Est. 2013 — Stratosphere Laboratory, Czech Technical University in Prague

Slips:
behavioral detection

Slips is the first free software behavioral machine learning-based IDS/IPS for endpoints. It reads input traffic, accumulates evidence until it triggers an alert. Can ingest live traffic, pcaps, Zeek, Suricata, Argus, and more.

How Slips decides

01

Models

Machine learning trained on malicious behavior over time windows. It catches targeted attacks and command-and-control channels that match nothing known.

02

Intel

40+ threat intelligence feeds, updated continuously. Modules can also pull context from VirusTotal and RiskIQ.

03

Heuristics

Expert rules where models fall short, including adaptive TLS/HTTPS anomaly detection with drift handling.

Evidence → Extensively trained thresholds decide when enough has accumulated. Then Slips alerts, notifies the desktop, and can block on Linux.

Quick start

One command.

Docker is the recommended way to run Slips. Needs Python 3.10.12 and 4 GB of RAM. Blocking features are Linux-only.

Pull and run
$ docker run --rm -d --net=host --cap-add=NET_ADMIN --name slips stratosphereips/slips:latest bash -c 'redis-server --daemonize yes && exec ./slips.py -i <interface> -w -o output'

Built to be extended

Modularity

One module, one detection

Python end to end, with a documented template for adding your own detection to the pipeline.

P2P

Peers that share indicators

Find other peers on the network and trade IoCs in a balanced, trusted manner. Optional.

Federated

Learning without pooling

Federated training across endpoints through the feel_project submodule.

Interfaces

Browser or terminal

Run with -w for the web interface on port 55000, or use Kalipso in the terminal.

Config

Yours to tune

Time window width, analysis direction, train or test the models, plug in your own Zeek script.

Contribute

Issues, PRs, GSoC

Report bugs, request features, or apply through Google Summer of Code — we host as an organisation.

On stage

Bring a
detection idea.

Slips is GPL-2.0 free software. Read the contributing guidelines, then open an issue, a pull request, or a whole new module.

Founder Sebastian Garcia
Main authors Sebastian Garcia, Alya Gomaa
+ Kamila Babayeva, Veronica Valeros, Frantisek Strasak, Dita Hollmannova, Ondřej Lukáš, Elaheh Biglar Beigi, Martin Řepa, arkamar, Maria Rigaki, Lukas Forst, Daniel Yang, Jan Sbovoda, David Otta