A critical vulnerability in flyto-core allows unauthenticated attackers to perform Server-Side Request Forgery and exfiltrate sensitive internal secrets.

This vulnerability involves a lack of authentication on the POST /run endpoint, which allows an attacker to supply a malicious callback URL. This triggers an outbound request containing the internal runner secret, leading to SSRF and credential exposure.

The exposure of internal runner secrets combined with SSRF capabilities presents a high risk of lateral movement and unauthorized access to cloud or internal resources. With a CVSS score of 9.3, this flaw poses a severe threat to the integrity and confidentiality of automation workflows, potentially allowing attackers to compromise the entire AI-agent infrastructure.

Immediate Action: Upgrade flyto-core to version 2.26.7 or later to implement required authentication and input validation.

Proactive Monitoring: Review access logs for suspicious POST requests to the /run endpoint and monitor for unexpected outbound traffic from the verification service.

Compensating Controls: Implement network-level egress filtering to restrict unauthorized outbound connections from the flyto-core service until the patch is applied.

Public Exploit Available: No (exploit_available: false)

Given the critical severity and the existence of a proof-of-concept, organizations must prioritize upgrading to version 2.26.7 immediately. Failure to patch this vulnerability leaves the automation kernel exposed to remote command and control by unauthorized actors.