From d8fa70e17b3ddf129e4c05bafe29555f7b36fded Mon Sep 17 00:00:00 2001
From: mistic96 <112645335+mistic96@users.noreply.github.com>
Date: Wed, 17 Jun 2026 09:46:19 +0000
Subject: [PATCH] Harden HTML output against U+2028 XSS + two
 crash-on-adversarial-input fixes (#1357)

Harden HTML export against U+2028/U+2029 script-breakout XSS + two crash-on-adversarial-input fixes (non-dict LLM JSON, _extract_parallel IndexError). Validated: full suite 2107 passed, HTML export smoke clean. Thanks @mistic96.
---
 graphify/extract.py   |  7 ++++---
 graphify/llm.py       | 12 ++++++++++--
 graphify/tree_html.py |  2 +-
 3 files changed, 15 insertions(+), 6 deletions(-)

diff --git a/graphify/extract.py b/graphify/extract.py
index fb7741d6..f40fa928 100644
--- a/graphify/extract.py
+++ b/graphify/extract.py
@@ -12164,16 +12164,17 @@ def _extract_parallel(
     try:
         with concurrent.futures.ProcessPoolExecutor(max_workers=max_workers) as pool:
             futures = {
-                pool.submit(_extract_single_file, item): item[0] for item in work_items
+                pool.submit(_extract_single_file, item): pos
+                for pos, item in enumerate(work_items)
             }
             for future in concurrent.futures.as_completed(futures):
                 try:
                     idx, result = future.result()
                     per_file[idx] = result
                 except Exception as exc:
-                    idx = futures[future]
+                    pos = futures[future]
                     print(
-                        f"  warning: worker failed for {work_items[idx][1]}: {exc}",
+                        f"  warning: worker failed for {work_items[pos][1]}: {exc}",
                         file=sys.stderr, flush=True,
                     )
                 done_count += 1
diff --git a/graphify/llm.py b/graphify/llm.py
index 335e1b3f..c0d294a5 100644
--- a/graphify/llm.py
+++ b/graphify/llm.py
@@ -721,7 +721,12 @@ def _parse_llm_json(raw: str) -> dict:
         else:
             stripped = after_fence.strip()
     try:
-        return json.loads(stripped)
+        parsed = json.loads(stripped)
+        if isinstance(parsed, dict):
+            return parsed
+        # Top-level array/scalar (common LLM output) is not a usable graph
+        # fragment; fall through to the next strategy rather than returning a
+        # non-dict that callers will try to subscript (e.g. result["input_tokens"]).
     except json.JSONDecodeError:
         pass
     # Strategy 2: extract the first balanced JSON object found anywhere in
@@ -751,7 +756,10 @@ def _parse_llm_json(raw: str) -> dict:
                 depth -= 1
                 if depth == 0:
                     try:
-                        return json.loads(stripped[start : i + 1])
+                        parsed = json.loads(stripped[start : i + 1])
+                        if isinstance(parsed, dict):
+                            return parsed
+                        break
                     except json.JSONDecodeError:
                         break
     print(
diff --git a/graphify/tree_html.py b/graphify/tree_html.py
index 3d825add..94b10df4 100644
--- a/graphify/tree_html.py
+++ b/graphify/tree_html.py
@@ -549,7 +549,7 @@ def emit_html(
 ) -> str:
     # Escape </script> sequences so embedded JSON cannot break out of the
     # <script> tag, and HTML-escape values that land in <title>/<h1>.
-    data_json = json.dumps(tree, ensure_ascii=False, separators=(",", ":")).replace("</", "<\\/")
+    data_json = json.dumps(tree, ensure_ascii=True, separators=(",", ":")).replace("</", "<\\/")
     return _HTML_TEMPLATE.format(
         title=_html.escape(title),
         header=_html.escape(header),