Critical Flyto-Core SSRF Flaw Found

Security researchers have identified a critical vulnerability, tracked as CVE-2026-67426, affecting the Flyto-Core verification service. The flaw allows unauthenticated remote attackers to perform server-side request forgery (SSRF) and exfiltrate internal runner secrets by sending malicious requests to the service's API.

The vulnerability exists because the flyto-verification service exposes a POST endpoint at /run without any authentication requirements. Furthermore, the service fails to validate the callback_url parameter, allowing it to be used for unauthorized outbound requests. Because the service unconditionally attaches an internal secret to these requests, an attacker can capture this sensitive data and use it to impersonate the service in subsequent operations.

What's at Risk

The vulnerability impacts all versions of pip/flyto-core prior to version 2.26.7. Organizations that have deployed this service in an internet-facing configuration are at the highest risk, as the service defaults to binding to all network interfaces on port 8344. This exposure allows any remote actor to interact with the service directly.

Systems that rely on this service for workflow verification or that store sensitive configuration data in the FLYTO_RUNNER_SECRET environment variable are particularly vulnerable. If an attacker successfully exfiltrates this secret, they can gain unauthorized control over the engine's callback mechanisms, effectively bypassing security controls meant to protect internal infrastructure.

How the Flaw Works

This vulnerability is a combination of SSRF and improper credential handling. SSRF occurs when an application accepts a user-supplied URL and uses it to make an outbound request without sufficient validation. In a typical scenario, this allows an attacker to interact with internal services that are not meant to be accessed from the public internet, such as cloud metadata services or internal management interfaces.

When an application also attaches internal authentication tokens or secrets to these outbound requests, the impact escalates significantly. Instead of just reading internal data, an attacker can use the stolen credentials to authenticate as the compromised service. This often leads to a complete breakdown of internal trust boundaries, allowing for lateral movement within a network or unauthorized access to backend systems that would otherwise reject unauthenticated requests.

How to Protect Your Systems

• Upgrade flyto-core to version 2.26.7 or higher immediately to apply the vendor-supplied patch.
• Restrict network access to the flyto-verification service by placing it behind a firewall or using a VPN to ensure it is not reachable from the public internet.
• Verify that the service is not bound to 0.0.0.0 and instead limit it to localhost if remote connectivity is not strictly required.
• Audit your environment for any unauthorized access logs or anomalous outbound traffic originating from the host running the verification service.
• Implement strict egress filtering to prevent internal servers from making unauthorized connections to external, attacker-controlled domains.

Given the CVSS score of 9.3, this vulnerability presents a significant risk to the integrity of affected environments. Promptly patching the software and restricting network exposure are the most effective ways to mitigate the risk of exploitation. Failure to address this flaw leaves systems open to both information disclosure and potential credential compromise, which can lead to further unauthorized access across the broader infrastructure.