[Image courtesy of DOE]
The skills make up the first set of technical resources developed through ModCon (Transformational AI Models and Data) and the American Science Cloud. Those programs build the shared AI tools and data for the national labs under the Genesis Mission, the Trump administration’s initiative to use AI to speed up scientific research. President Donald Trump has rebranded AI as “SI,” short for superintelligence. DOE’s release uses “Super Intelligence (SI)” throughout, in line with a September executive order requiring federal agencies to replace “artificial intelligence” with the term.
DOE frames this release as the first building block of the Genesis Mission’s shared research infrastructure. That building block works with commercial and open-source coding agents.
The skills follow the Agent Skills format, in which each skill is a folder built around a SKILL.md file whose metadata tells the agent when to use it. DOE says the package complements existing skill libraries that focus on coding and software engineering.
In essence, a skill is a folder of instructions, typically featuring files in Markdown format that an AI agent loads when performing a given task. Twenty-three of the 35 Genesis skills also include Python or shell scripts the agent can run; the other 12, including the job-scheduling guides for Frontier, Aurora and Perlmutter, are text only.
Researchers at seven national labs wrote them, according to attribution files in the repository. Matt Baughman of Princeton Plasma Physics Laboratory wrote the five supercomputer and scheduler skills; Ian Foster of Argonne wrote the literature-search skill; teams led from Lawrence Berkeley and Pacific Northwest national laboratories wrote the data and safety skills. Nathan Hodas of PNNL uploaded the first 291 files to GitHub on Aug. 5 in a commit titled “adding approved genesis skills.” The project’s contribution guide permits AI-assisted work with disclosure and requires a human reviewer on every pull request; eight of its 69 commits list Anthropic’s Claude or GitHub Copilot as a co-author.
The package includes 35 domain skills that instruct AI coding agents how to do lab work, such as drafting a job script for Oak Ridge National Laboratory’s Frontier supercomputer or generating a data card, a standard summary of what a dataset contains and how it was made. The project’s documentation says the skills work with Claude Code, Codex, Cursor, OpenCode, Gemini CLI and LangChain Deep Agents, though it notes that individual skills may still depend on client-specific features or external tools.
Anthropic, OpenAI and Google, the companies behind three of those agents, have all signed on as Genesis Mission partners.
The catalog includes skills for high-performance computing job schedulers and the Frontier, Perlmutter and Aurora supercomputers; data tools; AI safety analysis workflows; and clients for American Science Cloud services such as Globus Compute. There are also standalone skills for literature search and multi-agent systems. DOE says the package also covers agent evaluation and monitoring, data governance, scientific workflows and Super Intelligence (SI) safety.
The Baseline AI R&D Capabilities (BASE) team within ModCon developed the package. Pacific Northwest National Laboratory and Lawrence Berkeley National Laboratory co-lead the team.
How frontier AI labs fit into the Genesis Mission
Anthropic, OpenAI and Google were among the 24 organizations that signed memorandums of understanding with DOE for the Genesis Mission. On Oct. 8, at the White House summit, 11 companies pledged a combined $2.4 billion in “SI for science tools and compute credits,” including $200 million from OpenAI and $150 million from Google and Anthropic.
In December 2025, Anthropic announced a multi-year partnership with DOE to give the national labs access to Claude and Anthropic-built tools, including Claude Skills for scientific workflows. Anthropic’s three-year commitment, announced Oct. 8, makes Claude available to more than 15 agencies, including NASA, the National Institutes of Health and the National Science Foundation. The commitment also includes Claude, Claude Code and API credits for several hundred Genesis Mission projects.
Tell Us What You Think!
You must be logged in to post a comment.