Today we released OpenFastTrace 4.11.0 and OpenFastTrace Gradle Plugin 3.5.0. Both releases share one theme: making OFT extension plugins easier to use.

# OpenFastTrace 4.11.0

Until now, extension plugins had to be installed in OFT’s predefined plugin directory. With #610, embedders can now pass plugin JARs to OFT at runtime through the Java API:

Oft.builder().addPlugin("asciidoc", pluginJarFile).build();

This means plugins resolved through an external dependency mechanism — a build tool, a container image, or your own classloader — no longer need to touch the plugin directory at all.

Today, the OpenFastTrace repository reached 200 stars on GitHub.

We are especially proud of this milestone because requirement tracing is still a narrow niche in software development, even more so for a non-commercial open-source tool. It is encouraging to see that OFT resonates with people who care about keeping requirements, designs, code and tests connected.

Thank you to all our users and contributors who have gone with us through more than ten years of OFT. You are the best!

Last Sunday, the OpenFastTrace team met with Prof. Dr. Oliver Kopp, Felix Gorke and Lennard Solterbeck to talk about the background, history and design of OFT. It was a good opportunity to compare notes on traceability tools and the decisions that shape them.

# A Language Server for OFT

Felix donated the OpenFastTrace language server to the itsallcode.org organization. He originally developed it for his bachelor’s thesis, and we are very grateful that he is making it available to the community.

After feedback from our community, we have decided to release the AI skills under the MIT License.

To make this more clear, we moved the skills to a separate repository:

https://github.com/itsallcode/openfasttrace-ai-skills

Please note that we will replace the skills in the OpenFastTrace repository with a link to the new repository soon.

OpenFastTrace itself will remain licensed under the GNU General Public License v3.

AI skills often get heavily modified and customized by users, which can make it difficult to maintain a single, consistent codebase. By chosing the MIT license for the AI skills, we want to signal that we don’t expect users to contribute their modifications back to the original codebase. Of course, we still welcome improvements that everyone in the community can benefit from.

A lot happened in the world of OpenFastTrace in the last few weeks. OFT finally got its own website: https://openfasttrace.itsallcode.org. We use Jekyll to generate the site from the documentation in the OpenFastTrace repository.

We also have a draft for OFT’s own exchange and report format that will eventually replace the SpecObject format inherited from ReqM2.

Oh, and of course we have a fresh new release with 4.10.0.

But the biggest change is that we now also publish to Homebrew and have our own Debian repository so installing and updating OFT got a lot easier.