XML external entity (XXE) — how to fix unsafe XML parsing

Many XML parsers resolve external entities by default, which lets a crafted XML document read files off your server or make it send requests. If you parse XML that a user can supply, disable external entities or use a parser that does so by default, such as Python's defusedxml.

Why it's a problem

A malicious XML document can define an entity that points at a local file or an internal URL, and a default parser will happily fetch and include it, leaking file contents or enabling server-side request forgery. It only matters when the XML is attacker-controlled, but upload endpoints and API integrations often accept exactly that.

# Python default parser resolves external entities
import xml.etree.ElementTree as ET
tree = ET.parse(user_uploaded_file)

# defusedxml disables external entities and XML bombs
import defusedxml.ElementTree as ET
tree = ET.parse(user_uploaded_file)

Why AI tools write this

The standard library XML parser is the obvious completion for 'parse this XML,' and it works for well-formed documents. The external-entity risk is a default the assistant does not flag, because the code runs fine on the trusted files used while building.

The quick fix

• Use a hardened parser (defusedxml in Python) or explicitly disable external entities and DTDs.
• Never parse untrusted XML with a parser's default settings.
• If you only need data, consider accepting JSON instead of XML.