Zhihong Tian and Hui Lu found that XStream was vulnerable to remote code execution. A remote attacker could run arbitrary shell commands by manipulating the processed input stream. (CVE-2020-26217)
It was discovered that XStream was vulnerable to server-side forgery attacks. A remote attacker could request data from internal resources that are not publicly available only by manipulating the processed input stream. (CVE-2020-26258)
It was discovered that XStream was vulnerable to arbitrary file deletion on the local host. A remote attacker could use this to delete arbitrary known files on the host as long as the executing process had sufficient rights only by manipulating the processed input stream. (CVE-2020-26259)