CVE-2026-41863: LLM-influenced filename used unsanitized in Path.resolve before file write in Spring AI support for Anthropic Skills API

A CVE disclosure (CVE-2026-41863) has been identified in Spring AI's support for the Anthropic Skills API, where an LLM-influenced filename is used unsanitized in Path.resolve before a file write operation, creating a path traversal vulnerability. The actual advisory content was not fully accessible, but the CVE title indicates a security flaw where AI-generated input is not properly sanitized before being used in file system operations.