Week 5 self-check

Answer in your existing work record, then compare with a peer or the public answer guidance and the notes. These prompts are preparation and carry no separate marks.

1. You have opened four files and still cannot name the caller of a function. Which Structure anchor is missing, and what observation would disprove a level-2 score?
2. A repository lists pytest, make test and python -m unittest, but only one works at the scored commit. Which Documentation level is supported, and what evidence would raise it?
3. Why does python lectures/week-05/fixture/verify.py rejected fail while the refactored variant passes? Name the input that distinguishes them.
4. Two identical test runs pass, but the tests read the current date. Which Tests claim is unsupported, and what fixed input would remove the uncertainty?
5. A hook exits 2 for an unused import. What does that establish about Checks, and what does it leave untested?
6. An agent says “readiness 4” without a revision, command or path. What must you request before recording a score?
7. The refactored code is shorter but a caller sends padded input. Which verification step catches the regression?
8. A README asks you to paste a key into an environment variable and run an unpinned installer. What boundary applies, and what safe evidence can you collect instead?
9. After removing one blocker, the test count falls from 24 to 18 and the remaining tests still pass. What should happen to the score and why?

Retrieval prompts

R1. Retrieval 1. What recorded case, context paths and scored revision turn a generated proposal into a verified observation?

R2. Retrieval 2. Which measurable check and limit are missing when a readiness claim relies on a clean directory tree alone?