CVE-2026-42447: HTML injection — MEDIUM

CISO Take

jadx-gui, the GUI for the widely used jadx Dex-to-Java decompiler, fails to escape architecture and per-architecture-count strings pulled from native library (.so) file path components inside an analyzed APK, so a crafted, URL-encoded ZIP entry name renders as live HTML in the Summary panel. An analyst who opens a malicious APK can trigger out-of-band requests that disclose their IP address and can be used to probe or interact with locally exposed services and applications on the analyst's machine. The blast radius is narrow — CVSS 3.6 (local, user interaction required, no direct confidentiality/integrity/availability impact beyond IP disclosure), EPSS sits at 0.14%, it is not in CISA KEV, and no public exploit or scanner template exists, so this is not an urgent patch-now situation for most teams. It matters specifically for security, threat-intel, and AI red-team teams who routinely decompile untrusted Android apps (including ones bundling on-device AI/ML models or SDKs) as part of triage, since the malicious artifact being analyzed is the attack delivery mechanism. Action: upgrade jadx-gui to 1.5.6 or later before opening any APK from an untrusted or adversary-controlled source, and in the meantime analyze suspicious APKs in an isolated/network-segmented VM rather than on an analyst workstation with local services exposed.

Sources: NVD EPSS GitHub Advisory

What is the risk?

Low overall risk. Exploitability is gated by user interaction (the analyst must actively open a malicious APK in jadx-gui) and by local attack vector, keeping the CVSS score at 3.6. There is no privilege escalation, no code execution, and no data destruction — the confirmed impacts are HTML rendering, out-of-band network requests, IP disclosure, and interaction with locally exposed applications. EPSS (0.14%, ~96th percentile relative ranking but still a very low absolute score) and the absence of a KEV listing, public PoC, or Nuclei template all indicate this is not being actively weaponized. The realistic risk scenario is targeted: an adversary who knows a specific analyst or team uses jadx-gui to triage suspicious APKs (a common workflow in AI-app security review and malware/mobile threat research) could weaponize a decompiled sample specifically to deanonymize that analyst or pivot into their local network.

How does the attack unfold?

Malicious artifact delivery

Adversary crafts an APK with a URL-encoded, HTML-injecting ZIP entry name embedded in a native library (.so) path and delivers it to a target analyst for triage.

AML.T0011

User opens artifact

The analyst opens the malicious APK in jadx-gui and views the Summary tab, which renders the unescaped arch/perArchCount strings as live HTML.

Out-of-band execution

The rendered HTML triggers an out-of-band network request to an attacker-controlled endpoint, disclosing the analyst's IP address.

Local network reconnaissance

The injected HTML is further used to probe or interact with locally exposed applications or services on the analyst's machine or network.

Malicious artifact delivery

Adversary crafts an APK with a URL-encoded, HTML-injecting ZIP entry name embedded in a native library (.so) path and delivers it to a target analyst for triage.

AML.T0011

User opens artifact

The analyst opens the malicious APK in jadx-gui and views the Summary tab, which renders the unescaped arch/perArchCount strings as live HTML.

Out-of-band execution

The rendered HTML triggers an out-of-band network request to an attacker-controlled endpoint, disclosing the analyst's IP address.

Local network reconnaissance

The injected HTML is further used to probe or interact with locally exposed applications or services on the analyst's machine or network.

How severe is it?

CVSS 3.1

5.0 / 10

EPSS

0.2%

chance of exploitation in 30 days

Higher than 10% of all CVEs

Source: EPSS v3 — FIRST.org

Exploitation Status

Exploit Available

Exploitation: MEDIUM

Sophistication

Trivial

Exploitation Confidence

medium

○ CISA SSVC: Public PoC

○ Public PoC indexed (trickest/cve)

Composite signal derived from CISA KEV, VulnCheck KEV, CISA SSVC, EPSS, Metasploit, Exploit-DB, trickest/cve, Nuclei templates, and inthewild.io exploitation reports.

What is the attack surface?

AV Local

AC Low

PR None

UI Required

S Changed

C Low

I Low

A None

What should I do?

1 step

1) Upgrade jadx/jadx-gui to version 1.5.6 or later, which fixes the unescaped rendering in SummaryNode.java. 2) Until patched, avoid opening APKs from untrusted, unverified, or adversary-controlled sources directly in jadx-gui; prefer the CLI (jadx, non-GUI) for triage of unknown samples since the vulnerability is specific to the GUI's Summary panel rendering. 3) Analyze suspicious/untrusted APKs inside an isolated, network-egress-restricted VM or sandbox so any triggered out-of-band request cannot reach real internal services or disclose the analyst's true IP. 4) Detection: monitor for unexpected outbound HTTP/DNS requests originating from analyst workstations at the time jadx-gui is used to open a new sample, and treat unusual ZIP entry names (HTML/URL-encoded) inside submitted APKs as a suspicious indicator during static triage.

What does CISA's SSVC say?

Decision Track*

Exploitation poc

Automatable No

Technical Impact partial

Source: CISA Vulnrichment (SSVC v2.0). Decision based on the CISA Coordinator decision tree.

How is it classified?

Data Leakage Privacy Violation Framework AML.T0011 - User Execution

Which compliance frameworks are affected?

Compliance analysis pending. Sign in for full compliance mapping when available.

Frequently Asked Questions

What is CVE-2026-42447?

jadx-gui, the GUI for the widely used jadx Dex-to-Java decompiler, fails to escape architecture and per-architecture-count strings pulled from native library (.so) file path components inside an analyzed APK, so a crafted, URL-encoded ZIP entry name renders as live HTML in the Summary panel. An analyst who opens a malicious APK can trigger out-of-band requests that disclose their IP address and can be used to probe or interact with locally exposed services and applications on the analyst's machine. The blast radius is narrow — CVSS 3.6 (local, user interaction required, no direct confidentiality/integrity/availability impact beyond IP disclosure), EPSS sits at 0.14%, it is not in CISA KEV, and no public exploit or scanner template exists, so this is not an urgent patch-now situation for most teams. It matters specifically for security, threat-intel, and AI red-team teams who routinely decompile untrusted Android apps (including ones bundling on-device AI/ML models or SDKs) as part of triage, since the malicious artifact being analyzed is the attack delivery mechanism. Action: upgrade jadx-gui to 1.5.6 or later before opening any APK from an untrusted or adversary-controlled source, and in the meantime analyze suspicious APKs in an isolated/network-segmented VM rather than on an analyst workstation with local services exposed.

Is CVE-2026-42447 actively exploited?

Proof-of-concept exploit code is publicly available for CVE-2026-42447, increasing the risk of exploitation.

How to fix CVE-2026-42447?

1) Upgrade jadx/jadx-gui to version 1.5.6 or later, which fixes the unescaped rendering in SummaryNode.java. 2) Until patched, avoid opening APKs from untrusted, unverified, or adversary-controlled sources directly in jadx-gui; prefer the CLI (jadx, non-GUI) for triage of unknown samples since the vulnerability is specific to the GUI's Summary panel rendering. 3) Analyze suspicious/untrusted APKs inside an isolated, network-egress-restricted VM or sandbox so any triggered out-of-band request cannot reach real internal services or disclose the analyst's true IP. 4) Detection: monitor for unexpected outbound HTTP/DNS requests originating from analyst workstations at the time jadx-gui is used to open a new sample, and treat unusual ZIP entry names (HTML/URL-encoded) inside submitted APKs as a suspicious indicator during static triage.

What systems are affected by CVE-2026-42447?

This vulnerability affects the following AI/ML architecture patterns: mobile/edge AI app deployments (APKs bundling on-device ML runtimes or native inference libraries), AI/malware security tooling and reverse-engineering workflows.

What is the CVSS score for CVE-2026-42447?

CVE-2026-42447 has a CVSS v3.1 base score of 5.0 (MEDIUM). The EPSS exploitation probability is 0.20%.

What is the AI security impact?

Affected AI Architectures

mobile/edge AI app deployments (APKs bundling on-device ML runtimes or native inference libraries)AI/malware security tooling and reverse-engineering workflows

MITRE ATLAS Techniques

AML.T0011 User Execution

What are the technical details?

Original Advisory

jadx is a Dex to Java decompiler. Prior to 1.5.6, jadx-gui is affected by an HTML injection vulnerability in the Summary tab because SummaryNode.java appends arches and perArchCount values derived from .so file path components inside an APK into an HTML panel without escaping. A malicious APK with an HTML URL-encoded ZIP entry name can force rendering of arbitrary HTML, perform out-of-band requests, disclose the victim IP address, or interact with locally exposed applications. This issue is fixed in version 1.5.6.

Exploitation Scenario

An adversary crafts a malicious Android APK containing a native library path whose filename embeds URL-encoded HTML/JavaScript-equivalent markup. The APK is distributed to a target analyst — for instance, submitted as a 'suspicious app sample' to a security or AI-app-vetting team, or planted in a research dataset the team is expected to triage. When the analyst opens the APK in jadx-gui and views the Summary tab, the injected content renders as live HTML, issuing an out-of-band request to an attacker-controlled server. This reveals the analyst's public IP address and can be extended to probe or interact with services listening on localhost or the internal network (e.g., dev servers, internal dashboards), giving the adversary reconnaissance on the analyst's environment purely from a file the analyst chose to inspect.

Weaknesses (CWE)

CWE-79 Improper Neutralization of Input During Web Page Generation ('Cross-site Scripting') Primary CWE-79 Improper Neutralization of Input During Web Page Generation ('Cross-site Scripting')

CWE-79 — Improper Neutralization of Input During Web Page Generation ('Cross-site Scripting'): The product does not neutralize or incorrectly neutralizes user-controllable input before it is placed in output that is used as a web page that is served to other users.

[Architecture and Design] Use a vetted library or framework that does not allow this weakness to occur or provides constructs that make this weakness easier to avoid [REF-1482]. Examples of libraries and frameworks that make it easier to generate properly encoded output include Microsoft's Anti-XSS library, the OWASP ESAPI Encoding module, and Apache Wicket.
[Implementation, Architecture and Design] Understand the context in which your data will be used and the encoding that will be expected. This is especially important when transmitting data between different components, or when generating outputs that can contain multiple encodings at the same time, such as web pages or multi-part mail messages. Study all expected communication protocols and data representations to determine the required encoding strategies. For any data that will be output to another web page, especially any data that was received from external inputs, use the appropriate encoding on all non-alphanumeric characters. Parts of the same output document may require different encodings, which will vary depending on whether the output is in the: etc. Note that HTML Entity Encoding is only appropriate for the HTML body. Consult the XSS Prevention Cheat Sheet [REF-724] for more details on the types of encoding and escaping that are needed. HTML body Element attributes (such as src="XYZ") URIs JavaScript sections Casca

Source: MITRE CWE corpus.