CVE-2026-44489: Axios Prototype Pollution & Proxy Credential Injection

Technical summary

The vulnerability stems from unsafe object merging in utils.merge() that creates nested proxy configuration objects as plain JavaScript objects with Object.prototype in their prototype chain. The setProxy() function (lib/adapters/http.js, lines 209–223) reads proxy.username, proxy.password, and proxy.auth properties without using hasOwnProperty() checks to distinguish own properties from inherited ones. An attacker who can influence the application's prototype chain—through prototype pollution elsewhere in the codebase or supplied data—can set Object.prototype.username or Object.prototype.password. When setProxy() constructs the Proxy-Authorization header, it reads these polluted inherited properties and injects them into HTTP requests transiting a proxy. This violates the integrity of request headers and may leak or corrupt authentication credentials.

Business impact

For applications using Axios with HTTP proxies, this vulnerability creates a covert channel for header injection. The practical impact depends on proxy configuration and network topology. An attacker capable of prototype pollution could inject false proxy credentials, potentially causing authentication failures, triggering alerts, or—in some scenarios—enabling credential harvesting if logs or monitoring systems capture the injected headers. The severity is moderate because successful exploitation requires both prototype pollution and a proxied HTTP setup; however, if those conditions exist, the attacker gains persistent influence over proxy authentication for all subsequent requests without further interaction.

Affected systems

Axios versions 1.15.2 through 1.15.x are vulnerable. Version 1.16.0 and later include the fix. Any Node.js or browser-based application using Axios in the vulnerable version range is in scope. Risk is elevated for applications that (1) use Axios with explicit proxy configuration, (2) process untrusted JSON or deserialized objects that might be supplied by an attacker, or (3) run code from third-party libraries that could perform prototype pollution.

Exploitability

Exploitation requires two prerequisites: (1) the ability to pollute Object.prototype, typically through a prototype pollution gadget in application code or a vulnerable dependency, and (2) Axios configured to use an HTTP proxy. The CVSS score of 3.7 (LOW) reflects the high bar for achieving prototype pollution, which is not trivial in most well-written applications. However, given the widespread use of Axios and the prevalence of prototype pollution vulnerabilities in the npm ecosystem, opportunistic attacks targeting vulnerable supply chains or misconfigured applications are plausible. No public exploit code is known to be weaponized as of the advisory date.

Remediation

Update Axios to version 1.16.0 or later. No workaround mitigates the root cause if prototype pollution is possible; remediation must occur at the library level. After patching, verify that the application does not use object deserialization patterns that could still introduce prototype pollution from other sources.

Patch guidance

Upgrade Axios to version 1.16.0 or later. For Node.js projects, run `npm update axios` or `yarn upgrade axios`. For browser-based projects using Axios via CDN or bundler, ensure your build process pulls the latest version. Test the upgrade in a staging environment to confirm no breaking changes (the patch is backward compatible). If you cannot immediately upgrade, audit your application for prototype pollution vulnerabilities in dependent code and consider disabling proxy authentication where possible.

Detection guidance

Monitor for signs of prototype pollution in application logs and network traffic. Look for anomalous Proxy-Authorization headers in HTTP requests that do not match your configured proxy credentials. Implement Object.prototype integrity checks (e.g., Object.freeze(Object.prototype) in trusted startup code) to prevent runtime pollution, though this is a defense-in-depth measure and not a substitute for patching. Dependency scanning tools (npm audit, Snyk, etc.) should flag Axios < 1.16.0; integrate these into your CI/CD pipeline.

Why prioritize this

Although the CVSS score is LOW (3.7), the vulnerability merits swift attention because (1) Axios is a critical ecosystem dependency with millions of downloads weekly, (2) prototype pollution is a known attack vector in JavaScript, and (3) proxy misconfiguration is common in enterprise environments. Organizations using Axios with proxies should patch within their normal critical-dependency update window. For projects without proxy usage, the risk is negligible but upgrading incurs minimal cost.

Risk score, explained

The CVSS:3.1 score of 3.7 (LOW) reflects a network-accessible attack requiring high complexity (prototype pollution is non-trivial) and no privileges. The impact is limited to integrity (header injection) with no confidentiality or availability impact in isolation. However, context matters: organizations whose architecture involves both Axios and prototype pollution risk vectors should weight this higher internally.

Frequently asked questions