CS3704 — Agent Template

This repository configures Opencode with course-specific agents, skills, and helper scripts for CS3704 with Dr. Chris Brown.

Quick Start:

To use this template, you will need git, npm, and an ARC_API_KEY environment variable set to your API key from Virginia Tech's Advanced Research Computing.

Setup:

git clone https://github.com/CS3704-VT/agent-template   # clone this repository
cd agent-template                                       # navigate into the project directory
baker check CS3704-VT/profiles:agent-template.yml       # ensures you have necessary dependencies; verification will not be complete until after the next steps
baker bake .                                            # installs dependencies and configurations for this repo; defaults to your local machine
baker run setup                                         # runs setup scripts for starting the agent-template

First Time:

After the initial installation and setup, run the following in the root directory of the repository for each assignment:

opencode
@assignment analyze <assignment name or link to assignment> # Integrates assignmnt details into repository context
/skills -> /learning-goal                                   # Sets the learning goals for the given assignment scope (course, assignment, or team)

For subsequent usage, just navigate into the top-level directory of the project repository and run opencode to access the agents and skills for your assignment.

Basic Workflow:

• Use to switch between agents.
  • Plan is for brainstorming and ideation and cannot edit files.
  • Build can generate code, tests, etc. with full access to edit files.
  • More agents will be introduced as the semester goes on (see PHASES.md).
• Use @ to invoke any subagents.
  • @dcbrown to answer questions about the course, explain course content, and generate quizzes.
  • @assignment to analyze assignment requirements (@assignment analyze PM1.2 or @assignment analyze <link>), verify your work (@assignment validate), and package your submission (@assignment submit).
  • More subagents will be introduced as the semester goes on (see PHASES.md).
• To use skills, type /skills and select the approporiate skill.
  • To modify your course, assignment, or team learning goals, use learning-goals (should be invoked automatically when beginning an assignment).
  • If you would like to generate a quiz or ask a question about a concept, use the quiz or concept-explain skills. Any sub-agent can also invoke these skills (e.g., @dcbrown quiz me or @dcbrown explain Agile)
  • For more in-depth learning opportunities, use learning-opportunities (should be invoked automatically during development work).
• To submit, use prompts like "@assignment submit" or "@assignment submit my work" to package your work into the appropriate format (i.e., zip file, GitHub repo, etc.). You are responsible for checking the generated requirements (the per-assignment <name>.md files, e.g. PM1.2.md) meet the assignment expectations and uploading the generated package to Canvas to complete the submission!

Full setup and usage guide: docs/STUDENT_GUIDE.md.

For your submission, replace this README with a README specific to your assignment. The template guide will stay at docs/STUDENT_GUIDE.md and this README is copied in docs/README.md.