PHP2Shell: Read our blog on the critical RCE vulnerabilities affecting 70% of the Internet
Vulnerability DatabaseCVE-2026-111765

CVE-2026-111765: 
Java vulnerability analysis and mitigation

Overview

CVE-2026-111765 is a credential disclosure vulnerability in the A2A Java SDK (Agent2Agent Protocol) affecting the AndroidA2AHttpClient component. The flaw exists in versions >= 1.3.0.Final and < 1.3.2.Final of the org.a2aproject.sdk:a2a-java-sdk-http-client-android Maven package. It was published on October 20, 2026, with a patch released in version 1.3.2.Final. The vulnerability carries a CVSS v3.1 base score of 7.5 (High) (GitHub Advisory).

Technical details

The root cause (CWE-522: Insufficiently Protected Credentials) lies in a logic conflict within AndroidA2AHttpClient: the createConnection() method correctly sets setInstanceFollowRedirects(false) as a security measure, but the inner AndroidPostBuilder class declares private boolean followRedirects = true as its default, and the post() method subsequently calls connection.setInstanceFollowRedirects(followRedirects), overwriting the safe setting. JSONRPCTransport and RestTransport both inject authentication headers via AuthInterceptor — specifically X-API-Key, API-Key, X-Auth-Token, and X-Authentication — and invoke post() without explicitly calling followRedirects(false). Android's repackaged OkHttp strips the standard Authorization header on cross-authority redirects but retains these nonstandard headers, causing the full API key or token to be forwarded to the redirect destination (GitHub Advisory, Fix Commit). This vulnerability is a residual incomplete fix from the parent advisory GHSA-9rhm-2h4x-jwmx (Parent Advisory).

Impact

The primary impact is confidentiality loss through disclosure of a complete API key or authentication token to an unintended, potentially attacker-controlled authority. Autonomous A2A agent clients are exploitable without any human interaction, while user-invoked workflows may require a user to initiate a request. The advisory explicitly scopes the impact to credential disclosure only — there is no integrity, availability, code execution, or host compromise impact demonstrated (GitHub Advisory). Disclosed credentials could subsequently be used by an attacker to impersonate the legitimate client against the target A2A service, though that secondary abuse is outside the direct scope of this CVE.

Exploitability

There is no public proof-of-concept exploit code and no evidence of in-the-wild exploitation at this time (Feedly). The EPSS score is 0.0, and the vulnerability is not listed in the CISA Known Exploited Vulnerabilities (KEV) catalog. Exploitation requires the attacker to control or compromise the HTTP redirect destination (e.g., a malicious or compromised A2A agent endpoint), but autonomous clients require no human interaction once the redirect is issued, lowering the practical bar for exploitation in agent-to-agent workflows.

Exploitation steps

  1. Set up a malicious redirect server: Deploy two HTTP listeners — a primary endpoint (acting as the malicious or compromised A2A agent) and a credential collector endpoint at a different authority (e.g., a different host or port).
  2. Serve a redirect response: Configure the primary endpoint to respond to POST requests with an HTTP 302 redirect pointing to the collector endpoint.
  3. Target a vulnerable client: Identify an application using A2A Java SDK versions 1.3.0.Final or 1.3.1.Final with AndroidA2AHttpClient and configured API key credentials (via X-API-Key, API-Key, X-Auth-Token, or X-Authentication headers).
  4. Trigger the POST request: Cause the vulnerable client to send a synchronous POST request (via JSONRPCTransport or RestTransport) to the malicious primary endpoint. In autonomous agent workflows, no user interaction is required.
  5. Capture forwarded credentials: The AndroidPostBuilder defaults followRedirects=true, overriding the safe createConnection() setting. Android's OkHttp follows the redirect and forwards the nonstandard authentication headers to the collector endpoint, where the attacker captures the full API key or token (GitHub Advisory).

Indicators of compromise

  • Network: Unexpected outbound HTTP POST or GET requests from the A2A client to a second, unrecognized authority immediately following a request to the primary A2A agent endpoint; HTTP 302 responses from A2A agent endpoints redirecting to external or unexpected hosts.
  • Logs: Application or HTTP access logs showing a POST to the configured A2A agent URL followed by a subsequent request (method may change to GET) to a different host/port with authentication headers (X-API-Key, API-Key, X-Auth-Token, X-Authentication) present in the redirected request.
  • Behavioral: Unexpected API key usage or authentication events originating from IP addresses not associated with the legitimate client, suggesting a captured credential is being replayed (GitHub Advisory).

Mitigation and workarounds

Upgrade the A2A Java SDK to version 1.3.2.Final or later, which changes the AndroidPostBuilder default from followRedirects = true to followRedirects = false (Fix Commit, GitHub Advisory). As an immediate workaround for applications that cannot upgrade, explicitly call .followRedirects(false) on every PostBuilder instance before invoking .post(). Additionally, rotate any API keys or authentication tokens that may have been exposed in production environments running affected versions (1.3.0.Final or 1.3.1.Final).

Additional resources


Source: This report was generated using AI

Related Java vulnerabilities:

CVE ID

Severity

Score

Technologies

Component name

CISA KEV exploit

Has fix

Published date

CVE-2026-111765HIGH7.5
  • Java logoJava
  • org.a2aproject.sdk:a2a-java-sdk-http-client-android
NoYesOct 20, 2026
CVE-2026-102834HIGH7.5
  • Java logoJava
  • org.bytedeco:javacv
NoYesOct 16, 2026
CVE-2026-111763MEDIUM5.7
  • Java logoJava
  • org.a2aproject.sdk:a2a-java-sdk-parent
NoYesOct 20, 2026
CVE-2026-111762MEDIUM5.7
  • Java logoJava
  • org.a2aproject.sdk:a2a-java-sdk-transport-rest
NoYesOct 20, 2026
CVE-2026-113730MEDIUM5.6
  • Java logoJava
  • ognl:ognl
NoYesOct 21, 2026

Free Vulnerability Assessment

Benchmark your Cloud Security Posture

Evaluate your cloud security practices across 9 security domains to benchmark your risk level and identify gaps in your defenses.

Request assessment

Get a personalized demo

Ready to see Wiz in action?

"Best User Experience I have ever seen, provides full visibility to cloud workloads."
David EstlickCISO
"Wiz provides a single pane of glass to see what is going on in our cloud environments."
Adam FletcherChief Security Officer
"We know that if Wiz identifies something as critical, it actually is."
Greg PoniatowskiHead of Threat and Vulnerability Management