SeedPass/memory-update.md

Memory Update (2026-03-02)

TUI v2 mockup parity audit refreshed

• Added a dedicated audit doc: docs/tui_v2_mockup_gap_audit_2026-03-02.md.
• Linked the audit from docs/tui_v2_ui_refresh_plan.md so future sessions use it as the active parity reference.

High-impact findings (user-visible)

• Inspector panel currently pre-populates because first entry is auto-selected; this diverges from context-aware mockup behavior.
• Left nav is not yet a true hierarchical tree of fingerprint -> managed-account nodes.
• Layout density and panel proportions still differ from mockup geometry, especially on large displays.
• Reveal/QR pathways exist in code but user runtime still reports non-working v/QR behavior in some installs; treat as high-priority runtime validation gap.

Evidence artifacts to reuse

• Primary runtime evidence: artifacts/ui_eval/audit_20260302/state_dump.txt.
• Mockup baseline set: UI_mockups/PNG/*.png (all 1854x1080).

Recommended next implementation slice

1. Remove initial auto-selection and use a neutral inspector state.
2. Implement true left-side hierarchy model.
3. Rework dense header/footer strips to mockup rhythm.
4. Tune spacing/label verbosity for higher effective density before adding new feature surface.

Plan consolidation (2026-03-02)

• Created docs/tui_v2_integration_execution_plan_2026-03-02.md as the consolidated implementation roadmap.
• Captured user-priority requirements: context-aware inspector lifecycle, true fingerprint/managed/agent tree, collapsible sidebar, filter menu replacing always-visible kind toggles, and unified lexical+tag+semantic search.
• Reordered near-term execution toward structure and reliability first (Phases A-C, D, G, H), then semantic integration and final board-fidelity polish.

Session Update (2026-03-02, evening)

TUI v2 parity slices completed

• Added per-profile sidebar collapse persistence in src/seedpass/tui_v2/app.py (state now restores on profile-switch and profile-tree-open).
• Added search mode chips to grid heading and keyboard cycling (m) for keyword -> hybrid -> semantic.
• Improved reveal UX for confirm-required secrets (seed, managed_account, ssh, pgp): pressing v now arms confirm and pressing v again (within 8s) confirms reveal.
• Fixed profile tree label rendering glitch (fp-aSeed:...) by normalizing to | Seed: ... format.

Regression coverage added

• Sidebar state restore per profile.
• Search-mode chip rendering + m hotkey cycle behavior.
• Managed-account reveal double-confirm keyboard flow.
• Profile-tree seed label formatting check.

Verification status

• Focused TUI v2 gate is green at 45 passed for:
  • src/tests/test_tui_v2_textual_interactions.py
  • src/tests/test_tui_v2_action_matrix.py
• New UI artifacts captured:
  • artifacts/ui_eval/audit_20260302_phaseB2/
  • artifacts/ui_eval/audit_20260302_phaseE3/
  • artifacts/ui_eval/audit_20260302_phaseG2/
  • artifacts/ui_eval/audit_20260302_phaseB3/

Incremental parity progress (2026-03-02 late)

• Left profile tree now supports child-node traversal and open actions for managed/agent rows via ↑/↓/Ctrl+O.
• Inspector behavior is more context-aware: non-sensitive kinds keep secret panel hidden by default and report clean unsupported-action status for v/g.
• Tree rendering now includes explicit group headers (Managed Users, Agents) to match mockup hierarchy clarity.
• Focused TUI gate remains stable at 46 passed after these slices.
• Added CLI regression tests for legacy launch compatibility (seedpass legacy, seedpass --legacy-tui) in src/tests/test_typer_cli.py to guard installer/runtime fallback behavior.
• Added responsive high-resolution density mode for large viewports (notably 2256x1504 class), reducing chrome heights for brand/ribbon/status/action bars and heading rows to fit more working content.
• Tightened default vertical whitespace by removing extra top margin on status and TOTP board panels.
• Focused TUI gate after density updates: 47 passing tests.
• Added regression check for unsupported QR behavior on non-sensitive kinds (document) so g returns an explicit status message instead of ambiguous UI behavior.
• Added a unified right-pane mode switch path so document editor and 2FA board transitions remain mutually exclusive and state flags stay synchronized.
• Added regression test covering 2FA board -> document editor -> view transitions.
• Focused TUI gate now at 48 passing tests.
• Added context guard so opening an entry auto-closes 2FA board and blocks entry switching while document editor is active (prevents unsaved context drift).
• Added transition regression test for: 2FA board -> open entry (board closes) -> edit doc -> open other entry (blocked).
• Focused TUI gate now at 49 passing tests.
• Added dense high-res action strip mode with abbreviated global commands to reduce clipping and keep bottom controls readable on large viewports.
• Added regression assertions for dense action-strip activation/deactivation tied to responsive layout thresholds.
• Inspector heading is now context-aware and tracks right-pane mode: selected entry view, document editor, and 2FA board states.
• Added regression coverage for heading transitions across view/edit/2FA modes.
• Focused TUI gate now at 50 passing tests.
• Refined inspector board typography for password/document/TOTP kinds by grouping fields and operations into clearer compact sections (Login Fields, Document Fields, 2FA Fields) to reduce visual clutter.
• Captured visual/state evidence in artifacts/ui_eval/audit_20260302_phaseF6/.
• Bottom action strip shortcuts are now functional and test-covered via new shift-shortcut actions mapped to palette command prefixes plus reveal action.
• Added focused artifact: artifacts/ui_eval/audit_20260302_phaseH6/ for bottom action-strip validation.
• Added clickable action-strip routing: clicking top-row shortcut segments now triggers the same actions as keyboard shortcuts.
• Added regression test for action-strip click routing and captured artifact artifacts/ui_eval/audit_20260302_phaseH7/.
• Focused TUI gate now at 53 passing tests.
• Refined dense-mode top-left sidebar hierarchy: compact FP/profile/session scope lines plus reduced label noise while preserving tree navigation and status cues.
• Added artifact artifacts/ui_eval/audit_20260302_phaseH8/ documenting the updated hierarchy.
• Continued board-rhythm parity pass: aligned TOTP field labels (Period :, Digits :) to match password/document row style and improve scan consistency.
• Added artifacts audit_20260302_phaseF7 and audit_20260302_phaseF8 for board layout comparison against mockups.
• Extended dense-mode action verbosity reduction across additional boards (seed/managed, ssh, pgp, nostr) for consistent low-noise inspector language.
• Improved profile-tree depth readability with active-branch counts (M:x A:y) and clearer child group connectors/labels.
• Added verification artifact artifacts/ui_eval/audit_20260302_phaseB6/ for tree hierarchy scan quality.

Session Update (2026-03-03)

TUI v2 managed-session + action-strip reliability fixes

• Added managed-session breadcrumb tracking in src/seedpass/tui_v2/app.py so left header now shows nested path context:
  • Path: <root-fingerprint> > <managed-fingerprint> > ...
• Managed session state now tracks nested loads (managed-load) with stack semantics and restores the previous layer on each managed-exit.
• managed-load now normalizes kind tokens (managed-account, managed account, managed_account) to reduce false negatives.
• Action strip row-2 is now clickable (context actions), including:
  • v reveal
  • g QR
  • e edit
  • a archive
  • 6 2FA board
  • managed-load / managed-exit when managed account is selected
• Added robust kind normalization for context rendering so action hints stay aligned with selected entry kind.

Regression tests added/updated

• test_tui2_textual_action_strip_click_routes_to_shortcuts now validates row-2 click reveal.
• Added test_tui2_textual_action_strip_context_updates_for_selected_kind to ensure bottom context row updates by selection.
• Extended test_tui2_textual_managed_account_session_palette_commands with nested managed-load + breadcrumb assertions.

Verification

• Focused and gate suite green:
  • poetry run pytest -q src/tests/test_tui_v2_textual_interactions.py src/tests/test_tui_v2_action_matrix.py
  • Result: 56 passed.

Session Update (2026-03-03, Slice 2)

UI-board density and hierarchy polish

• Dense high-resolution mode now keeps #grid-heading at 4 rows instead of 3 to avoid header/divider clipping in large viewports.
• Refined dense top-left header hierarchy to reduce line noise:
  • retained explicit Fingerprint and Path
  • consolidated session/profile/managed state into one line
  • kept scope line focused on active filters only

Regression updates

• Updated test_tui2_textual_hires_density_compacts_vertical_chrome to assert the new dense grid-heading height (4).

Verification

• Focused TUI v2 gate remains green:
  • poetry run pytest -q src/tests/test_tui_v2_textual_interactions.py src/tests/test_tui_v2_action_matrix.py
  • Result: 56 passed.

Session Update (2026-03-03, Slice 3)

Board geometry parity pass

• Added reusable ASCII card renderer for inspector sections (_board_card).
• Password and stored-password boards now render framed Credentials and Quick Actions blocks.
• Note/document boards now render framed Content and Quick Actions blocks.
• Increased note/document preview length (120 -> 180 chars) for better mockup-like content emphasis.

Regression coverage

• Extended note metadata board test to assert card framing.
• Added password board card-section test.

Verification

• Focused TUI v2 gate:
  • poetry run pytest -q src/tests/test_tui_v2_textual_interactions.py src/tests/test_tui_v2_action_matrix.py
  • Result: 57 passed.

Session Update (2026-03-03, Slice 4)

UI-board density rebalancing

• Added dynamic dense/high-res vertical split:
  • idle: top-work=9fr, right=3fr
  • active inspect/edit/2FA: top-work=8fr, right=4fr
• Added _refresh_layout_balance() and invoked it during key state transitions so layout rebalance is state-driven, not resize-driven.

Regression coverage

• Added test_tui2_textual_hires_density_rebalances_idle_vs_selected.

Verification

• Focused TUI v2 gate:
  • poetry run pytest -q src/tests/test_tui_v2_textual_interactions.py src/tests/test_tui_v2_action_matrix.py
  • Result: 58 passed.

Session Update (2026-03-03, Slice 5)

Action-strip microcopy polish

• Standardized context action wording to verb-first with key hints across entry kinds.
• Preserved managed account session actions while aligning adjacent wording style.
• Expanded row-2 click parser to recognize verb-led labels so copy changes do not break mouse interactions.

Regression coverage

• Updated action-strip context tests to match polished copy.

Verification

• Focused TUI v2 gate:
  • poetry run pytest -q src/tests/test_tui_v2_textual_interactions.py src/tests/test_tui_v2_action_matrix.py
  • Result: 58 passed.

Session Update (2026-03-03, Slice 6)

Universal card framing

• Extended _board_card to TOTP (Parameters), SSH/PGP/Nostr (Keys), and Seed/Managed (Seed Info) boards.
• All 9 board types now use consistent card-section visual language.
• Updated regression tests for SSH/PGP/Nostr, Seed/Managed, and TOTP boards.

Verification

• Focused TUI v2 gate: 58 passed.

Next

• Capture fresh state dump/screenshots for updated boards.
• Update strict closeout scoreboard (Password/Stored Password/Note now at card parity; remaining Minor Gap items are density/rhythm only).

Session Update (2026-03-03, Slice 7)

Board density tightening

• Removed redundant section headers above card frames (saves 2-3 lines per board).
• Removed "Compact: Notes/Tags shown inline." noise from compact mode.
• Capped note preview at 100 chars (was 180) to prevent oversized Content cards.
• Password board now 17 lines (was ~21). All boards similarly tighter.

Verification

• Focused TUI v2 gate: 58 passed.

Session Update (2026-03-03, Slice 8)

Action strip + hint line tightening

• Normal-mode global row now uses verb-first labels matching context row style.
• Seed/TOTP Actions hints shortened from raw palette commands to user-facing verbs.
• All action hint lines now consistently use human-readable language.

Verification

• Focused TUI v2 gate: 58 passed.

Session Update (2026-03-03, Slice 9)

Final micro-polish

• _board_card now has max_width=72 cap with ellipsis truncation for long rows.
• Nostr action hint capitalized: QR Private (confirm).
• Compact mode global strip verb-first: Settings (S) Add (A) ....
• All three modes (compact/dense/normal) now share consistent verb-first label style.

Verification

• Focused TUI v2 gate: 58 passed.