Answer Key and Calibration

Expected judgments

The fresh restored file has Pages: 12 and Next: summarize one claim; the altered working file may remain at 120. The incomplete copy fails the intended-content check even if its first line is right. The comparison must inspect content in the restored location rather than accidentally reopening the backup.

Recovery codes are secret authentication material, not debugging evidence. Broad folder access is unnecessary for this local task. Use a known complete copy and local file operations. The state diagram distinguishes editor state, saved working file and independent copy. Same-device copies share a device-loss failure mode.

Dimension calibration

DS1: filenames alone score at most 1; correct state and full-location reasoning with the fresh example reaches 2. DS2: “the link has HTTPS” alone scores at most 1; explaining destination, scope and withheld secrets reaches 2. DS3: a copied file without reopening/content validation scores 1; actual recovery plus the detected incomplete control and honest failure boundary reaches 2.

A well-written plan without the practical observation remains pending. Repair with a third practice artifact; do not erase the first attempt.

The example hostnames are help.example.org and help.example.org.attacker.invalid; the latter is a different host, not a path at the first host. In the wrong-file incident, the altered working copy remains intact while the learner opens and checks the restored version. Require this path/content observation for DS1, not merely a definition of a file.