openssl_encrypt CVE-2026-81700

Exploitation requires that the attacker control a GPG signing key that the victim's keyring already trusts AND that is now in a revoked or expired state - specifically a compromised-then-revoked signing key or an expired project key - because the bug is that verify_detached accepts precisely these otherwise-invalid keys. The attacker must also be able to get the target to load a plugin whose detached signature is checked by the vulnerable gpg_runner.verify_detached path. What LIMITS exploitation: the key must have been previously valid/trusted (an attacker cannot use an arbitrary self-generated untrusted key, since that would not produce VALIDSIG), and the deployment must actually use openssl_encrypt's GPG-signed plugin verification to make a trust decision; installations that do not load externally supplied signed plugins are not exposed. No target-side authentication or user interaction is required per the CVSS vector (PR:N/UI:N).

The signals are mostly aligned toward a genuinely serious flaw, with one important nuance. The CVSS 4.0 vector (AV:N/AC:L/PR:N/UI:N with VC:H/VI:H/VA:H) describes an unauthenticated, network-reachable path to full confidentiality, integrity, and availability compromise via code execution in the host process, which supports the 9.3 rating. The tempering factor absent from the raw vector is the real prerequisite: an attacker must hold a signing key that was once trusted by the victim and is now revoked or expired - this is a meaningful gate, though revoked public keys are frequently recoverable and expired project keys are common, so the barrier is lower than it first appears. There is no EPSS score, no POC, and no KEV entry in the provided data, so exploitation is currently theoretical rather than observed; this suggests the risk is elevated by exposure and impact rather than by demonstrated in-the-wild activity, and even a mid-sized operator that ships plugins signed by rotating or retired project keys should treat this as urgent because their older signatures may now verify against exactly the kind of expired key this bug accepts. Overall this is a real priority for anyone relying on openssl_encrypt's plugin signature gate for a security decision, not merely a high-CVSS-but-low-real-risk finding - but it is not (yet) evidence of active exploitation.

An attacker obtains a signing key that a target's openssl_encrypt deployment previously trusted - for example a project key that has since expired, or a compromised key that was subsequently revoked - and uses it to sign a malicious plugin. They deliver that plugin through whatever channel the application loads plugins from; because verify_detached only checks for VALIDSIG and ignores the revocation/expiry status, the signature is accepted and the malicious plugin executes in the host process. No public POC is identified at time of analysis, and the low attack complexity means no additional exploitation hurdles once a suitable revoked or expired key is in hand.