The U.S. Department of Energy released Genesis Agent Skills on October 15 as a public package of technical resources for scientific AI work. DOE says it is the first such release developed through ModCon and the American Science Cloud for the Genesis Mission.
The package brings together instructions, workflows and resources from the national laboratory system. Its stated topics include AI safety, agent evaluation and monitoring, data governance, scientific workflows and multi-agent systems. This is a resource collection rather than the launch of a new foundation model.
DOE describes using plain-language requests with AI coding tools to obtain guidance, review code or explore scientific problems. Its examples include data-handling tasks and code-safety evaluation. The examples describe intended uses, not evidence that generated code is correct or that a safety review is complete.
The Baseline AI R&D Capabilities team within ModCon developed the package. DOE describes ModCon as a collaboration connecting national laboratories, industry and academia, and presents the release as an early part of shared research infrastructure.
The agency says the package is freely available to researchers, developers and the public through GitLab and invites contributions and adaptation. The announcement does not establish that every external model or coding service used with it is free, approved for sensitive data or compatible with a particular institution's policies.
What to check before use: the selected resource, its current documentation, the coding tool's permissions and the data allowed in the workflow. Reusable guidance can support a research process, but the release does not demonstrate autonomous scientific reliability or replace validation of outputs and security controls.