APPENDIX H: RESUME PORTFOLIO BUILDING CAPSTONE PROJECTS

CHECKPOINTS Checkpoint 1, after step 7: the budget and retry helpers behave correctly alone. Checkpoint 2, after step 13: a full run shows one retry, one redraft, and one escalation, and the audit log tells the story. Checkpoint 3, after step 16: both automated tests pass. Checkpoint 4, after step 21: the kill switch stops work and can be turned back off. Checkpoint 5, after step 31: the runbook has been drilled once and repaired. Checkpoint 6, after step 38: the deployment plan has every heading and an honest statement of its limits. COMMON MISTAKES AND HOW TO FIX THEM Mistake: the run takes a long time. Fix: the retry for ticket T3 really waits half a second, so a short pause is normal, and the tests avoid it by using a fake sleeper. Mistake: the kill switch seems ignored. Fix: set the variable in the same terminal window where you run the program, and remember that the value must be exactly 1. Mistake: a ticket loops forever and your computer is slow. Fix: confirm that every worker calls charge on each pass through its loop, because a loop that never calls charge cannot be stopped by a budget. Mistake: the supervisor swallows every error. Fix: catch only the named errors, and let unexpected ones appear in the log so that you hear about new problems. Mistake: the runbook is vague. Fix: rewrite each instruction as an action a tired person could follow at three in the morning, with exact words to type.