CVE-2026-18141 Overview
A flaw in aap-gateway, a component of Red Hat Ansible Automation Platform's Event-Driven Ansible (EDA), allows unauthenticated remote attackers to bypass mutual Transport Layer Security (mTLS) authentication on event streams. Attackers manipulate the event stream URL and forge the HTTP Subject header to impersonate a trusted client. Error messages returned by the service disclose the expected certificate subject, reducing attacker guesswork. Successful exploitation lets an attacker inject arbitrary events into EDA and potentially trigger automated workflows across managed infrastructure. The weakness is tracked under [CWE-295] (Improper Certificate Validation).
Critical Impact
Unauthenticated attackers can inject forged events into Event-Driven Ansible and trigger automation workflows, undermining the integrity of downstream operations.
Affected Products
- Red Hat Ansible Automation Platform (aap-gateway component)
- Event-Driven Ansible (EDA) event stream endpoints
- Deployments referenced in Red Hat advisories RHSA-2026:50336, RHSA-2026:50340, and RHSA-2026:50479
Discovery Timeline
- 2026-07-31 - CVE-2026-18141 published to NVD
- 2026-08-04 - Last updated in NVD database
Technical Details for CVE-2026-18141
Vulnerability Analysis
Event-Driven Ansible exposes event stream endpoints protected by mTLS. The aap-gateway component enforces client identity by inspecting an HTTP Subject header rather than deriving the subject directly from the validated TLS peer certificate. Because the header is attacker-controllable, a request can carry any subject string while the underlying TLS channel bypasses proper peer verification. Combined with URL manipulation of the event stream path, unauthenticated network attackers reach an endpoint that should require a trusted client certificate.
The issue is compounded by verbose error handling. When authentication fails, the service returns messages that disclose the exact expected certificate subject value. Attackers use this leaked identifier to craft a forged Subject header that satisfies the flawed check on the next request. Once accepted, injected events flow into EDA rulebooks and can drive automated actions such as playbook execution.
Root Cause
The root cause is improper certificate validation. Authentication logic trusts a client-supplied HTTP header instead of binding the identity check to the cryptographically validated TLS client certificate. Error responses further weaken the control by disclosing the expected subject, turning a hardened mTLS channel into a header-based access decision.
Attack Vector
Exploitation is network-based, requires no privileges, and needs no user interaction. An attacker reaches the EDA event stream endpoint over the network, sends a probe to harvest the expected subject from the returned error message, then replays the request with a forged Subject header and a manipulated event stream URL. The request is accepted and arbitrary event payloads are ingested by EDA.
No public proof-of-concept code is available. See the Red Hat CVE Analysis for CVE-2026-18141 and Red Hat Bug Report #2508155 for vendor technical detail.
Detection Methods for CVE-2026-18141
Indicators of Compromise
- Inbound HTTP requests to EDA event stream endpoints that carry an explicit Subject header from external or untrusted networks.
- Repeated 4xx responses from aap-gateway containing certificate subject strings, followed shortly by a successful 2xx response from the same source.
- EDA rulebook activations tied to events that lack a corresponding upstream producer or source system record.
Detection Strategies
- Alert on any request to EDA event stream URLs where the client TLS certificate subject does not match the Subject header value.
- Correlate aap-gateway access logs with mTLS handshake logs to flag requests that were accepted without a valid client certificate chain.
- Hunt for anomalous playbook or automation executions triggered outside of change windows or without a matching upstream event source.
Monitoring Recommendations
- Enable verbose logging on the reverse proxy or ingress fronting aap-gateway to capture TLS peer certificate details alongside HTTP headers.
- Forward EDA and gateway logs to a central SIEM and retain them long enough to reconstruct workflow execution chains.
- Monitor for error responses that echo certificate subjects and treat repeated occurrences from a single source as reconnaissance.
How to Mitigate CVE-2026-18141
Immediate Actions Required
- Apply the fixes referenced in RHSA-2026:50336, RHSA-2026:50340, and RHSA-2026:50479 to all Ansible Automation Platform installations.
- Restrict network exposure of EDA event stream endpoints to trusted producers only, using firewall rules or private network segments.
- Rotate any client certificates whose subjects may have been disclosed through error messages during the exposure window.
Patch Information
Red Hat has released patched builds of Ansible Automation Platform through the advisories listed above. Refer to the Red Hat CVE Analysis for CVE-2026-18141 for the affected package versions and the fixed releases mapped to each supported product stream. Apply updates through your normal Red Hat subscription channels and restart the aap-gateway service after upgrade.
Workarounds
- Place EDA event stream endpoints behind a reverse proxy that strips or overrides the incoming Subject header before it reaches aap-gateway.
- Enforce network-layer allow lists so only known event producer IP ranges can reach the event stream endpoints until patches are applied.
- Disable or suppress detailed error responses at the proxy layer to prevent disclosure of expected certificate subjects.
Disclaimer: This content was generated using AI. While we strive for accuracy, please verify critical information with official sources.
