OpenClaw before 2026.3.22 has a flaw that lets an attacker rebind replies sent by a webhook to an unintended user. The bug arises because the system matches recipients by mutable usernames instead of stable numeric identifiers. An attacker can point replies to a different private or public account, potentially delivering sensitive content to the wrong user or hijacking communication flows. The weakness aligns with a common authorization flaw.

The affected product is OpenClaw; all installations using OpenClaw prior to version 2026.3.22 are vulnerable. The issue surfaces within Synology Chat deployments that rely on OpenClaw for webhook functionality.

The CVSS score of 6 indicates moderate severity. The EPSS score is not available, and the vulnerability is not listed in CISA's KEV catalog. The likely attack vector is through manipulation of webhook configuration or by changing a user’s username, enabling the attacker to redirect reply traffic. An attacker would need access to the webhook or the ability to alter usernames, but no remote code execution or privilege escalation is required.