AI Integration¶
FrontISTR is pursuing initiatives that combine AI agents with a variety of software-related tasks, including environment setup, running analyses, and improving code. The range of situations in which AI can be used is expected to expand further, including support for model creation and integration with external tools. This page introduces these AI integration initiatives.
Fabrica — A Workshop for Using AI Agents¶
FrontISTR-Fabrica is a workshop (Fabrica) for using AI agents, intended for both FrontISTR users and developers. Fill in the requirements in a prompt template prepared for each task and give it to an AI agent; the AI then creates a work plan, executes the task while requesting confirmation as needed, and summarizes the results in a report. This allows the user to focus on reviewing the results and making decisions.
Using the templates, you can interactively carry out tasks such as the following.
- Environment setup — From investigating build settings through compilation and operation checks
- Running analyses — Checking meshes, creating input files, running the solver, and organizing results
- Feature additions and bug fixes — From code investigation through implementation, testing, and commits
- Documentation — Writing manuals and checking consistency with the code
- Contributing — Drafting issues and merge requests, and guidance on pushing to a fork
To use Fabrica, you need an environment in which FrontISTR can be built and run, as well as an AI agent such as GitHub Copilot, Claude Code, or Codex. For installation procedures, prerequisites, and instructions for each template, see the repository README.
The user is responsible for the final decision on whether to accept commands or code proposed by the AI. The templates are intended to make that review easier; they do not guarantee the correctness of AI-generated output.