NetProbe parses untrusted input by design — packets from the wire and attacker-supplied PCAP/PCAPNG files. Bugs in that path are security bugs, and reports are very welcome.
Please report vulnerabilities privately via GitHub's security advisory form: Report a vulnerability. Do not open a public issue for anything you believe is exploitable.
NetProbe is maintained by one person. You will normally get an acknowledgement within a week; a fix timeline depends on severity, and you'll be kept informed. Please allow a fix to ship before public disclosure.
Most valuable reports, roughly in order:
- Memory-safety issues (crashes, overreads, use-after-free) reachable from a crafted capture file or crafted live traffic — the parser chain from link layer through tunnels to TLS/QUIC/DNS/HTTP is the primary attack surface.
- Issues in the QUIC Initial decryption path (Mbed TLS usage, HKDF/AES).
- Flow-export output that could inject into downstream consumers (CSV/JSON).
Out of scope: issues requiring a compromised local machine, vulnerabilities in the Npcap/libpcap drivers themselves (report those upstream), and denial of service against the GUI from pathological-but-valid traffic volumes.
Every push is built with AddressSanitizer and UndefinedBehaviorSanitizer over the full untrusted-bytes path, the protocol parsers run under a libFuzzer harness in CI, and the test suite includes adversarial malformed-packet fixtures. A finding that slips past those layers is exactly the kind of report this policy is for.
| Version | Supported |
|---|---|
| Latest release (1.2.x) | Yes |
| Older releases | No — please reproduce on the latest release |