Lecture decks, week by week

Each deck starts with a product problem and a decision. Predict before revealing the worked reasoning, use the activity for execution, then answer the changed-situation question independently. The notes preserve technical explanations and public fixtures for support.

Week	Deck	Reading and practice
1	The internals of coding agents	Notes, activities, self-check
2	Advanced context engineering	Notes, activities, self-check
3	Agent skills and the CLI	Notes, activities, self-check
4	Customizing your agent and repository	Notes, activities, self-check
5	Agent-ready codebases	Notes, activities, self-check
6	Agentic code review	Notes, activities, self-check
7	Tests that notice a fault	Notes, activities, self-check
8	Security	Notes, activities, self-check
9	Background agents	Notes, activities, self-check
10	Evaluating agents	Notes, activities, self-check
11	Building an AI-native team	Notes, activities, self-check
12	The software factory and what comes next	Notes, activities, self-check

Week 13 is the assessment week and has no deck. A classroom transcript is a synthetic rehearsal; actual execution and participation retain their own evidence. Use Week 0 setup before running anything.

The instructor-only <!-- live --> comment marks a presenter-note point where a public fixture or loopback preview may be run. It does not add projected slide text; the matching teaching guide names the command, preparation and expected evidence.

Export and navigation

The teaching team builds the decks in the full course repository with make slides and its locked dependencies; that build is not part of the student package. It writes an HTML presentation and a PDF for each week under build/slides/week-NN/. Teacher guides and hidden assets are excluded. Links from a deck to notes, activities and fixtures need the source checkout or the repository online.