OpenClaw Authentication Token Reuse Vulnerability in Owner ID Prompt Hashing

A vulnerability exists in OpenClaw versions prior to 2026.2.22, where the gateway authentication token is reused as a fallback hash secret for obfuscating owner IDs. This occurs when the owner display is set to hash and no specific owner display secret is provided. The dual use of authentication tokens creates a security risk, as attackers with access to system prompts sent to third-party model providers can extract the gateway authentication token from the hash outputs, undermining gateway authentication security.

Impact

The vulnerability leads to a dual-use of authentication secrets, mixing gateway authentication with prompt metadata hashing. While it does not directly disclose the plaintext token, it allows for the extraction of the gateway authentication token from hashed outputs, potentially compromising authentication security, especially if weak tokens are used.

Reproduction

To reproduce this vulnerability, set the 'commands.ownerDisplay' configuration to 'hash' without specifying an 'ownerDisplaySecret'. When prompts are sent to third-party model providers, the gateway authentication token will be derived from the hash output, exposing the token and compromising authentication security.

Remediation

The vulnerability has been patched in OpenClaw version 2026.2.22. Users should update to this version.