OpenClaw versions prior to 2026.5.7 are vulnerable because the Matrix allowFrom feature lets authenticated users modify display name metadata. An attacker who can change a display name can cause the system to match policy entries intended for another Matrix identity, effectively gaining agent access that was not originally authorized. This flaw can allow an attacker to obtain higher privileges than permitted, potentially exposing sensitive data or control over the system. The weakness is identified as CWE‑290, representing an authentication or authorization issue that leads to improper validation of user credentials.

The affected product is OpenClaw from the OpenClaw vendor. Versions earlier than 2026.5.7 are impacted; no further version specifics are provided.

The vulnerability carries a CVSS score of 7.7, indicating high severity. EPSS data is not available and the issue is not listed in the CISA KEV catalog, suggesting that exploitation is not currently observed at scale. The probable attack vector requires an authenticated account with write access to display names within the Matrix allowFrom configuration. Once such an account is compromised or misused, the attacker can manipulate the display name to match an existing policy entry, thereby obtaining unintended agent-level permissions. Given the high CVSS score and the administrative‑level impact, the risk of severe privilege escalation is significant.