AI Agent Engineering 101

Seoul National University of Science and Technology, Fall 2026. This repository holds the lecture notes and is where all coursework is submitted.

Instructor: Jaegyu Lee (@Q00)

If this repository is useful to you, please star it ⭐ before you fork — it helps other students and future cohorts find it.

Lecture notes

index.html is the course home; week-NN.html is that week's note. Weeks 01 through 03 are up; the rest are published as the semester goes on.

How submission works: fork and PR

Every submission in this course is a GitHub pull request. Two reasons:

1. PR timestamps cannot be forged. Commit times can be rewritten locally, but the time a PR was opened is recorded by GitHub's servers. Deadlines are judged by PR open time.
2. Commit history is evidence of process. This course grades the process as much as the result. What you tried and what you threw away must be visible in the history.

One-time setup

# 1. Fork this repository (Fork button on GitHub)
# 2. Clone your fork
git clone https://github.com/<your-id>/ai-agent-engineering-101.git
cd ai-agent-engineering-101
git remote add upstream https://github.com/Q00/ai-agent-engineering-101.git

Every week

git fetch upstream && git merge upstream/main   # pull the latest assignment
git switch -c week-01                           # one branch per week
# ... work ...
git push origin week-01
# Open a PR to upstream. Title: [week-01] 23512345

Repository layout

index.html         Course home. week-NN.html is that week's lecture note.
roster/            One file per student. Created in the week-01 lab.
weeks/week-NN/     Assignment specs and starter code. Read-only.
submissions/       Your work. submissions/<student-id>/week-NN/ is your territory.
scripts/           The same checks CI runs, runnable locally.

One rule: do not touch anything outside submissions/<your-student-id>/. PRs that modify weeks/ or another student's submission directory are rejected by the checks.

Automated checks

Opening a PR triggers structural checks on GitHub Actions: the code parses, required files exist, and the spec (e.g. tool count) is met. Run the same checks locally before pushing:

python scripts/check_week01.py submissions/<student-id>/week-01

Passing checks is the minimum bar for submission, not a grade. Actual grading — does the code solve a real task, is the process honestly recorded — is done by a human.

Commit rules

• Do not squash. Leave broken intermediate commits as they are. A messy history is not a penalty; it is grading evidence.
• Commit run logs (console output) to submissions/<student-id>/week-NN/logs/.
• Never commit API keys. Environment variables only.

LLM policy

LLMs and agent tools (Claude Code, Codex, opencode, anything) are free to use on every assignment. AGENTS.md and CLAUDE.md in this repository exist so those tools understand the rules here. The one condition: what you asked for and what you discarded must remain in the commit history and logs.

License

The lecture notes, figures, assignment specs, and course code in this repository are licensed under CC BY-NC-SA 4.0. You may read, copy, modify, and share them for study and other non-commercial purposes, as long as you credit the source and keep the same license on anything you derive from them.

Using this material to teach a course, whether at a school, a company, or a paid program, is not covered by that license. If you want to do that, email jaegyu.lee@ourolabs.ai first and describe the setting. Commercial use without permission is a copyright infringement and will be pursued.

submissions/ is excluded: each student owns the work in their own directory.