CVE-2026-49869 is a critical authentication-bypass vulnerability leading to unauthenticated remote code execution in Kestra, an open-source, event-driven orchestration platform for data, AI, and infrastructure workflows (27,000+ GitHub stars). Kestra's AuthenticationFilter uses a suffix match - checking whether a request path ends with "/configs" - to whitelist its public configuration endpoint from Basic Authentication, rather than validating the full path. Because the check only looks at the path's ending, an unauthenticated attacker can reach any API route that happens to end in "/configs", including routes for creating, executing, and deleting workflows, and for writing key-value secrets, none of which were intended to be exposed without credentials. Since Kestra ships with script-execution plugins (shell, Python, Node.js, and others) enabled by default, an attacker who reaches these endpoints can create and run an arbitrary workflow containing a malicious script task, achieving remote code execution as root inside the Kestra worker container without ever presenting a valid credential.

The following version(s) are affected:

Kestra before 1.0.45 – (Affected)
Kestra 1.1.0 and later, before 1.3.21 – (Affected)
Kestra 1.0.45, and Kestra 1.3.21 and later – (Not affected, patched)

Kestra is typically self-hosted (via Docker, Helm chart, or a standalone JAR) rather than centrally auto-updated by the vendor, so this fix will not reach an existing deployment on its own.

Typical attack flow:

1. Reach the authentication-bypass endpoint — An attacker sends an unauthenticated HTTP request to a reachable Kestra instance, targeting an API path ending in "/configs" - for example PUT /api/v1/main/flows/{namespace}/configs - which Kestra's Authentication Filter incorrectly whitelists based on the path's suffix alone, bypassing Basic Authentication entirely.
2. Create and execute a malicious workflow — Using the same authentication-bypass pattern, the attacker creates a workflow ("flow") containing a task that runs a shell, Python, or Node.js script - plugins Kestra enables by default - then triggers its execution via a similarly-bypassed endpoint (e.g. POST /api/v1/main/executions/{namespace}/configs), and retrieves the command output from the execution logs, achieving remote code execution as root inside the worker container without ever authenticating.

Attack vectors:

• A network-based, unauthenticated request against any internet- or network-reachable Kestra instance (typically listening on port 8080), requiring no prior access or account.
• No user interaction, no privileges, and no special access conditions are required (CVSS AV:N/AC:L/PR:N/UI:N) - a small number of crafted HTTP requests to documented, publicly-known endpoints is sufficient to achieve full compromise.

Successful exploitation may allow attackers to:

• Achieve remote code execution as root inside the Kestra worker container, and from there pivot into any systems, credentials, or data the worker has network access to.
• Perform server-side request forgery against internal network services and cloud metadata endpoints (AWS/GCP/Azure), potentially harvesting cloud credentials; overwrite or delete existing workflows and key-value secrets; and delete audit logs to cover evidence of the intrusion.

CERTVU recommends the following:

1. Apply the Vendor Patch Without Delay

   Upgrade Kestra to version 1.0.45 or 1.3.21 (whichever branch applies to your deployment), where this vulnerability is fixed (GitHub Security Advisory GHSA-5vc5-wxxq-3fjx). Because Kestra is self-hosted, this requires a deliberate redeploy - confirm the running instance has actually reached a fixed version rather than assuming an update has applied.

2. Block Network Access Until Patched

   Until patched, block network access to the instance from untrusted networks via firewall or reverse-proxy rules, since the vendor advisory documents no other workaround.

3. Audit the Estate for Kestra Deployments

   Audit the estate for Kestra deployments.

4. Treat Unpatched Instances as Potentially Compromised

   Treat any internet-reachable, unpatched instance as potentially compromised.

5. Rotate Credentials and Secrets

   Rotate credentials and secrets.

Report suspected compromise to CERTVU at This email address is being protected from spambots. You need JavaScript enabled to view it. or on telephone (678) 33380.