[luannn010]

Summary

• Add pinned engine WASM, account entry, encrypted local storage and Markdown transfer.
• Foundation of the beta stack; merge before the UI and integration PRs. No public beta or deployment.
• Risk: high; account trust and local keys. Undo: revert the merge before release; preserve browser data.
• First-device recovery trusts the account server; established-device approval verifies the engine ceremony.

Testing

• From web/: npm ci, npm run check:data, npm run test:data, npm run verify:engine.
• Actual-engine OPAQUE/TLS-mailbox account test passes (1/1); invalid/expired/used/concurrent invite checks pass.
• Not tested: production delivery or full second-factor/paired-device HTTP ceremonies.

Co-authors

none

AI usage

• Written by hand, with no AI
• Harness and version: Codex CLI 0.160.0 through Codex desktop; PowerShell, Node/npm, Git, Vitest, Chromium, WSL Rust/PostgreSQL; no Claude.
• Models: gpt-6.1-sol (coordination, implementation, separate review).
• Effort: medium.
• Human role: approved beta scope, parallel ownership, isolated test dependencies and PR preparation; retains merge/deployment decisions.
• Sources: orbyn-engine and unchanged generated artifacts, MIT OR Apache-2.0 with retained notices. No Planner application source copied. Dependencies and versions are in the lockfile.

Checklist

• Works on Windows, macOS and Linux (Windows/WSL checked; hosted Linux pending)
• Generated engine contracts and pin recorded
• No secrets, credentials, personal data or AI attribution in commits/title/branch