Pull Request (Recommended)
Generates assessment questions whenever a student opens or updates a pull request. Questions are posted as a PR comment so the instructor can see them inline alongside the submitted code.
Push to Branch
Generates assessment questions on every push to a non-default branch. Useful when students work directly on a feature or personal branch without opening a pull request. The output file is written back to the repository.
Post to GitHub Issues
Creates a GitHub Issue for each assessment run. Each issue title includes the branch name and head commit SHA so runs are distinct and searchable.
Post to GitHub Discussions
Creates a GitHub Discussion for each assessment run. Useful for class-wide visibility or when instructors prefer a forum-style record of assessments.
Third-Party AI Provider (OpenAI / OpenRouter)
Uses an external AI provider instead of GitHub Models. Useful when a higher question count, a specific model, or a different provider is required.
Azure OpenAI Provider
Uses an Azure OpenAI deployment for question generation. Suitable when institutional policy requires data to stay within an Azure tenant, or when a specific Azure deployment is already provisioned.
All Inputs
A fully annotated workflow showing every available input. Inputs that don't apply to the pull-request scenario are commented out with an explanation of when they would be used instead.
Assignment Context
Automatically injects assignment files (README, brief, rubric, style guide, etc.) into the AI prompt so questions are targeted to the specific requirements of the assignment — without manually copying content into additional_context.
Instructor Repository Delivery
Automatically stores a private, instructor-only copy of every assessment — including both questions and answers — in a dedicated repository. Students only ever see their student-facing report; the instructor copy is written separately using a PAT that students cannot access.