This folder contains the two small public repositories I split out of the larger research workspace that I used while investigating bitsquatting and spontaneous higher-layer corruption. They are related, but they answer different questions.
This is the deterministic point-to-point UDP measurement harness I used for the Hetzner transit experiment. It sends fixed-size packets at a controlled rate, regenerates the expected packet at the receiver from the sequence number and shared key, and records corruption, duplicates, missing packets, host telemetry, and optional rolling packet captures.
Use this one if you want to reproduce:
- the long-duration sender/receiver WAN test
- local loopback control runs
- end-to-end packet integrity experiments over real networks
This is the standalone browser-side anomaly rig I used for the local browser experiment. It drives Chrome or Chromium through a local controller page and explicit proxy, generates deterministic unique URLs, and classifies the requests the browser actually makes so malformed, duplicated, stale, or otherwise unexpected browser-side behavior becomes visible.
Use this one if you want to reproduce:
- the offline local browser experiment
- browser-originated request anomaly testing
- the Chrome renderer stability issue discovered during the experiment
The two rigs were built to probe different parts of the same broader question:
point-to-point/asks whether corruption survives a real packet path and reaches user space.browser-lab/asks whether a browser on a consumer machine can itself originate corrupted or otherwise wrong requests without user error.
Taken together, they cover both a network-path experiment and a browser-side experiment.
- If you want the network/transit experiment, start with
point-to-point/README.md. - If you want the browser/local-host experiment, start with
browser-lab/README.md.
Each subfolder is meant to stand on its own as a small public repository, with its own README, configs, scripts, and source tree.