🛡️ Sentinel: [HIGH] Fix XXE vulnerabilities in XML parsing

.jules/sentinel.md

@@ -7,3 +7,8 @@
 **Vulnerability:** SQL Injection via string-interpolated subqueries with LIMIT. The `query` function in `optional-skills/mcp/fastmcp/templates/database_server.py` wrapped user SQL in a subquery: `SELECT * FROM ({sql}) LIMIT N`. This allowed malicious users to bypass simple checks (e.g. ensuring it starts with SELECT) and inject additional clauses or statements by manipulating the closing parenthesis.
 **Learning:** SQLite does not natively support parameterization for the FROM clause (e.g., subqueries or table names). Attempting to string-interpolate user input into a subquery creates an injection vector, especially when trying to enforce a LIMIT clause on user-provided queries.
 **Prevention:** To prevent SQL injection when applying limits to user-provided SQL queries, execute the raw user query directly and restrict the output rows in Python using `cursor.fetchmany(limit)` instead of trying to wrap the query in another SELECT with a LIMIT clause.
+
+## 2024-05-26 - Security Enhancement: XXE Prevention
+**Vulnerability:** XML External Entity (XXE) vulnerability via `xml.etree.ElementTree`.
+**Learning:** `xml.etree.ElementTree` is vulnerable to XML External Entity (XXE) attacks when parsing untrusted or external XML data. This vulnerability was found in scripts making external requests (e.g. `skills/research/arxiv/scripts/search_arxiv.py`, `optional-skills/devops/watchers/scripts/watch_rss.py`) and receiving callback requests (`gateway/platforms/wecom_callback.py`).
+**Prevention:** Always use `defusedxml.ElementTree` (or `defusedxml.minidom`) when parsing XML data from untrusted sources to mitigate XXE and billion laughs attacks.

gateway/platforms/wecom_callback.py

@@ -17,7 +17,7 @@
 import socket as _socket
 import time
 from typing import Any, Dict, List, Optional
-from xml.etree import ElementTree as ET
+import defusedxml.ElementTree as ET
 
 try:
     from aiohttp import web

optional-skills/devops/watchers/scripts/watch_rss.py

@@ -19,7 +19,7 @@
 import urllib.error
 import urllib.request
 from pathlib import Path
-from xml.etree import ElementTree as ET
+import defusedxml.ElementTree as ET
 
 sys.path.insert(0, str(Path(__file__).parent))
 from _watermark import Watermark, format_items_as_markdown  # type: ignore

skills/research/arxiv/scripts/search_arxiv.py

@@ -13,7 +13,7 @@
 import sys
 import urllib.request
 import urllib.parse
-import xml.etree.ElementTree as ET
+import defusedxml.ElementTree as ET
 
 NS = {'a': 'http://www.w3.org/2005/Atom'}