During internal penetration tests, one of the things often checked is which outbound ports a network allows, to see how data could be exfiltrated. The usual approach is to scan a server with every port open, like letmeoutofyour.net or portquiz.net, and see which ports come back open.
The problem I kept running into was that the results weren't always accurate. Proxies and other devices in between would sometimes respond themselves, making a port look open when it wasn't. Other times the traffic did reach the destination, but the scanner still reported the port as closed.
So I built portleak.link. It doesn't respond to any TCP SYNs or UDP packets, so your scanner will show everything as closed. Instead, the web interface logs all traffic that actually hits the server. If your traffic appears in the log, it definitely made it out.
To use it, enter your public IP in the filter, run your scan, and the results show up in the table. You can export them to CSV or XLSX. There's also a scans page with ready-made scripts for Windows and Linux.
Xtrato•56m ago
The problem I kept running into was that the results weren't always accurate. Proxies and other devices in between would sometimes respond themselves, making a port look open when it wasn't. Other times the traffic did reach the destination, but the scanner still reported the port as closed.
So I built portleak.link. It doesn't respond to any TCP SYNs or UDP packets, so your scanner will show everything as closed. Instead, the web interface logs all traffic that actually hits the server. If your traffic appears in the log, it definitely made it out.
To use it, enter your public IP in the filter, run your scan, and the results show up in the table. You can export them to CSV or XLSX. There's also a scans page with ready-made scripts for Windows and Linux.
I wrote up a bit more background here: https://ja.meswoolley.co.uk/portleak-link/
It's free to use. I'd be interested to hear any feedback or ideas for improvements.