You can build anything with AI. So how do you prove it?
AI has changed software development forever. One developer can now build systems that used to require a team. When output becomes easier to generate, output alone stops telling you how the work was done, how decisions were made, or how effectively someone can use AI to solve a real problem.
What matters now is how someone works with their tools: how they break down problems, guide an agent, decide what is correct, and recover when it isn't. Those patterns show up over time in real projects, not in a single artifact or an artificial test.
Skillcraft is built for this new world. It works with your favourite AI coding agents like OpenCode and OpenAI Codex to turn git commits from real projects into verifiable evidence, so developers can earn credentials that prove their AI coding skills.
Instead of asking people to describe their capabilities, Skillcraft captures evidence from real development. It ties claims to commits, records the workflow context around the work, and verifies the result against explicit requirements. Because the evidence is anchored in git, it stays local, inspectable, and reproducible.
You can follow that chain yourself without relying on a separate system. A claim points to commits, commits reference evidence, and the evidence is evaluated against explicit requirements. If the evidence satisfies the requirements, a credential is issued. If it doesn't, it fails. The result is a different kind of signal. Instead of treating skills as static labels, Skillcraft treats them as patterns of execution. Over time, those patterns form a clearer picture of capability through verifiable credentials. Better than a resume, portfolio, or generated output ever could.
Nothing about how you work needs to change. You still build with AI, commit code, and iterate as usual. Skillcraft just turns that activity into verifiable evidence.