Skip to content

AI turns documents or topics into real, native PowerPoint decks—with native shapes, transitions and animations, data-backed charts and tables on demand, audio narration from speaker notes, and support for your own .pptx templates. · by Hugo He

fix(tooling): prevent silent state loss and false-success results

hugohe3627a41460d8d
diff --git a/skills/ppt-master/scripts/image_search.py b/skills/ppt-master/scripts/image_search.py
index 0dac009f..adea9ae0 100644
--- a/skills/ppt-master/scripts/image_search.py
+++ b/skills/ppt-master/scripts/image_search.py
@@ -280,13 +280,22 @@ def _try_provider(
name: str,
request: ImageSearchRequest,
license_tier_filter: str,
+ *,
+ provider_is_explicit: bool = False,
) -> tuple[Optional[list[AssetCandidate]], Optional[str]]:
- """Run one provider while preserving whether it errored or returned no rows."""
+ """Run one provider while preserving whether it errored or returned no rows.
+
+ An explicitly selected provider is required; a missing key is retryable
+ instead of an optional-provider skip.
+ """
try:
module = _load_provider(name)
return module.search(request, license_tier_filter=license_tier_filter), None
except RuntimeError as exc:
- if _is_keyed_provider_unconfigured(name, exc):
+ if (
+ not provider_is_explicit
+ and _is_keyed_provider_unconfigured(name, exc)
+ ):
print(
f" [{name}] skipped: {exc}",
file=sys.stderr,
@@ -611,6 +620,7 @@ def search_and_download(
strict_no_attribution: bool,
save_candidates: bool = False,
max_candidates: int = 4,
+ provider_is_explicit: bool = False,
) -> SearchDownloadResult:
"""Find a candidate AND successfully download it.
@@ -621,6 +631,8 @@ def search_and_download(
Returns a structured result so batch mode can keep transient/provider
failures retryable while treating a complete no-match as terminal.
+ ``provider_is_explicit`` distinguishes a required provider from an optional
+ member of the default fallback chain.
"""
license_filters: list[str] = (
["no-attribution-only"] if strict_no_attribution else ["all"]
@@ -634,7 +646,12 @@ def search_and_download(
ranked: list[tuple[float, str, AssetCandidate]] = []
for provider_name in providers:
print(f" -> trying {provider_name} ({stage}) ...", file=sys.stderr)
- candidates, provider_error = _try_provider(provider_name, request, stage)
+ candidates, provider_error = _try_provider(
+ provider_name,
+ request,
+ stage,
+ provider_is_explicit=provider_is_explicit,
+ )
if provider_error:
provider_errors.append(provider_error)
if not candidates:
@@ -1469,6 +1486,7 @@ def _search_one_item(
strict_no_attribution=strict,
save_candidates=save_candidates,
max_candidates=max_candidates,
+ provider_is_explicit=bool(pinned),
)
if result.candidate is None:
return (
@@ -2044,6 +2062,7 @@ def main(argv: Optional[list[str]] = None) -> int:
strict_no_attribution=args.strict_no_attribution,
save_candidates=args.save_candidates,
max_candidates=args.max_candidates,
+ provider_is_explicit=bool(args.provider),
)
if result.candidate is None:
diff --git a/skills/ppt-master/scripts/pptx_intake.py b/skills/ppt-master/scripts/pptx_intake.py
index f9366be9..97abd1a5 100644
--- a/skills/ppt-master/scripts/pptx_intake.py
+++ b/skills/ppt-master/scripts/pptx_intake.py
@@ -200,6 +200,36 @@ def build_source_profile(
SOURCE_INDEX_NAME = "source_profile.json"
+def _load_source_index(index_path: Path) -> dict[str, Any]:
+ """Load and validate an existing multi-deck source index."""
+ if not index_path.exists():
+ return {}
+ if not index_path.is_file():
+ raise RuntimeError(f"Source index is not a file: {index_path}")
+
+ try:
+ loaded = json.loads(index_path.read_text(encoding="utf-8"))
+ except (json.JSONDecodeError, UnicodeError) as exc:
+ raise RuntimeError(
+ f"Source index contains invalid JSON and was left unchanged: {index_path}"
+ ) from exc
+ except OSError as exc:
+ raise RuntimeError(f"Cannot read source index: {index_path}: {exc}") from exc
+
+ if not isinstance(loaded, dict) or not isinstance(loaded.get("decks"), list):
+ raise RuntimeError(
+ f"Source index must be a JSON object with a decks array and was left unchanged: "
+ f"{index_path}"
+ )
+ for index, deck in enumerate(loaded["decks"]):
+ if not isinstance(deck, dict):
+ raise RuntimeError(
+ f"Source index decks[{index}] must be an object and was left unchanged: "
+ f"{index_path}"
+ )
+ return loaded
+
+
def upsert_source_index(output_dir: Path, digest: dict[str, Any]) -> Path:
"""Merge one deck's digest into the single multi-deck index `source_profile.json`.
@@ -209,14 +239,7 @@ def upsert_source_index(output_dir: Path, digest: dict[str, Any]) -> Path:
with the same stem replaces its entry in place.
"""
index_path = output_dir / SOURCE_INDEX_NAME
- index: dict[str, Any] = {}
- if index_path.is_file():
- try:
- loaded = json.loads(index_path.read_text(encoding="utf-8"))
- if isinstance(loaded, dict) and isinstance(loaded.get("decks"), list):
- index = loaded
- except (json.JSONDecodeError, OSError):
- index = {}
+ index = _load_source_index(index_path)
stem = digest.get("stem")
decks = [d for d in index.get("decks", []) if d.get("stem") != stem]
decks.append(digest)
diff --git a/skills/ppt-master/scripts/project_manager.py b/skills/ppt-master/scripts/project_manager.py
index 57638a65..bd042df8 100644
--- a/skills/ppt-master/scripts/project_manager.py
+++ b/skills/ppt-master/scripts/project_manager.py
@@ -218,6 +218,17 @@ class ProjectManager:
) -> str:
base_path = Path(base_dir) if base_dir else self.base_dir
+ if (
+ not project_name
+ or project_name in {".", ".."}
+ or Path(project_name).is_absolute()
+ or "/" in project_name
+ or "\\" in project_name
+ ):
+ raise ValueError(
+ "Project name must be a single, non-absolute path component"
+ )
+
normalized_format = normalize_canvas_format(canvas_format)
if normalized_format not in self.CANVAS_FORMATS:
available = ", ".join(sorted(self.CANVAS_FORMATS.keys()))
@@ -236,6 +247,10 @@ class ProjectManager:
project_dir_name = f"{project_name}_{normalized_format}_{date_str}"
project_path = base_path / project_dir_name
+ if not is_within_path(project_path, base_path):
+ raise ValueError(
+ f"Project directory must stay within the base directory: {base_path}"
+ )
if project_path.exists():
raise FileExistsError(f"Project directory already exists: {project_path}")
diff --git a/skills/ppt-master/scripts/source_to_md/web_to_md.py b/skills/ppt-master/scripts/source_to_md/web_to_md.py
index 4dfbe9ef..b000933c 100644
--- a/skills/ppt-master/scripts/source_to_md/web_to_md.py
+++ b/skills/ppt-master/scripts/source_to_md/web_to_md.py
@@ -912,7 +912,7 @@ def _write_emit_result(result_file: str, url: str, markdown_path: str) -> None:
print(f" [WARN] Could not write --emit-result: {exc}")
-def main() -> None:
+def main(argv: list[str] | None = None) -> int:
"""Run the CLI entry point."""
parser = argparse.ArgumentParser(
description="Web to Markdown Converter (Python)")
@@ -926,7 +926,7 @@ def main() -> None:
help="On success, write the saved output path as JSON to this file "
"(single-URL dispatcher use, so a title-named file can be located)")
- args = parser.parse_args()
+ args = parser.parse_args(argv)
if args.dir:
CONFIG["output_dir"] = args.dir
@@ -942,11 +942,16 @@ def main() -> None:
and not l.strip().startswith("#")]
targets.extend(lines)
else:
- print(f"Error: File {args.file} not found")
+ print(f"Error: File {args.file} not found", file=sys.stderr)
+ return 1
if not targets:
- parser.print_help()
- sys.exit(0)
+ parser.print_usage(sys.stderr)
+ print(
+ "web_to_md.py: error: at least one URL or --file is required",
+ file=sys.stderr,
+ )
+ return 2
results = []
for i, url in enumerate(targets):
@@ -970,10 +975,12 @@ def main() -> None:
for r in results:
if not r[0]:
print(f" - {r[1]}: {r[2]}")
+ return 1
+ return 0
if __name__ == "__main__":
# Disable warnings for verify=False if needed, though often useful to see
import urllib3
urllib3.disable_warnings(urllib3.exceptions.InsecureRequestWarning)
- main()
+ raise SystemExit(main())
diff --git a/skills/ppt-master/scripts/svg_editor/server.py b/skills/ppt-master/scripts/svg_editor/server.py
index eeadc3ce..71306e71 100644
--- a/skills/ppt-master/scripts/svg_editor/server.py
+++ b/skills/ppt-master/scripts/svg_editor/server.py
@@ -597,8 +597,10 @@ def create_app(
logger.warning('slide parse failed: %s: %s', svg_file.name, exc)
_cache_put(_LIST_CACHE, _LIST_CACHE_LOCK, path_str, mtime, disk_count)
- mem_count = len(annotations.get(svg_file.name, {}))
- annotation_count = max(disk_count, mem_count)
+ if svg_file.name in annotations:
+ annotation_count = len(annotations[svg_file.name])
+ else:
+ annotation_count = disk_count
slides.append({
'name': svg_file.name,
@@ -614,16 +616,43 @@ def create_app(
def _safe_svg_path(name: str):
"""Validate slide name and return safe path. Returns None if invalid.
- The early string checks reject obvious bad inputs; the resolve()+startswith()
+ The early string checks reject obvious bad inputs; the resolve()+relative_to()
check is the authoritative path traversal guard.
"""
if '/' in name or '\\' in name or '..' in name:
return None
svg_file = (svg_dir / name).resolve()
- if not str(svg_file).startswith(str(svg_dir.resolve())):
+ try:
+ svg_file.relative_to(svg_dir.resolve())
+ except ValueError:
return None
return svg_file
+ def _get_annotation_snapshot(name: str):
+ """Return the page's complete staged annotation state, loading it once."""
+ annotations = app.config['ANNOTATIONS']
+ if name in annotations:
+ return annotations[name], None
+
+ svg_file = _safe_svg_path(name)
+ if svg_file is None:
+ return None, (jsonify({'error': 'Invalid slide name'}), 400)
+ if not svg_file.exists():
+ return None, (jsonify({'error': 'Slide not found'}), 404)
+
+ try:
+ root = ET.parse(str(svg_file)).getroot()
+ except ET.ParseError as exc:
+ logger.warning('slide parse failed: %s: %s', name, exc)
+ return None, (jsonify({'error': f'Failed to parse SVG: {exc}'}), 500)
+
+ assign_temp_ids(root)
+ annotations[name] = {
+ item['element_id']: item['annotation']
+ for item in parse_annotations(root)
+ }
+ return annotations[name], None
+
@app.route('/api/slide/<name>')
def get_slide(name: str):
svg_file = _safe_svg_path(name)
@@ -686,11 +715,13 @@ def create_app(
(content, warnings, disk_annotations, id_to_tag),
)
- mem_annotations = app.config['ANNOTATIONS'].get(name, {})
- merged: dict[str, str] = {}
- for ann in disk_annotations:
- merged[ann['element_id']] = ann['annotation']
- merged.update(mem_annotations)
+ if name in app.config['ANNOTATIONS']:
+ merged = dict(app.config['ANNOTATIONS'][name])
+ else:
+ merged = {
+ ann['element_id']: ann['annotation']
+ for ann in disk_annotations
+ }
annotations_list = [
{
@@ -728,29 +759,28 @@ def create_app(
if len(annotation) > 10000:
return jsonify({'error': 'Annotation too long (max 10000 chars)'}), 400
- if name not in app.config['ANNOTATIONS']:
- app.config['ANNOTATIONS'][name] = {}
+ annotations, error = _get_annotation_snapshot(name)
+ if error is not None:
+ return error
- app.config['ANNOTATIONS'][name][element_id] = annotation
+ annotations[element_id] = annotation
return jsonify({
'status': 'ok',
- 'annotations_count': len(app.config['ANNOTATIONS'][name]),
+ 'annotations_count': len(annotations),
})
@app.route('/api/slide/<name>/annotate/<element_id>', methods=['DELETE'])
def delete_annotate(name: str, element_id: str):
- annotations = app.config['ANNOTATIONS']
- # Ensure the file key exists so save-all knows to rewrite this file
- # even if no new annotations were added (pure delete path).
- if name not in annotations:
- annotations[name] = {}
- if element_id in annotations[name]:
- del annotations[name][element_id]
+ annotations, error = _get_annotation_snapshot(name)
+ if error is not None:
+ return error
+ if element_id in annotations:
+ del annotations[element_id]
return jsonify({
'status': 'ok',
- 'annotations_count': len(annotations.get(name, {})),
+ 'annotations_count': len(annotations),
})
@app.route('/api/slide/<name>/edit', methods=['POST'])
@@ -896,6 +926,7 @@ def create_app(
filenames = sorted(set(annotations.keys()) | set(pending_edits.keys()))
for filename in filenames:
+ has_staged_annotations = filename in annotations
anns = annotations.get(filename, {})
edits = pending_edits.get(filename, [])
# anns may be empty when the user deleted all annotations — still
@@ -921,6 +952,8 @@ def create_app(
item['element_id']: item['annotation']
for item in parse_annotations(root)
}
+ if not has_staged_annotations:
+ anns = old_annotations
# Clear all existing annotations from the file before writing current state
for elem in root.iter():
diff --git a/skills/ppt-master/scripts/update_spec.py b/skills/ppt-master/scripts/update_spec.py
index ed3ee434..ce4c1813 100755
--- a/skills/ppt-master/scripts/update_spec.py
+++ b/skills/ppt-master/scripts/update_spec.py
@@ -97,7 +97,10 @@ def replace_color_in_svgs(
"""
if not HEX_RE.match(old_hex) or not HEX_RE.match(new_hex):
raise ValueError(f"not a HEX color: old={old_hex!r} new={new_hex!r}")
- pattern = re.compile(re.escape(old_hex), re.IGNORECASE)
+ pattern = re.compile(
+ rf"{re.escape(old_hex)}(?![0-9A-Fa-f])",
+ re.IGNORECASE,
+ )
planned: list[tuple[Path, str, int]] = []
for svg in sorted(svg_dir.glob("*.svg")):
text = svg.read_text(encoding="utf-8")