller.is_display_active = False
+        controller._dim_checked_minute = None
+        controller._tz = None
+
+        controller._check_dim_schedule()
+        # Early-exit path does not set the minute key
+        assert controller._dim_checked_minute is None
+
+    def test_dim_cached_target_brightness_updated_after_full_evaluation(self, controller):
+        """After a full evaluation _cached_target_brightness reflects the result."""
+        controller.config["dim_schedule"] = {
+            "enabled": True,
+            "start_time": "22:00",
+            "end_time": "06:00",
+        }
+        controller.is_display_active = True
+        controller._dim_checked_minute = None   # force full re-evaluation
+        controller._tz = None
+
+        result = controller._check_dim_schedule()
+        assert controller._cached_target_brightness == result
+
+    # ── timezone lazy init ───────────────────────────────────────────────────
+
+    def test_tz_starts_none(self, controller):
+        assert controller._tz is None
+
+    def test_tz_lazily_initialised_on_first_schedule_check(self, controller):
+        """_tz is None until _check_schedule or _check_dim_schedule is called."""
+        controller.config["schedule"] = {
+            "enabled": True,
+            "start_time": "00:00",
+            "end_time": "23:59",
+        }
+        controller._tz = None
+        controller._schedule_checked_minute = None
+
+        controller._check_schedule()
+        assert controller._tz is not None
+
+    def test_tz_shared_between_schedule_and_dim(self, controller):
+        """Both methods use the same cached _tz instance."""
+        controller.config["schedule"] = {"enabled": True, "start_time": "00:00", "end_time": "23:59"}
+        controller.config["dim_schedule"] = {"enabled": True, "start_time": "22:00", "end_time": "06:00"}
+        controller.is_display_active = True
+        controller._tz = None
+        controller._schedule_checked_minute = None
+        controller._dim_checked_minute = None
+
+        controller._check_schedule()
+        tz_after_schedule = controller._tz
+
+        controller._check_dim_schedule()
+        assert controller._tz is tz_after_schedule, (
+            "_check_dim_schedule should reuse the _tz set by _check_schedule"
+        )

test/test_store_manager_caches.py

@@ -58,19 +58,15 @@ class TestGitInfoCache(unittest.TestCase):
         (self.plugin_path / ".git" / "HEAD").write_text("ref: refs/heads/main\n")
 
     def _fake_subprocess_run(self, *args, **kwargs):
-        # Return different dummy values depending on which git subcommand
+        # _get_local_git_info now reads branch and remote_url directly from
-        # was invoked so the code paths that parse output all succeed.
+        # .git/HEAD and .git/config (no subprocess) and uses a single
+        # ``git log --format=%H%n%cI`` call that returns SHA on line 1 and
+        # ISO date on line 2.  Adjust the fake accordingly.
         cmd = args[0]
         result = MagicMock()
         result.returncode = 0
-        if "rev-parse" in cmd and "HEAD" in cmd and "--abbrev-ref" not in cmd:
+        if "log" in cmd:
-            result.stdout = "abcdef1234567890\n"
+            result.stdout = "abcdef1234567890\n2026-04-08T12:00:00+00:00\n"
-        elif "--abbrev-ref" in cmd:
-            result.stdout = "main\n"
-        elif "config" in cmd:
-            result.stdout = "https://example.com/repo.git\n"
-        elif "log" in cmd:
-            result.stdout = "2026-04-08T12:00:00+00:00\n"
         else:
             result.stdout = ""
         return result
@@ -84,7 +80,8 @@ class TestGitInfoCache(unittest.TestCase):
             self.assertIsNotNone(first)
             self.assertEqual(first["short_sha"], "abcdef1")
             calls_after_first = mock_run.call_count
-            self.assertEqual(calls_after_first, 4)
+            # Production code now uses a single ``git log`` call.
+            self.assertEqual(calls_after_first, 1)
 
             # Second call with unchanged HEAD: zero new subprocess calls.
             second = self.sm._get_local_git_info(self.plugin_path)
@@ -105,7 +102,8 @@ class TestGitInfoCache(unittest.TestCase):
             os.utime(head, (new_time, new_time))
 
             self.sm._get_local_git_info(self.plugin_path)
-            self.assertEqual(mock_run.call_count, calls_after_first + 4)
+            # One new ``git log`` call after cache invalidation.
+            self.assertEqual(mock_run.call_count, calls_after_first + 1)
 
     def test_no_git_directory_returns_none(self):
         non_git = self.plugins_dir / "no_git"
@@ -192,14 +190,11 @@ class TestGitInfoCache(unittest.TestCase):
             result = MagicMock()
             result.returncode = 0
             cmd = args[0]
-            if "rev-parse" in cmd and "--abbrev-ref" not in cmd:
+            # Production code now uses a single ``git log --format=%H%n%cI``.
-                result.stdout = branch_file.read_text().strip() + "\n"
+            # Branch and remote_url are read directly from .git/HEAD/.git/config.
-            elif "--abbrev-ref" in cmd:
+            if "log" in cmd:
-                result.stdout = "main\n"
+                sha = branch_file.read_text().strip()
-            elif "config" in cmd:
+                result.stdout = f"{sha}\n2026-04-08T12:00:00+00:00\n"
-                result.stdout = "https://example.com/repo.git\n"
-            elif "log" in cmd:
-                result.stdout = "2026-04-08T12:00:00+00:00\n"
             else:
                 result.stdout = ""
             return result

test/test_web_api.py

@@ -617,7 +617,8 @@ class TestDottedKeyNormalization:
             'leagues': {'eng.1': {'enabled': True, 'favorite_teams': []}},
         }
         schema_mgr.merge_with_defaults.side_effect = lambda config, defaults: {**defaults, **config}
-        schema_mgr.validate_config_against_schema.return_value = []
+        # Must be a (bool, list) tuple: the endpoint does is_valid, errors = validate_config_against_schema(...)
+        schema_mgr.validate_config_against_schema.return_value = (True, [])
         api_v3.schema_manager = schema_mgr
 
         request_data = {
@@ -679,7 +680,7 @@ class TestDottedKeyNormalization:
             'leagues': {'eng.1': {'favorite_teams': []}},
         }
         schema_mgr.merge_with_defaults.side_effect = lambda config, defaults: {**defaults, **config}
-        schema_mgr.validate_config_against_schema.return_value = []
+        schema_mgr.validate_config_against_schema.return_value = (True, [])
         api_v3.schema_manager = schema_mgr
 
         request_data = {

test/web_interface/test_state_reconciliation.py

@@ -224,20 +224,14 @@ class TestStateReconciliation(unittest.TestCase):
         with open(manifest_path, 'w') as f:
             json.dump({"version": "1.0.0", "name": "Plugin 1"}, f)
         
-        # Mock save_config to track calls
-        saved_configs = []
-        def save_config(config):
-            saved_configs.append(config)
-        
-        self.config_manager.save_config = save_config
-        
         # Run reconciliation
         result = self.reconciler.reconcile_state()
-        
+
-        # Verify fix was attempted
+        # config.json is the source of truth for enabled state. The fix syncs
+        # the state manager to match config (config says True → state set True),
+        # rather than overwriting the config with the stale state value.
         self.assertEqual(len(result.inconsistencies_fixed), 1)
-        self.assertEqual(len(saved_configs), 1)
+        self.state_manager.set_plugin_enabled.assert_called_once_with("plugin1", True)
-        self.assertEqual(saved_configs[0]["plugin1"]["enabled"], False)
     
     def test_multiple_inconsistencies(self):
         """Test reconciliation with multiple inconsistencies."""

test/web_interface/test_systemctl_sudoers_alignment.py

@@ -0,0 +1,76 @@
+"""Guards that every privileged systemctl call the web interface makes is
+covered by a passwordless-sudo grant in configure_web_sudo.sh.
+
+The web interface runs headless (no TTY), so any `sudo` call that is not
+matched by a NOPASSWD rule in /etc/sudoers.d/ledmatrix_web falls back to a
+password prompt and fails with:
+
+    sudo: a terminal is required to read the password
+
+sudo matches the command line by exact string, so `systemctl start ledmatrix`
+and `systemctl start ledmatrix.service` are NOT interchangeable. This test
+parses both the production blueprint and the sudoers-generator script and
+asserts the (verb, unit) pairs line up, catching the suffix-mismatch class of
+bug before it ships.
+"""
+
+import ast
+import re
+from pathlib import Path
+
+PROJECT_ROOT = Path(__file__).resolve().parents[2]
+API_V3 = PROJECT_ROOT / "web_interface" / "blueprints" / "api_v3.py"
+SUDOERS_SCRIPT = PROJECT_ROOT / "scripts" / "install" / "configure_web_sudo.sh"
+
+
+def _sudo_systemctl_calls(source: str) -> set[tuple[str, str]]:
+    """Return (verb, unit) for every list literal beginning with
+    ['sudo', 'systemctl', ...] passed to a subprocess call in the source."""
+    calls: set[tuple[str, str]] = set()
+    for node in ast.walk(ast.parse(source)):
+        if not isinstance(node, ast.List):
+            continue
+        elts = node.elts
+        if len(elts) < 4:
+            continue
+        if not all(isinstance(e, ast.Constant) and isinstance(e.value, str) for e in elts[:4]):
+            continue
+        if elts[0].value == "sudo" and elts[1].value == "systemctl":
+            calls.add((elts[2].value, elts[3].value))
+    return calls
+
+
+def _granted_systemctl_rules(script: str) -> set[tuple[str, str]]:
+    """Return (verb, unit) for each `$SYSTEMCTL_PATH <verb> <unit>` NOPASSWD
+    grant emitted by the sudoers-generator script."""
+    rules: set[tuple[str, str]] = set()
+    for match in re.finditer(r"\$SYSTEMCTL_PATH\s+(\S+)\s+(\S+)", script):
+        verb, unit = match.group(1), match.group(2).rstrip('"')
+        rules.add((verb, unit))
+    return rules
+
+
+def test_every_sudo_systemctl_call_is_granted() -> None:
+    calls = _sudo_systemctl_calls(API_V3.read_text())
+    rules = _granted_systemctl_rules(SUDOERS_SCRIPT.read_text())
+
+    assert calls, "expected to find sudo systemctl calls in api_v3.py"
+
+    uncovered = {c for c in calls if c not in rules}
+    assert not uncovered, (
+        "These sudo systemctl calls have no matching NOPASSWD grant in "
+        "configure_web_sudo.sh; they will fail headless with "
+        "'sudo: a terminal is required to read the password': "
+        + ", ".join(f"systemctl {v} {u}" for v, u in sorted(uncovered))
+    )
+
+
+def test_units_are_fully_qualified() -> None:
+    """Privileged systemctl calls must name the unit as <name>.service so they
+    match the sudoers grants, which use the fully-qualified unit name."""
+    calls = _sudo_systemctl_calls(API_V3.read_text())
+    unqualified = {(v, u) for v, u in calls if not u.endswith(".service")}
+    assert not unqualified, (
+        "sudo systemctl calls must use fully-qualified .service unit names: "
+        + ", ".join(f"systemctl {v} {u}" for v, u in sorted(unqualified))
+    )

web_interface/app.py

@@ -391,6 +391,22 @@ def captive_portal_redirect():
     # Redirect to lightweight captive portal setup page (not the full UI)
     return redirect(url_for('pages_v3.captive_setup'), code=302)
 
+# Append a content-version query param (file mtime) to every static URL so the
+# long-lived `immutable` cache (see add_security_headers below) is actually safe:
+# when a static file changes its URL changes, so browsers refetch it. Without
+# this, edited JS/CSS were served immutable under an unchanging URL and never
+# reached clients until a manual cache clear.
+@app.url_defaults
+def add_static_version(endpoint, values):
+    if endpoint == 'static' and values.get('filename'):
+        try:
+            file_path = os.path.join(app.static_folder, values['filename'])
+            values['v'] = int(os.path.getmtime(file_path))
+        except OSError:
+            # File missing (e.g. plugin asset not yet installed) — skip versioning.
+            pass
+
+
 # Add security headers and caching to all responses
 @app.after_request
 def add_security_headers(response):

web_interface/blueprints/api_v3.py

@@ -190,7 +190,7 @@ def _ensure_display_service_running():
     if status.get('active'):
         status['started'] = False
         return status
-    result = _run_systemctl_command(['sudo', 'systemctl', 'start', 'ledmatrix'])
+    result = _run_systemctl_command(['sudo', 'systemctl', 'start', 'ledmatrix.service'])
     service_status = _get_display_service_status()
     result['started'] = result.get('returncode') == 0
     result['active'] = service_status.get('active')
@@ -199,7 +199,7 @@ def _ensure_display_service_running():
 
 def _stop_display_service():
     """Stop the ledmatrix display service."""
-    result = _run_systemctl_command(['sudo', 'systemctl', 'stop', 'ledmatrix'])
+    result = _run_systemctl_command(['sudo', 'systemctl', 'stop', 'ledmatrix.service'])
     status = _get_display_service_status()
     result['active'] = status.get('active')
     result['status'] = status
@@ -1461,7 +1461,7 @@ def execute_system_action():
                 # For on-demand modes, we would need to integrate with the display controller
                 # For now, just start the display service
                 try:
-                    result = subprocess.run(['sudo', 'systemctl', 'start', 'ledmatrix'],
+                    result = subprocess.run(['sudo', 'systemctl', 'start', 'ledmatrix.service'],
                                          capture_output=True, text=True, timeout=10)
                 except subprocess.TimeoutExpired as e:
                     logger.error("start_display (%s) timed out: %s", mode, e)
@@ -1478,16 +1478,16 @@ def execute_system_action():
                     resp['stderr'] = result.stderr.strip()
                 return jsonify(resp)
             else:
-                result = subprocess.run(['sudo', 'systemctl', 'start', 'ledmatrix'],
+                result = subprocess.run(['sudo', 'systemctl', 'start', 'ledmatrix.service'],
                                      capture_output=True, text=True, timeout=10)
         elif action == 'stop_display':
-            result = subprocess.run(['sudo', 'systemctl', 'stop', 'ledmatrix'],
+            result = subprocess.run(['sudo', 'systemctl', 'stop', 'ledmatrix.service'],
                                  capture_output=True, text=True, timeout=10)
         elif action == 'enable_autostart':
-            result = subprocess.run(['sudo', 'systemctl', 'enable', 'ledmatrix'],
+            result = subprocess.run(['sudo', 'systemctl', 'enable', 'ledmatrix.service'],
                                  capture_output=True, text=True, timeout=10)
         elif action == 'disable_autostart':
-            result = subprocess.run(['sudo', 'systemctl', 'disable', 'ledmatrix'],
+            result = subprocess.run(['sudo', 'systemctl', 'disable', 'ledmatrix.service'],
                                  capture_output=True, text=True, timeout=10)
         elif action == 'reboot_system':
             result = subprocess.run(['sudo', 'reboot'],
@@ -1568,11 +1568,11 @@ def execute_system_action():
                 'message': pull_message,
             })
         elif action == 'restart_display_service':
-            result = subprocess.run(['sudo', 'systemctl', 'restart', 'ledmatrix'],
+            result = subprocess.run(['sudo', 'systemctl', 'restart', 'ledmatrix.service'],
                                  capture_output=True, text=True, timeout=10)
         elif action == 'restart_web_service':
             # Try to restart the web service (assuming it's ledmatrix-web.service)
-            result = subprocess.run(['sudo', 'systemctl', 'restart', 'ledmatrix-web'],
+            result = subprocess.run(['sudo', 'systemctl', 'restart', 'ledmatrix-web.service'],
                                  capture_output=True, text=True, timeout=10)
         else:
             return jsonify({'status': 'error', 'message': 'Unknown action'}), 400
@@ -2412,6 +2412,13 @@ def reconcile_plugin_state():
 
         from src.plugin_system.state_reconciliation import StateReconciliation
 
+        # Parse optional `force` flag from request body, guarding against
+        # non-dict bodies (bare string, array, null) that would raise AttributeError.
+        payload = request.get_json(silent=True)
+        if not isinstance(payload, dict):
+            payload = {}
+        force = _coerce_to_bool(payload.get('force', False))
+
         reconciler = StateReconciliation(
             state_manager=api_v3.plugin_state_manager,
             config_manager=api_v3.config_manager,
@@ -2419,7 +2426,7 @@ def reconcile_plugin_state():
             plugins_dir=Path(api_v3.plugin_manager.plugins_dir)
         )
 
-        result = reconciler.reconcile_state()
+        result = reconciler.reconcile_state(force=force)
 
         return success_response(
             data={
@@ -2846,6 +2853,89 @@ def update_plugin():
             status_code=500
         )
 
+def _do_transactional_uninstall(plugin_id, preserve_config):
+    """Execute an uninstall with snapshot-based rollback.
+
+    Order of operations:
+      1. Snapshot main config + secrets (abort on unexpected errors, proceed on expected I/O errors).
+      2. Clean up plugin config (abort with 500 if this raises — avoids orphaned files).
+      3. Unload plugin from runtime if loaded (rollback + 500 if this raises).
+      4. Remove plugin files (rollback + 500 if this returns False or raises).
+      5. Finish (remove state, invalidate caches).
+
+    Rollback restores the config snapshot and, if the plugin had been
+    loaded before unload, calls load_plugin to restore runtime state.
+
+    Returns (True, None) on success or (False, error_message) on failure.
+    """
+    from src.exceptions import ConfigError
+
+    # --- Step 1: snapshot main + secrets ---
+    main_snapshot = None
+    secrets_snapshot = None
+    try:
+        main_snapshot = api_v3.config_manager.get_raw_file_content('main')
+    except (OSError, ConfigError):
+        pass  # Proceed without snapshot; narrow catch preserves TypeError/AttributeError
+    try:
+        secrets_snapshot = api_v3.config_manager.get_raw_file_content('secrets')
+    except (OSError, ConfigError):
+        pass
+
+    # --- Step 2: cleanup config first (abort before touching filesystem) ---
+    if not preserve_config:
+        api_v3.config_manager.cleanup_plugin_config(plugin_id, remove_secrets=True)
+
+    # Record whether the plugin was running before we touch anything.
+    was_loaded = (
+        api_v3.plugin_manager is not None
+        and plugin_id in api_v3.plugin_manager.plugins
+    )
+
+    def _rollback(reload_plugin):
+        if main_snapshot is not None:
+            try:
+                api_v3.config_manager.save_raw_file_content('main', main_snapshot)
+            except Exception as restore_err:
+                logger.error("Failed to restore main config snapshot for %s: %s", plugin_id, restore_err)
+        if secrets_snapshot is not None:
+            try:
+                api_v3.config_manager.save_raw_file_content('secrets', secrets_snapshot)
+            except Exception as restore_err:
+                logger.error("Failed to restore secrets snapshot for %s: %s", plugin_id, restore_err)
+        if reload_plugin and api_v3.plugin_manager is not None:
+            try:
+                api_v3.plugin_manager.load_plugin(plugin_id)
+            except Exception as reload_err:
+                logger.error("Failed to reload plugin %s during rollback: %s", plugin_id, reload_err)
+
+    # --- Step 3: unload ---
+    if was_loaded:
+        try:
+            api_v3.plugin_manager.unload_plugin(plugin_id)
+        except Exception as unload_err:
+            _rollback(reload_plugin=False)  # unload failed — runtime state unchanged
+            return False, f"Failed to unload plugin {plugin_id}: {unload_err}"
+
+    # --- Step 4: remove files ---
+    try:
+        success = api_v3.plugin_store_manager.uninstall_plugin(plugin_id)
+    except Exception as remove_err:
+        _rollback(reload_plugin=was_loaded)
+        return False, f"Failed to remove plugin {plugin_id}: {remove_err}"
+
+    if not success:
+        _rollback(reload_plugin=was_loaded)
+        return False, f"Failed to uninstall plugin {plugin_id}"
+
+    # --- Step 5: finish ---
+    if api_v3.schema_manager:
+        api_v3.schema_manager.invalidate_cache(plugin_id)
+    if api_v3.plugin_state_manager:
+        api_v3.plugin_state_manager.remove_plugin_state(plugin_id)
+    return True, None
+
+
 @api_v3.route('/plugins/uninstall', methods=['POST'])
 def uninstall_plugin():
     """Uninstall plugin"""
@@ -2865,19 +2955,13 @@ def uninstall_plugin():
         plugin_id = data['plugin_id']
         preserve_config = data.get('preserve_config', False)
 
-        # Use operation queue if available
+        # Both queued and direct paths use the same transactional helper so
+        # snapshot/rollback behaviour is consistent regardless of deployment.
         if api_v3.operation_queue:
             def uninstall_callback(operation):
-                """Callback to execute plugin uninstallation."""
+                """Callback to execute plugin uninstallation via transactional helper."""
-                # Unload the plugin first if it's loaded
+                success, error_msg = _do_transactional_uninstall(plugin_id, preserve_config)
-                if api_v3.plugin_manager and plugin_id in api_v3.plugin_manager.plugins:
-                    api_v3.plugin_manager.unload_plugin(plugin_id)
-
-                # Uninstall the plugin
-                success = api_v3.plugin_store_manager.uninstall_plugin(plugin_id)
-
                 if not success:
-                    error_msg = f'Failed to uninstall plugin {plugin_id}'
                     if api_v3.operation_history:
                         api_v3.operation_history.record_operation(
                             "uninstall",
@@ -2885,24 +2969,7 @@ def uninstall_plugin():
                             status="failed",
                             error=error_msg
                         )
-                    raise Exception(error_msg)
+                    raise Exception(error_msg or f'Failed to uninstall plugin {plugin_id}')
-
-                # Invalidate schema cache
-                if api_v3.schema_manager:
-                    api_v3.schema_manager.invalidate_cache(plugin_id)
-
-                # Clean up plugin configuration if not preserving
-                if not preserve_config:
-                    try:
-                        api_v3.config_manager.cleanup_plugin_config(plugin_id, remove_secrets=True)
-                    except Exception as cleanup_err:
-                        logger.warning("Failed to cleanup config after uninstall: %s", cleanup_err)
-
-                # Remove from state manager
-                if api_v3.plugin_state_manager:
-                    api_v3.plugin_state_manager.remove_plugin_state(plugin_id)
-
-                # Record in history
                 if api_v3.operation_history:
                     api_v3.operation_history.record_operation(
                         "uninstall",
@@ -2910,7 +2977,6 @@ def uninstall_plugin():
                         status="success",
                         details={"preserve_config": preserve_config}
                     )
-
                 return {'success': True, 'message': 'Plugin uninstalled successfully'}
 
             # Enqueue operation
@@ -2925,31 +2991,10 @@ def uninstall_plugin():
                 message='Plugin uninstallation queued'
             )
         else:
-            # Fallback to direct uninstall
+            # Direct (non-queued) transactional uninstall
-            # Unload the plugin first if it's loaded
+            success, error_msg = _do_transactional_uninstall(plugin_id, preserve_config)
-            if api_v3.plugin_manager and plugin_id in api_v3.plugin_manager.plugins:
-                api_v3.plugin_manager.unload_plugin(plugin_id)
-
-            # Uninstall the plugin
-            success = api_v3.plugin_store_manager.uninstall_plugin(plugin_id)
 
             if success:
-                # Invalidate schema cache
-                if api_v3.schema_manager:
-                    api_v3.schema_manager.invalidate_cache(plugin_id)
-
-                # Clean up plugin configuration if not preserving
-                if not preserve_config:
-                    try:
-                        api_v3.config_manager.cleanup_plugin_config(plugin_id, remove_secrets=True)
-                    except Exception as cleanup_err:
-                        logger.warning("Failed to cleanup config after uninstall: %s", cleanup_err)
-
-                # Remove from state manager
-                if api_v3.plugin_state_manager:
-                    api_v3.plugin_state_manager.remove_plugin_state(plugin_id)
-
-                # Record in history
                 if api_v3.operation_history:
                     api_v3.operation_history.record_operation(
                         "uninstall",
@@ -2957,7 +3002,6 @@ def uninstall_plugin():
                         status="success",
                         details={"preserve_config": preserve_config}
                     )
-
                 return success_response(message='Plugin uninstalled successfully')
             else:
                 if api_v3.operation_history:
@@ -2965,12 +3009,11 @@ def uninstall_plugin():
                         "uninstall",
                         plugin_id=plugin_id,
                         status="failed",
-                        error='Plugin uninstall failed'
+                        error=error_msg
                     )
-
                 return error_response(
                     ErrorCode.PLUGIN_UNINSTALL_FAILED,
-                    'Plugin uninstall failed',
+                    error_msg or 'Plugin uninstall failed',
                     status_code=500
                 )
 
@@ -3666,13 +3709,14 @@ def _parse_form_value_with_schema(value, key_path, schema):
                 except ValueError:
                     return prop.get('default', 0.0)
 
-        # Try parsing as number (fallback)
+        # Try parsing as number (fallback) — skip when schema explicitly says string
-        try:
+        if not (prop and prop.get('type') == 'string'):
-            if '.' in stripped:
+            try:
-                return float(stripped)
+                if '.' in stripped:
-            return int(stripped)
+                    return float(stripped)
-        except ValueError:
+                return int(stripped)
-            pass
+            except ValueError:
+                pass
 
         # Return as string
         return value
@@ -4217,7 +4261,9 @@ def save_plugin_config():
                             nested_dict = config_dict.get(prop_key)
 
                         if isinstance(nested_dict, dict):
-                            fix_array_structures(nested_dict, prop_schema['properties'], nested_prefix)
+                            # Pass no prefix: config_dict is already the navigated sub-dict,
+                            # so path segments from the parent would mis-navigate it.
+                            fix_array_structures(nested_dict, prop_schema['properties'])
 
             # Also ensure array fields that are None get converted to empty arrays
             def ensure_array_defaults(config_dict, schema_props, prefix=''):
@@ -4277,7 +4323,8 @@ def save_plugin_config():
                                 nested_dict = config_dict[prop_key]
 
                         if isinstance(nested_dict, dict):
-                            ensure_array_defaults(nested_dict, prop_schema['properties'], nested_prefix)
+                            # Pass no prefix: config_dict is already navigated.
+                            ensure_array_defaults(nested_dict, prop_schema['properties'])
 
             if schema and 'properties' in schema:
                 # First, fix any dict structures that should be arrays
@@ -4377,6 +4424,21 @@ def save_plugin_config():
             defaults = schema_mgr.generate_default_config(plugin_id, use_cache=True)
             plugin_config = schema_mgr.merge_with_defaults(plugin_config, defaults)
 
+        # After merging defaults, replace any None array values with their schema defaults.
+        # merge_with_defaults gives user config higher priority, so a None submitted by
+        # the client can survive the merge — this pass cleans those up.
+        def _fix_none_arrays(cfg, props):
+            for k, pschema in props.items():
+                if pschema.get('type') == 'array':
+                    if isinstance(cfg, dict) and (k not in cfg or cfg[k] is None):
+                        cfg[k] = pschema.get('default', [])
+                elif pschema.get('type') == 'object' and 'properties' in pschema:
+                    if isinstance(cfg, dict) and isinstance(cfg.get(k), dict):
+                        _fix_none_arrays(cfg[k], pschema['properties'])
+
+        if schema and 'properties' in schema and isinstance(plugin_config, dict):
+            _fix_none_arrays(plugin_config, schema['properties'])
+
         # Ensure enabled state is preserved after defaults merge
         # Defaults should not overwrite an explicitly preserved enabled value
         if preserved_enabled is not None:

web_interface/static/v3/js/widgets/array-table.js

@@ -440,7 +440,7 @@
                 <h3 class="text-base font-semibold text-gray-900">Advanced Properties</h3>
                 <button type="button" onclick="window.closeArrayTableRowEditor()"
                         class="text-gray-400 hover:text-gray-600"><i class="fas fa-times"></i></button>
-            </div>`;
+            </div>`);
 
         const body = document.createElement('div');
         body.className = 'px-5 py-4 space-y-4';
@@ -512,7 +512,7 @@
             <button type="button" onclick="window.closeArrayTableRowEditor()"
                     class="px-4 py-2 text-sm text-gray-700 border border-gray-300 rounded-md hover:bg-gray-100">Cancel</button>
             <button type="button" id="array-row-editor-save"
-                    class="px-4 py-2 text-sm bg-blue-600 hover:bg-blue-700 text-white rounded-md">Save</button>`;
+                    class="px-4 py-2 text-sm bg-blue-600 hover:bg-blue-700 text-white rounded-md">Save</button>`);
 
         // Save handler
         footer.querySelector('#array-row-editor-save').onclick = function() {

web_interface/templates/v3/base.html

@@ -352,15 +352,14 @@
                         }
                     });
 
-                    // Set data-loaded on tab containers after HTMX settles their content,
+                    // Mark tab containers as loaded once their content settles, so switching
-                    // preventing repeated re-fetches on every tab switch.
+                    // away and back doesn't re-fetch. Scoped to the "loadtab" trigger (tab
-                    // Scoped to elements with hx-trigger="revealed" (tab containers only) so
+                    // containers only) so modals and plugin config panels can still reload.
-                    // modals and plugin config panels that legitimately reload are unaffected.
                     document.body.addEventListener('htmx:afterSettle', function(event) {
                         if (event.detail && event.detail.target) {
                             var target = event.detail.target;
                             var trigger = target.getAttribute('hx-trigger') || '';
-                            if (trigger.includes('revealed')) {
+                            if (trigger.includes('loadtab')) {
                                 target.setAttribute('data-loaded', 'true');
                             }
                         }
@@ -867,7 +866,7 @@
     <link rel="stylesheet" href="https://cdnjs.cloudflare.com/ajax/libs/font-awesome/6.0.0/css/all.min.css">
 
     <!-- Custom v3 styles -->
-    <link rel="stylesheet" href="{{ url_for('static', filename='v3/app.css') }}?v=20260216b">
+    <link rel="stylesheet" href="{{ url_for('static', filename='v3/app.css') }}">
 </head>
 <body x-data="app()" class="bg-gray-50 min-h-screen">
     <!-- Header -->
@@ -1030,7 +1029,7 @@
         <div id="tab-content" class="space-y-6">
             <!-- Overview tab -->
             <div x-show="activeTab === 'overview'" x-transition>
-                <div id="overview-content" hx-get="/v3/partials/overview" hx-trigger="revealed" hx-swap="innerHTML" hx-on::htmx:response-error="loadOverviewDirect()">
+                <div id="overview-content" hx-get="/v3/partials/overview" hx-trigger="loadtab" hx-swap="innerHTML" hx-on::htmx:response-error="loadOverviewDirect()">
                     <div class="animate-pulse">
                         <div class="bg-white rounded-lg shadow p-6">
                             <div class="h-4 bg-gray-200 rounded w-1/4 mb-4"></div>
@@ -1098,7 +1097,7 @@
 
             <!-- General tab -->
             <div x-show="activeTab === 'general'" x-transition>
-                <div id="general-content" hx-get="/v3/partials/general" hx-trigger="revealed" hx-swap="innerHTML">
+                <div id="general-content" hx-get="/v3/partials/general" hx-trigger="loadtab" hx-swap="innerHTML">
                     <div class="animate-pulse">
                         <div class="bg-white rounded-lg shadow p-6">
                             <div class="h-4 bg-gray-200 rounded w-1/4 mb-4"></div>
@@ -1116,7 +1115,7 @@
             <div x-show="activeTab === 'wifi'" x-transition>
                 <div id="wifi-content" 
                      hx-get="/v3/partials/wifi" 
-                     hx-trigger="revealed" 
+                     hx-trigger="loadtab" 
                      hx-swap="innerHTML"
                      hx-on::htmx:response-error="loadWifiDirect()">
                     <div class="animate-pulse">
@@ -1167,7 +1166,7 @@
 
             <!-- Schedule tab -->
             <div x-show="activeTab === 'schedule'" x-transition>
-                <div id="schedule-content" hx-get="/v3/partials/schedule" hx-trigger="revealed" hx-swap="innerHTML">
+                <div id="schedule-content" hx-get="/v3/partials/schedule" hx-trigger="loadtab" hx-swap="innerHTML">
                     <div class="animate-pulse">
                         <div class="bg-white rounded-lg shadow p-6">
                             <div class="h-4 bg-gray-200 rounded w-1/4 mb-4"></div>
@@ -1183,7 +1182,7 @@
 
             <!-- Display tab -->
             <div x-show="activeTab === 'display'" x-transition>
-                <div id="display-content" hx-get="/v3/partials/display" hx-trigger="revealed" hx-swap="innerHTML">
+                <div id="display-content" hx-get="/v3/partials/display" hx-trigger="loadtab" hx-swap="innerHTML">
                     <div class="animate-pulse">
                         <div class="bg-white rounded-lg shadow p-6">
                             <div class="h-4 bg-gray-200 rounded w-1/4 mb-4"></div>
@@ -1198,7 +1197,7 @@
 
             <!-- Backup & Restore tab -->
             <div x-show="activeTab === 'backup-restore'" x-transition>
-                <div id="backup-restore-content" hx-get="/v3/partials/backup-restore" hx-trigger="revealed" hx-swap="innerHTML">
+                <div id="backup-restore-content" hx-get="/v3/partials/backup-restore" hx-trigger="loadtab" hx-swap="innerHTML">
                     <div class="animate-pulse">
                         <div class="bg-white rounded-lg shadow p-6">
                             <div class="h-4 bg-gray-200 rounded w-1/4 mb-4"></div>
@@ -1210,7 +1209,7 @@
 
             <!-- Config Editor tab -->
             <div x-show="activeTab === 'config-editor'" x-transition>
-                <div id="config-editor-content" hx-get="/v3/partials/raw-json" hx-trigger="revealed" hx-swap="innerHTML">
+                <div id="config-editor-content" hx-get="/v3/partials/raw-json" hx-trigger="loadtab" hx-swap="innerHTML">
                     <div class="animate-pulse">
                         <div class="bg-white rounded-lg shadow p-6">
                             <div class="h-4 bg-gray-200 rounded w-1/4 mb-4"></div>
@@ -1225,7 +1224,7 @@
 
             <!-- Plugins tab -->
             <div x-show="activeTab === 'plugins'" x-transition>
-                <div id="plugins-content" hx-get="/v3/partials/plugins" hx-trigger="revealed" hx-swap="innerHTML"
+                <div id="plugins-content" hx-get="/v3/partials/plugins" hx-trigger="loadtab" hx-swap="innerHTML"
                      hx-on::response-error="loadPluginsDirect()">
                     <div class="animate-pulse">
                         <div class="bg-white rounded-lg shadow p-6">
@@ -1242,7 +1241,7 @@
 
             <!-- Fonts tab -->
             <div x-show="activeTab === 'fonts'" x-transition>
-                <div id="fonts-content" hx-get="/v3/partials/fonts" hx-trigger="revealed" hx-swap="innerHTML">
+                <div id="fonts-content" hx-get="/v3/partials/fonts" hx-trigger="loadtab" hx-swap="innerHTML">
                     <div class="animate-pulse">
                         <div class="bg-white rounded-lg shadow p-6">
                             <div class="h-4 bg-gray-200 rounded w-1/4 mb-4"></div>
@@ -1257,7 +1256,7 @@
 
             <!-- Logs tab -->
             <div x-show="activeTab === 'logs'" x-transition>
-                <div id="logs-content" hx-get="/v3/partials/logs" hx-trigger="revealed" hx-swap="innerHTML">
+                <div id="logs-content" hx-get="/v3/partials/logs" hx-trigger="loadtab" hx-swap="innerHTML">
                     <div class="animate-pulse">
                         <div class="bg-white rounded-lg shadow p-6">
                             <div class="h-4 bg-gray-200 rounded w-1/4 mb-4"></div>
@@ -1269,7 +1268,7 @@
 
             <!-- Cache tab -->
             <div x-show="activeTab === 'cache'" x-transition>
-                <div id="cache-content" hx-get="/v3/partials/cache" hx-trigger="revealed" hx-swap="innerHTML">
+                <div id="cache-content" hx-get="/v3/partials/cache" hx-trigger="loadtab" hx-swap="innerHTML">
                     <div class="animate-pulse">
                         <div class="bg-white rounded-lg shadow p-6">
                             <div class="h-4 bg-gray-200 rounded w-1/4 mb-4"></div>
@@ -1281,7 +1280,7 @@
 
             <!-- Operation History tab -->
             <div x-show="activeTab === 'operation-history'" x-transition>
-                <div id="operation-history-content" hx-get="/v3/partials/operation-history" hx-trigger="revealed" hx-swap="innerHTML">
+                <div id="operation-history-content" hx-get="/v3/partials/operation-history" hx-trigger="loadtab" hx-swap="innerHTML">
                     <div class="animate-pulse">
                         <div class="bg-white rounded-lg shadow p-6">
                             <div class="h-4 bg-gray-200 rounded w-1/4 mb-4"></div>
@@ -1861,28 +1860,53 @@
                 },
 
                 loadTabContent(tab) {
-                    // Try to load content for the active tab
+                    const contentEl = document.getElementById(tab + '-content');
-                    if (typeof htmx !== 'undefined') {
+                    // data-loaded: already fetched. data-loading: a fetch is queued or in
-                        const contentId = tab + '-content';
+                    // flight. Both guard against re-entry so a panel loads exactly once, even
-                        const contentEl = document.getElementById(contentId);
+                    // if the tab is reopened before an in-progress (or polling) load settles.
-                        if (contentEl && !contentEl.hasAttribute('data-loaded')) {
+                    if (!contentEl || contentEl.hasAttribute('data-loaded') || contentEl.hasAttribute('data-loading')) return;
-                            // Trigger HTMX load
+                    const url = contentEl.getAttribute('hx-get');
-                            htmx.trigger(contentEl, 'revealed');
+                    if (!url) return;
+
+                    contentEl.setAttribute('data-loading', 'true');
+
+                    // htmx.ajax issues the request and swaps the response into the panel
+                    // directly, so it works even before htmx has wired up the element's
+                    // hx-trigger listeners. data-loaded is stamped on success so the panel
+                    // loads once; the activeTab check drops loads for a tab the user navigated
+                    // away from while htmx was still loading (avoids fetching hidden panels).
+                    const swap = contentEl.getAttribute('hx-swap') || 'innerHTML';
+                    const load = () => {
+                        if (this.activeTab !== tab || contentEl.hasAttribute('data-loaded')) {
+                            contentEl.removeAttribute('data-loading');
+                            return;
                         }
-                    } else {
+                        return htmx.ajax('GET', url, { target: contentEl, swap: swap })
-                        // HTMX is still loading asynchronously — retry when it signals ready,
+                            .then(() => contentEl.setAttribute('data-loaded', 'true'))
-                        // or fall back to direct fetch if it fails to load entirely.
+                            .catch(() => {}) // leave unstamped on failure so it can retry
-                        const self = this;
+                            .finally(() => contentEl.removeAttribute('data-loading'));
-                        function onReady() { window.removeEventListener('htmx-load-failed', onFailed); self.loadTabContent(tab); }
+                    };
-                        function onFailed() {
+
-                            window.removeEventListener('htmx:ready', onReady);
+                    if (typeof htmx !== 'undefined') {
+                        load();
+                        return;
+                    }
+
+                    // htmx is loaded from a CDN and may not be ready yet. Poll until it is,
+                    // then load; if it never arrives, fall back to a direct fetch.
+                    let tries = 0;
+                    const timer = setInterval(() => {
+                        if (typeof htmx !== 'undefined') {
+                            clearInterval(timer);
+                            load();
+                        } else if (++tries > 100) { // ~10s
+                            clearInterval(timer);
+                            contentEl.removeAttribute('data-loading');
                             if (tab === 'overview' && typeof loadOverviewDirect === 'function') loadOverviewDirect();
                             else if (tab === 'wifi' && typeof loadWifiDirect === 'function') loadWifiDirect();
                             else if (tab === 'plugins' && typeof loadPluginsDirect === 'function') loadPluginsDirect();
                         }
-                        window.addEventListener('htmx:ready', onReady, { once: true });
+                    }, 100);
-                        window.addEventListener('htmx-load-failed', onFailed, { once: true });
-                    }
                 },
 
                 async loadInstalledPlugins() {