OpenClaw présente une vulnérabilité SSRF (Server-Side Request Forgery) via redirection de citation de recherche web avec une politique private-network-allowing

The impact of this SSRF vulnerability lies in the ability of an attacker to initiate requests from the OpenClaw host to internal network resources. By manipulating citation redirect targets, an attacker could potentially access sensitive data residing on internal servers, interact with internal APIs, or even attempt to pivot to other systems within the network. While the description doesn't detail specific data at risk, the potential for internal network scanning and data exfiltration is significant. This vulnerability could be exploited to gain unauthorized access to internal services and data, potentially leading to a broader compromise of the affected environment.

Organizations utilizing OpenClaw for web search and citation processing are at risk, particularly those with internal services accessible via the network. Environments with less stringent network segmentation or those relying on untrusted citation sources are especially vulnerable. Any deployment of OpenClaw versions prior to 2026.3.1 should be considered at risk.

• nodejs / server:

  journalctl -u openclaw | grep -i "citation redirect"

• generic web:

  curl -I <openclaw_endpoint> | grep -i "X-Forwarded-For"

• generic web:

  grep -r "web_search" /path/to/openclaw/source/code

1. 2026-03-02Disclosure

   disclosure

1. Réservé2026-03-10
2. Publiée2026-03-02
3. Modifiée2026-03-20
4. EPSS mis à jour2026-03-26

The primary mitigation for CVE-2026-31989 is to upgrade OpenClaw to version 2026.3.1 or later. This version implements a stricter SSRF policy that blocks redirects to localhost and private/internal network destinations. If upgrading immediately is not feasible, consider implementing temporary workarounds such as restricting outbound network access from the OpenClaw host to only necessary destinations. Review and validate all citation redirect sources to ensure they are trustworthy. There are no specific WAF rules or detection signatures readily available for this vulnerability, so proactive monitoring of network traffic for unusual outbound requests is recommended.