Whenever you're ready, walk me through how you'd approach this.
How would you build a system to continuously scan Google repositories for known vulnerabilities? Design a practical pipeline covering repository change detection, dependency and manifest extraction, vulnerability matching, incremental rescans, durable findings storage, and alert delivery. Explain how you would handle idempotency, false positives, newly disclosed vulnerabilities, scan failures, access control, monitoring, and backfills using Google Cloud services.