Offline. Deterministic.
By architecture.

SNF (Shadow Network Fingerprinting Engine) is a 100% offline, air-gap-native passive network intelligence platform written entirely in Rust. It captures raw packets, reconstructs TCP/UDP flows, and runs them through 14 deterministic protocol analyzers - producing structured NDJSON output, court-admissible evidence bundles, and forensic baseline reports without ever making a network call.

The core guarantee: F(dataset, config, version) = identical NDJSON output every run. Same PCAP, same config, same SNF version - SHA-256 identical output on any machine, any analyst, any time. This is the determinism contract, and every architectural decision in SNF is subordinate to it.

Zero network calls. Ever. Not as a setting, not as a mode - as an architectural constraint. SNF does not phone home, does not require internet access, and does not function differently in an air-gapped room than it does connected to the internet. Because it was designed that way from day one.
~50K
Lines of Rust
200+
Source Files
24
Phases Complete
98.8%
Test Coverage
1.25G
bps Single-Core
14.9M
Packets - Zero Crashes

What existing tools
cannot do.

The tools that dominate network security today - Zeek, Suricata, Darktrace, Vectra, CrowdStrike - were built for connected environments. They assume cloud access for updates, telemetry, and analysis. In a classified network, a nuclear facility, a defense air-gap, or an OT environment, they are either illegal to deploy or architecturally incapable of operating.

Beyond connectivity, they share a deeper problem: non-determinism. Wireshark is analyst-dependent. Zeek output varies by configuration drift. No tool in mainstream use can prove that two analysts running the same PCAP will get the same result - which makes their output inadmissible as court evidence without extensive manual verification.

How it's built.

SNF uses a two-binary design: snf_core (the capture and analysis engine) and snf_report (the reporting and output engine). The analysis path runs 14 protocol analyzers in a locked execution order - DNS, TLS, QUIC, HTTP/1.1, HTTP/2, DHCP, ICMP, SMB, and five ICS/SCADA protocols among them - producing deterministic events regardless of threading model or platform.

Four operation modes cover the full deployment surface: Forensic (multi-threaded, full output, DFIR post-mortem), Monitor (live 24/7 SOC sensor), Replay (single-threaded, deterministic, court evidence), and Stealth (silent NDJSON, zero console output, covert sensor). A fifth mode, Redact, produces GDPR/HIPAA-safe PCAP with consistent IP/MAC anonymization and a full audit manifest - enabling PCAP sharing without legal exposure.

The security constraint system formalized 28 non-negotiable rules enforced at the code level: bounded allocations, capped loops, zero unwrap/expect in production paths, pre-allocated flow tables, atomic file writes, path traversal protection, and self-learning gated by mode so Replay and Stealth produce zero disk side effects. These are auditable, not aspirational.

What SNF is not.

SNF occupies a specific and narrow space. Understanding what it is not is as important as understanding what it is.

Who builds it.

SNF is developed and maintained by SNF Labs, an independent division of RSAAT Labs. SNF Labs carries its own brand identity, targets a distinct audience - security researchers, incident responders, DFIR teams, and critical infrastructure operators - and operates independently in terms of product, go-to-market, and technical roadmap.

The parent organization, RSAAT Labs, is an independent software lab building tools across developer productivity, consumer SaaS, and deep security engineering. Both entities share the same founding philosophy: software should do what it says it does, with precision and no overhead.

Parent Organization
RSAAT Labs

An independent software lab building developer tools, passive network security engines, and consumer SaaS products. No bloat, no dark patterns, no nonsense. Every product that ships is built to be genuinely useful, technically sound, and respectful of the people using it. Privacy is architecture, not policy.

rsaatlabs.com

Built by one person.

SNF is a single-author project - designed, architected, and implemented from the ground up. ~50,000 lines of Rust, 24 implementation phases, 200+ source files, validated against 18 real-world PCAPs including MAWI backbone traffic at 14.9 million packets. Not a team project. Not a prototype. A production-grade forensics engine built with the kind of focus that distributed teams rarely achieve.

Tejas Padigel
Founder, SNF Labs & RSAAT Labs

Designer and developer of SNF - Shadow Network Fingerprinting Engine. B.Tech CSE (Cybersecurity). Builder of tools that work in the real world, under constraints, without compromise.