This flaw is a CWE‑918 Server‑Side Request Forgery vulnerability. The vulnerable SitemapLoader applies the restrict_to_same_domain control only to direct URL entries and omits validation for nested sitemap references. As a result, an attacker who can influence the sitemap being processed can point a nested sitemap entry at an internal address. The loader fetches the specified internal resource without domain comparison, parses its contents, and returns the data as part of its output Documents, thereby disclosing internal responses to an external caller.

The vulnerability impacts the Langchain Community library, specifically the SitemapLoader component within the langchain_community package. All releases of this library that expose the restrict_to_same_domain option are potentially affected; no specific version range is listed, so any deployment using the current library is at risk.

The flaw carries a CVSS score of 7.7, classifying it as a high severity Server‑Side Request Forgery. No EPSS score is available, and the vulnerability is not included in CISA's KEV catalog. The likely attack vector is inferred to be exploitation of a maliciously crafted sitemap; an attacker only needs control over the sitemap data that the application processes. If successful, the server will fetch the victim’s internal resources, parsing and delivering their content to the attacker in the form of Documents.