These scripts stress this host only (and localhost networking). Do not point them at other machines.
Terminal A — collect metrics:
python -m sysspectogram collect --out data/anomaly.csv --duration 1200Terminal B — pick a generator:
python simulations/cpu_miner.py --duration 900
python simulations/net_flood_local.py --duration 900 --rate 100
python simulations/mem_pressure.py --duration 300 --mb 512
python simulations/disk_thrash.py --duration 300Or:
chmod +x simulations/run_all.sh
./simulations/run_all.shCollect a separate data/normal.csv during normal use for Class 0.
Equivalent CLI entrypoints:
python -m sysspectogram simulate cpu|mem|disk|net ...| Flag | Default | Description |
|---|---|---|
--duration |
900 | Seconds of CPU burn |
--workers |
auto (~half CPUs) | Worker processes |
| Flag | Default | Description |
|---|---|---|
--duration |
300 | Seconds |
--mb |
512 | Megabytes to allocate and touch |
| Flag | Default | Description |
|---|---|---|
--duration |
300 | Seconds |
--block-mb |
32 | Write/read block size |
--path |
/tmp/sysspectogram_disk_thrash.bin |
Temporary file path |
| Flag | Default | Description |
|---|---|---|
--duration |
900 | Seconds |
--rate |
80 | TCP connections per second to a local listener on 127.0.0.1 |
Environment overrides: DURATION_CPU, DURATION_NET, DURATION_MEM, DURATION_DISK (seconds).
- CPU / memory / disk: local resource burn only
- Network: bind and flood
127.0.0.1only - No remote scanning tools
Perimeter scenarios. No attacks on third-party hosts. External TEST-NET beacon requires --lab.
python simulations/intrusion/ssh_bruteforce_local.py --count 12 --out /tmp/fake_auth.log
python simulations/intrusion/port_scan_local.py --count 40
python simulations/intrusion/egress_beacon.py --local-sink --duration 8
python simulations/intrusion/lab_runner.py --lab egress_beaconRun alongside python -m sysspectogram watch-perimeter or guard.