sebaandrada-cell published PcapAnalyzer with 27 detectors for .pcap and .pcapng files. Security coverage includes port scans, cleartext logins, Telnet, SMBv1 and ARP spoofing. An attacker with passive network access can exploit the cleartext flaws it flags.
The pitch is seconds, not hours.
Before this, analysts ran Wireshark filters and tshark commands by hand. PcapAnalyzer bundles that triage into a self-hosted app under Apache 2.0, sebaandrada-cell says. Per GitHub, the project has 67 stars and 117 commits as of . Setup runs through Docker Compose and start.sh with a one-time token on 127.0.0.1 only.
TCP, path, DNS, application, tunnel and capture-quality checks run on every upload. Findings include SYN to RST blocks, retransmissions, PMTUD black holes, slow DNS, HTTP 5xx errors and one-way WireGuard or IPsec handshakes. AI investigation is optional and uses your own Anthropic key, with cited packets and follow-up questions. Without a key, demo mode runs the full flow at no cost.
PacketSnitch lists protocol decoding, GeoIP and threat intel plus AI on its site, a commercial desktop contrast to this local-first approach. PcapAnalyzer keeps captures encrypted and deletes them after 60 minutes of inactivity or on manual delete. It never stores or shows credentials from cleartext logins. The app warns when results are limited by mirror-port duplicates, one-direction traffic or truncated packets.
Testing so far used synthetic fixtures and public samples, sebaandrada-cell says. The repo warns against exposing the API without an authenticated reverse proxy. Beta feedback now runs through GitHub Issues for missed faults or false positives.





