Module 1 · Prove

Module completion checklist

Module 1 is complete when all of these are true. Build and the quiz are the scored gates; the rest are required evidence. Answer quality isn't auto-graded in Module 1; Module 3 adds that.

0 of 5 doneTracked in this browser

Build: all six Build tasks pass, and the live tests pass
Your first model call, Log every call, Write the assistant's rules, Give it the policy, Compare two models (rated), Record your model decision, plus pytest -m live (Lesson 4).
Break: the broken assistant's tests pass and your diagnosis is written
Lesson 7 · tests/test_break_01.py offline and live (three runs, all with the current limit), and docs/break-01.md.
Decide: all six requests answered and submitted
Lesson 8 · the pattern is scored; your reasoning is for discussion.
Quiz: 7/10 or higher
Unlimited retakes.
Day 1: docs/day-1-design.md is submitted
Ungraded, stored for Module 7.

The Build gate runs on your machine, against a real model:

pytest -m "not live"
pytest -m live