Skip to content

Wallet Privacy Testkit 0.2.0

Latest

Choose a tag to compare

@ztsalexey ztsalexey released this 06 Sep 04:24
· 2 commits to main since this release
3ccf62c

Wallet Privacy Testkit 0.2.0 packages the recovery verifier, continuous-traffic analysis, metadata forwarding, and a reproducible two-wallet privacy study.

python -m pip install wallet-privacy-testkit==0.2.0
  • Five recovery scenarios cover lost responses, a wallet crash, indexer restart, outage, and switching indexers, with actual balance and confirmation observations and a live negative control.
  • The expanded study freezes a calibration-only detector before repeated sessions, emulated network conditions, and evaluation with the independent zcash-devtool development wallet.
  • All 45 payments in 15 study sessions confirmed. The 9,207-byte frozen cutoff missed every independent-wallet send-active window: its largest client records were three bytes below the cutoff. This is a detector transfer failure, not a wallet privacy ranking.
  • Forty tests pass on Python 3.11–3.13 across Linux and macOS. Fresh OrbStack ARM64 and native Linux Docker runs validate recovery; the exact release wheel also passes all 40 tests and independently reproduces the study.

Complete study results and retained metadata · Methodology and limits · Docker/OrbStack lab

This remains a research preview. Scheduled sends align with scoring windows; sessions share wallet/node state; the independent development wallet has a different process lifecycle; and stream conditioning is not a real WAN. The results do not establish identity inference, public-transaction linkage, or real-world wallet privacy. The source archive includes the lab and evidence; the wheel installs the CLI and library.

The wheel and source archive were audited and tested before publication. SHA256SUMS matches the checksum file committed in this tag. PyPI publication uses Trusted Publishing with attestations. Licensed MIT OR Apache-2.0.

Post-publication verification completed: both PyPI artifact metadata hashes and downloaded bytes match the tagged checksums; both publication attestations name this repository, publish.yml, and the pypi environment. A clean, uncached PyPI installation passed all 40 tests and pip check, reproduced the complete study report, and verified its recovery observations. Successful publication workflow.