LibreSSL ocspcheck CVE-2026-118372

Exploitation requires that LibreSSL ocspcheck (version ≤4.4.0) be used to process an attacker-controlled certificate containing an AIA OCSP URI whose path includes CR/LF control characters; the specific trigger is that url2host() copies the URI path without rejecting control characters and http_open() places it unescaped into the HTTP POST request line sent to the OCSP responder. Limiting factors: the attacker cannot reach the vulnerable code directly but depends on the victim invoking ocspcheck against a malicious certificate (reflected by AT:P in the CVSS 4.0 vector), impact is confined to HTTP integrity (header injection / request splitting) against the responder host with no confidentiality or availability effect, and systems on LibreSSL 4.4.0 or later with the control-character filter are not affected.

This is a medium-severity (CVSS 4.0 base 6.3) CRLF/HTTP-header-injection flaw (CWE-93) in LibreSSL's ocspcheck utility, and the real-world risk is moderate and conditional rather than urgent. The supplied CVSS 4.0 vector (AV:N/AC:L/AT:P/PR:N/UI:N) encodes the key nuance: although it is network-reachable and needs no privileges, AT:P marks a genuine attack requirement - the attacker does not talk to ocspcheck directly but must get it to process an attacker-controlled X.509 certificate whose AIA OCSP URI contains embedded CR/LF. Impact is limited to integrity (VC:N/VI:L/VA:N): the attacker can inject headers into or split the HTTP POST that ocspcheck sends to the OCSP responder, with no confidentiality or availability loss and no scope change to a subsequent system (SC/SI/SA:N). Signals are consistent and modest: no KEV listing is present, no EPSS score was provided, and no verified public exploit code is supplied - the VulnCheck entry is an advisory reference, not working exploit code. The vendor-assigned priority is 'this_month,' aligning with the medium score and the conditional, integrity-only impact. Net: a legitimate patch-management item, not a high-CVSS-but-overhyped critical nor an emergency. EPSS and KEV data are absent and should be confirmed before final prioritization.

An attacker crafts an X.509 certificate whose Authority Information Access OCSP URI path embeds CR/LF sequences, then induces a victim running LibreSSL ocspcheck to validate that certificate. Because url2host() copies the path without rejecting control characters and http_open() places it in the POST request line, the injected bytes split or add headers in the OCSP request sent to the responder. No verified public exploit code is provided - only a VulnCheck advisory reference exists.