scan_masscan_nmap() and scan_geo_scout() were running nmap serially —
one host at a time with a 60s timeout, turning a 1h estimate into 24h+
for large scans.
Both modes now use ThreadPoolExecutor with 10 workers (configurable via
WEBRUNNER_NMAP_WORKERS env var). geo-scout uses an inner scan_one()
helper to batch nmap + probe_service() per host within the same future.
Also fixed probe_service() bytes format antipattern and removed dead
targets_arg variable from scan_nmap_only().
run_scan.yml converted from command: > folded scalar to command: argv:
list form to prevent argument splitting on space-containing values.
n0mad1k
4b5db31b6b