mirror of
https://github.com/yogeshojha/rengine.git
synced 2026-10-20 23:31:31 +02:00
Four stages write ip_addresses for one scan and each reported its own "ips" count, so a single-address scan of 192.168.0.88 rolled up as two addresses. subdomains had the same shape across subdomain_discovery, vhost and asset_seed. In the other direction origin_probe writes http_assets but reports no count key, so that column under-reported every origin it found. open_ports and endpoints were already measured from their tables for exactly this reason. Every headline is now measured the same way, so a figure on the scan equals the rows it opens. The per-stage summing layer is gone rather than left in place: it was unconditionally overwritten by the derived values, so editing COUNT_TO_COLUMN to add a metric would have looked correct and done nothing. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01K6LcH8uUddjJbKrahf34ys