Claude Operator: Prompt to Autonomy · 60 min · 15 XP
Capstone: autonomous operations loop
One test-only workflow carrying every control this level taught.
Build one complete workflow against test accounts and test data only. It gathers bounded inputs, proposes work, executes only the low-risk actions you approved, verifies each side effect, logs what it did, and stops on budget or policy limits.
Every control from this level should be present and demonstrable: a dry-run mode, structured output, turn and spend ceilings, idempotency, checkpoints and resumption, a retry table with backoff, read-back verification, an approval queue, a kill switch, redacted logs, the six metrics, and four alerts with owners.
Do it now, in your own tools. Demonstrate it end to end and keep the evidence: the dry-run report, the approval queue and what you approved, the execution log, the verification records, the metrics, and a kill-switch drill you actually performed. Finish with a short retrospective — which control you'd add first if this handled something that mattered, and which one you were tempted to skip. That temptation is the most useful thing you'll learn here.
Claude Code CLI reference · https://code.claude.com/docs/en/cli-reference
Loading your workspace…