OpenClaw < 2026.2.2 - Exec Allowlist Umgehung über Befehlssubstitution in doppelten Anführungszeichen

The impact of this vulnerability is severe. Successful exploitation allows an attacker to gain complete control over the affected system. By injecting command substitution syntax (using $() or backticks) within double-quoted strings, an attacker can bypass the intended allowlist and execute arbitrary commands with the privileges of the OpenClaw process. This could lead to data breaches, system compromise, and potential lateral movement within the network. The ability to execute arbitrary commands effectively grants the attacker root access, enabling them to install malware, modify system configurations, or steal sensitive data.

Systems running OpenClaw, particularly those exposed to untrusted networks or user input, are at significant risk. Environments utilizing OpenClaw for critical infrastructure or sensitive data processing should prioritize patching. Shared hosting environments where OpenClaw is installed could also be affected if the host's security practices are inadequate.

• linux / server:

journalctl -u openclaw | grep -i "command substitution"
ps aux | grep openclaw | grep -i "$()"

1. 2026-03-05Disclosure

   disclosure

1. Reserviert2026-02-27
2. Veröffentlicht2026-03-05
3. Geändert2026-04-22
4. EPSS aktualisiert2026-03-23

The primary mitigation for CVE-2026-28470 is to immediately upgrade OpenClaw to version 2026.2.2 or later. If upgrading is not immediately feasible, consider implementing strict input validation and sanitization on any user-supplied data that is processed by OpenClaw. While a direct WAF rule is unlikely to be effective due to the nature of the command injection, carefully reviewing and restricting the allowed characters within the exec approvals allowlist can provide a limited layer of defense. After upgrading, verify the fix by attempting to inject command substitution syntax into an OpenClaw process and confirming that the commands are not executed.