Axios EUVD-2026-88387

Exploitation requires that the same Node.js/browser process ALREADY be compromised by a separate prototype-pollution vulnerability capable of writing Object.prototype.headers - this CVE alone is a non-exploitable gadget. The application must also use axios's fetch adapter (fetch-based transport) in the affected range 1.7.0 through 1.19.x. Given that precondition, no additional user interaction or authentication against axios is needed. The dominant limiting factor is the mandatory chained prototype-pollution primitive; without it, the sanitized Request headers are never overridden.

This is a GADGET vulnerability, not a directly-triggerable flaw: the axios fetch adapter constructs a Request from sanitized options but then passes the original fetchOptions as fetch's second argument, so header sanitization can be undone - but ONLY if a SEPARATE same-process prototype-pollution flaw has already populated Object.prototype.headers. The description states this dependency explicitly, and the vendor CVSS 4.0 vector reflects it with AT:P (Attack Requirements: Present). No EPSS score, no CISA KEV listing, and no public exploit code are present in the input, so this is 'no public exploit identified at time of analysis' - the fix ships as config-hardening rather than in response to observed attacks. The vendor scored it 6.9 (CVSS 4.0) and mapped impact entirely to the subsequent system (SC:L/SI:H, with VC/VI/VA:N), consistent with header injection that corrupts a downstream request/service rather than crashing axios itself. Real-world priority is LOW-to-MEDIUM despite the ubiquity of axios: the gadget is inert without a chained prototype-pollution primitive elsewhere in the same process, which sharply narrows exploitability. The 'Information Disclosure' tag understates the primary integrity/header-injection nature (though a confidentiality angle exists via leaked authorization headers or metadata-service access). Note: the CVSS 4.0 vector shows PR:N/AC:L but AT:P - the real limiting factor is the required prototype-pollution precondition, not privileges, so raw base metrics overstate ease of standalone exploitation.

In an application that already contains a separate prototype-pollution bug, an attacker pollutes Object.prototype.headers so that axios's fetchOptions.headers resolves through inheritance. When the app then makes an axios request via the fetch adapter, the polluted headers override the sanitized Request headers - letting the attacker inject Authorization, caching, or cloud metadata-service (e.g., IMDS) headers into an otherwise legitimate outbound request. No public exploit code is identified, and exploitation is impossible without the chained pollution primitive.