OpenSSL 3.6 added LMS verification, with a narrower scope than signing

OpenSSL's October 1, 2025 version 3.6 release announcement added support for LMS signature verification in its default and FIPS providers. The operative word is verification. The announcement does not claim that the added feature supplies an LMS signing workflow.

That is a small wording distinction with a large migration consequence. A system that checks signatures and a system that creates them need different evidence of readiness.

Put the operation beside the algorithm

NIST's SP 800-208, an older background standard published in 2020, covers stateful hash-based signature schemes including LMS. It supplies the standards context for the feature named in the release.

An inventory entry that only says “LMS supported” is therefore incomplete. An operator needs to know whether the component verifies, signs, manages keys or merely recognizes an identifier. The software version and selected provider belong beside that description.

The release's mention of a FIPS provider also should not be converted into an unsupported claim about the validation status of a particular deployed module. This article has not checked a deployment's module certificate or configuration.

A focused evaluation

For a team considering a verification feature, a useful test plan would identify representative signed artifacts, expected acceptances and rejections, and the application's behavior when verification fails. That is an editorial suggestion, not a report of tests performed for this article.

The signing side needs its own implementation and key-management review. A successful verifier test cannot fill that gap, just as installing a library does not establish that an application calls the new functionality.

This is a retrospective on a release within the past year, not a recommendation to install version 3.6 as the newest or preferred branch. Current support and upgrade decisions require a separate review.

The lasting lesson from the release is concrete: write the cryptographic operation into the feature record. Doing so makes a narrowly useful addition visible without allowing it to imply a complete post-quantum migration.