NETRA reads a packet capture of email traffic and reports what cryptography two machines actually negotiated — TLS version, cipher, forward secrecy, key-exchange group, certificate validity, STARTTLS outcome. Every finding names the exact frame and byte offset that proves it.
It decrypts nothing, contacts no server, and runs entirely offline.
Open the tool Source on GitHubWhy passive
There are three ways to assess email cryptography. Put a component inside the mail server, scan the servers from outside, or read a capture of traffic that has already happened.
The first two are easier and both are forbidden where this matters most. An inline component sits in the path of production mail, so a bug delays real messages. A scanner opens connections to production infrastructure and needs written authorisation. In classified and air-gapped networks, neither is usually permitted. Reading a capture changes nothing, contacts nothing, and needs no credentials.
What makes it different
Observed risk says how bad what we saw was. Evidence coverage says how much we could see. They are never combined, because a single grade lets missing evidence look like safety.
Every fact is labelled observed, deduced, inferred, or not observable. TLS 1.3 encrypts the certificate, so no passive tool can see it — we say so rather than reporting no problems found.
Each finding carries a frame number, a byte offset, and a Wireshark display filter. Paste it into Wireshark and land on the same bytes.
Exports a CycloneDX 1.6 CBOM built from observed traffic. Most CBOM tooling scans source code; only traffic shows what was actually negotiated.
What it does not do
- It never decrypts anything.
- One capture is not a configuration — it reports what this client and this server negotiated, not everything the server would accept.
- It holds no state between runs. The report is the artifact, anchored to the capture's hash.
About
Built by team Grey Hats for Smart India Hackathon 2026, problem statement SIH26159. The engine has been through an architecture audit and an adversarial security review; the findings and the fixes that resolved them are held in the team's working repository and can be shown on request.