Skip to content

Week 10 - Jul 20-26, 2026

Focus: Direction 1 implementation and Draft PR review

What I did

  • Took the selected wallet-summary direction from the corrected navigation audit into a focused implementation.
  • Changed the generated-mnemonic summary to show Continue and Wallet Options. Passphrase and Customize now sit inside Wallet Options, while existing mnemonic loading keeps its current menu.
  • Added focused login tests, translations, a user-guide update, and a changelog entry.
  • Split the work into focused feature, test, translation, documentation, and screenshot commits after review made the history harder to inspect than it needed to be.
  • Added a repeatable generated-mnemonic simulator sequence and new documentation images for the summary and Wallet Options. I kept the existing loaded-mnemonic image because that flow has not changed in the PR.
  • Opened Draft PR #909 against develop. It addresses open issue #829.

Verification

  • 1,121 tests passed with 97% coverage.
  • Pylint, formatting, Vulture, translation validation, and GitHub CI passed.
  • French simulator checks covered the longest Wallet Options translation on TZT, M5StickV, and Amigo without clipping or overlap.
  • TZT hardware checks covered Continue, Back and decline recovery, Wallet Options, and the return paths from Passphrase and Customize.
  • The documentation sequence produced clean generated-summary and Wallet Options captures on Amigo, TZT, and M5StickV.

What I found

The smaller hierarchy change makes the common next step explicit without removing the useful wallet summary. Optional actions remain available, but they no longer have the same prominence as the usual path.

The review thread exposed a real wording trade-off. Continue fits the generated flow because the mnemonic already exists, while Load Wallet is familiar to existing users but can suggest a narrower wallet type. I tested a consistent loaded-mnemonic hierarchy, but I did not add it to the PR while that follow-up still needs maintainer confirmation.

Current boundary

The navigation findings still describe the audited Krux behaviour and its original evidence limits. The proposed Figma screens, simulator checks, TZT results, and Draft PR are later implementation evidence and remain labelled separately.

PR #811 also moved from design work into an implemented Draft PR this week. It keeps the 5x6 keypad size and returns scanned text to the editable buffer before the user presses Go. Its four implementation commits credit my layout, interaction, and glyph contribution. This remains separate from the three audit-derived directions.

Next step

  • Respond to review on Draft PR #909 without expanding its scope.
  • Wait for maintainer confirmation before changing the loaded-mnemonic flow.
  • Review the implemented QR-keypad flow against the agreed device and recovery behaviour.
  • Keep the other two audit-derived directions as possible follow-ups rather than parallel implementation work.
  • Finish the evidence summary and final audit write-up.