# Changelog

All notable changes to this project will be documented in this file.

The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),
and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).

## [Unreleased]

## [0.5.12] - 2026-07-27

### Dependencies

- `serde`: 1.0.228 → 1.0.229 (#242)
- `regex`: 1.12.4 → 1.13.1 (#241, #249)
- `toml`: 1.1.2 → 1.1.3 (#247)
- `tokio`: 1.52.3 → 1.53.1 (#244, #252)
- `anyhow`: 1.0.103 → 1.0.104 (#245)
- `rust-i18n`: 4.1.0 → 4.2.1 (#248)
- `futures`: 0.3.32 → 0.3.33 (#243)
- `thiserror`: 2.0.18 → 2.0.19 (#246)
- `serde_json`: 1.0.150 → 1.0.151 (#253)

### CI

- Bump `lewagon/wait-on-check-action` from 1.8.1 to 1.9.0 (#250)
- Bump `actions/labeler` from 6 to 7 (#251)

## [0.5.11] - 2026-07-07

### Dependencies

- `tokio`: 1.52.1 → 1.52.3 (#220)
- `serde_json`: 1.0.149 → 1.0.150 (#221)
- `rust-i18n`: 4.0.0 → 4.1.0 (#222)
- `fsrs`: 5.2.0 → 6.6.1 (#223, #225, #230)
- `ratatui`: 0.30.0 → 0.30.2 (#226, #234)
- `chrono`: 0.4.44 → 0.4.45 (#227)
- `tui-big-text`: 0.8.4 → 0.8.8 (#229, #233)
- `regex`: 1.12.3 → 1.12.4 (#231)
- `anyhow`: 1.0.102 → 1.0.103 (#236)
- `rand`: 0.10.1 → 0.10.2 (#238)
- `crossbeam-epoch`: 0.9.18 → 0.9.20 (transitive, fixes RUSTSEC-2026-0204) (#238)

### CI

- Bump `codecov/codecov-action` from 6 to 7 (#224)
- Bump `lewagon/wait-on-check-action` from 1.7.0 to 1.8.1 (#228, #237)
- Bump `actions/checkout` from 6 to 7 (#232)
- Bump `actions/cache` from 5 to 6 (#235)

## [0.5.10] - 2026-04-21

### Fixed

- `Excellent` performance tier (80-99%) is now reachable with all existing scenarios by lowering the threshold from 90% to 80% (#201)
- PerformanceTracker (FSRS) data now persisted between sessions (#196)
- `commands_executed` counter now incremented on scenario completion (#196)
- Daily quests now generated on first launch (#195)
- Escape key no longer counts as a game action when dismissing the hint popup (#194)
- Arcade sub-mode headers now display the correct title: Survival shows `SURVIVAL MODE`, Daily Challenge shows `DAILY CHALLENGE` (#197)
- Collapse nested if into match guard for clippy compliance in gamification module

### Dependencies

- `rust-i18n`: 3.1.5 → 4.0.0 (#217)
- `tokio`: 1.49.0 → 1.52.x (#187, #209, #216)
- `rand`: 0.10.0 → 0.10.1 (#215)
- `toml`: 1.0.3 → 1.1.2 (#188, #193, #207, #208)
- `tui-big-text`: 0.8.2 → 0.8.4 (#205, #210)
- `proptest`: 1.10.0 → 1.11.0 (#206)
- `tempfile`: 3.26.0 → 3.27.0 (#189)
- `tracing-subscriber`: 0.3.22 → 0.3.23 (#190)
- `rodio`: 0.22.1 → 0.22.2 (#186)

### CI

- Bump `codecov/codecov-action` from 5 to 6 (#203)
- Bump `lewagon/wait-on-check-action` from 1.5.0 to 1.7.0 (#204, #212, #218)
- Bump `softprops/action-gh-release` from 2 to 3 (#211)
- Bump `actions/github-script` from 8 to 9 (#214)
- Bump `dependabot/fetch-metadata` from 2 to 3 (#213)

## [0.5.9] - 2026-03-07

### Fixed

- Fix double keypress on Windows by filtering `KeyEventKind::Release` events (#182)

## [0.5.8] - 2026-02-09

### Added

- **Arrow key cursor movement** — Optional user configuration to enable arrow keys for cursor movement (#153)

### Changed

- CI: Added dependabot auto-merge workflow for patch-level dependency updates
- CI: Fixed codecov upload to not fail workflow on upload errors
- CI: Fixed labeler workflow to use `pull_request_target` event

### Dependencies

- `rand`: 0.9.2 → 0.10.0 (adapted to `RngExt` trait rename)
- `anyhow`: 1.0.100 → 1.0.101
- `proptest`: 1.9.0 → 1.10.0
- `regex`: 1.12.2 → 1.12.3
- `criterion`: 0.8.1 → 0.8.2
- `tui-big-text`: 0.8.1 → 0.8.2
- `chrono`: 0.4.42 → 0.4.43
- `thiserror`: 2.0.17 → 2.0.18
- `time`: bumped to latest compatible
- `bytes`: bumped to latest compatible

### Quality

- **Tests**: 1857 (up from 1842)
- **Clippy**: Zero warnings

## [0.5.7] - 2026-01-15

### Added

- **Category Filters Screen** — Filter scenarios by category with popup UI (#143):
  - Quick category selection with `f` key from scenario list
  - 9 categories: Movement, Editing, Selection, Clipboard, Search, Text Objects, Surround, Multi-cursor, Insert
  - Toggle categories with Space/Enter, navigate with j/k

- **Multi-cursor Support** — Full support for Helix multi-cursor operations (#142):
  - `C` (copy_selection_on_next_line) and `Alt-C` (copy_selection_on_prev_line)
  - Multi-cursor scenarios for learning cursor multiplication

- **Bracket Preview Highlight** — Visual preview for surround commands (#136):
  - Shows bracket type before confirming surround operation
  - Supports all bracket pairs: `()`, `[]`, `{}`, `<>`, quotes

- **Scenario Progress Tracking** — Display current scenario number in title (#135)

### Fixed

- **Search Command `*` Behavior** — Now correctly matches real Helix behavior (#150):
  - `*` selects current word and sets search pattern without jumping
  - Use `n`/`N` to navigate between matches after `*`

- **Search `N` Command** — Fixed `find_prev` returning same match at cursor end (#148):
  - Proper backward navigation with correct match boundary handling

- **Removed Invalid `#` Command** — Removed Vim-ism that doesn't exist in Helix (#146):
  - `#` (search_word_backward) is not a Helix command
  - Use `*` followed by `N` for backward word search

- **Alt Modifier Key Support** — Proper handling of Alt-key combinations (#140)

### Quality

- **Tests**: 1842 (up from 1626)
- **Clippy**: Zero warnings

## [0.5.6] - 2026-01-10

### Added

- **Embedded Scenarios & Quests** — All 136 scenarios and 55 quests now compiled into binary:
  - Zero runtime file I/O for scenarios/quests
  - Instant startup with no file loading
  - Single-binary distribution (no assets folder needed)
  - `include_str!` macro for compile-time embedding

- **Menu Position Persistence** — Scenario list remembers position when returning from training:
  - Selected item and scroll offset preserved
  - Seamless navigation between training and menu
  - Session-scoped memory (resets on restart)

- **GitHub Auto Labeler** — Automatic PR labeling in CI:
  - File-path based labels (ui, helix-core, game, etc.)
  - Conventional commit prefix labels (feat→enhancement, fix→bug, etc.)

### Changed

- **README** — Added training mode GIF demo in hero section

### Quality

- **Tests**: 1626 (up from 1355)
- **Coverage**: 91% (up from 86%)
- **Clippy**: Zero warnings

## [0.5.5] - 2026-01-10

### Added

- **Audio Feedback System** — Sound effects for mini-game events:
  - Complete/Failed scenario sounds
  - Multiplier up, Level up, Life lost sounds
  - Game over and countdown sounds
  - Timer warning alerts
  - OGG Vorbis format via rodio library (CC0 licensed from Kenney.nl)

- **Game Mode Selection** — Three distinct mini-game modes:
  - **Arcade**: Classic 60-second timed session with 3 lives
  - **Survival**: Endless mode with 1 life, escalating time pressure (10s → 3s per level)
  - **Daily Challenge**: Fixed 10 scenarios, 3 attempts per day, same for all players

- **Sound Toggle** — Global sound on/off with `M` key across all screens

- **Mode Selection Submenu** — Visual popup for choosing game mode when starting Arcade

### Changed

- CI: Added `libasound2-dev` installation for Linux audio support

### Quality

- **Tests**: 1355 (up from 1347)
- **Clippy**: Zero warnings

## [0.5.4] - 2026-01-10

### Added

- **FSRS-based Arcade Mode** — Intelligent scenario selection prioritizes commands needing practice:
  - Overdue commands increase scenario urgency
  - Weak commands (low success rate) appear more often
  - Novel commands get moderate priority for discovery
  - Weighted random selection maintains variety

- **Statistics Screen Improvements** — Real FSRS data integration:
  - Command Mastery Distribution with progress bars
  - Review Status showing due reviews and success rates
  - Commands Needing Practice section (top 5 weak commands)
  - Scenario Mastery distribution (Learning/Proficient/Mastered)

- **Profile Screen Learning Progress** — New section showing FSRS learning metrics

- **Review Session UX** — Improved review experience:
  - Command descriptions from registry
  - Session complete notifications with XP earned
  - "No reviews due" informative message

- **Mastery Level-Up Notifications** — Visual feedback when command mastery improves

- **Navigation Shortcuts** — `m` and `Esc` keys navigate back to mode selection from scenario list

- **Review Shortcut** — `r` key starts review session from all main screens

### Changed

- **Codebase Architecture** — Comprehensive refactoring:
  - Split `input/typestate.rs` (2747 → ~400 lines max per file)
  - Centralized test utilities in `src/testing/` module
  - Split simulator tests into 8 organized modules
  - Removed 21 dead code items

### Quality

- **Tests**: 1268 (up from 1239)
- **Clippy**: Zero warnings
- **Coverage**: 86%

## [0.5.3] - 2026-01-09

### Added

- **Syntax highlighting for scenario code** — Rust code in scenarios now displays with full syntax highlighting using syntect
- **Adaptive difficulty system** — Mini-game difficulty adjusts based on player performance metrics

### Changed

- **136 scenarios with realistic Rust code** — All scenarios updated from minimal synthetic text to real, idiomatic Rust code snippets
- **Paragraph commands use Helix keybindings** — Changed from Vim-style `{`/`}` to Helix-style `[p`/`]p` paragraph navigation

### Fixed

- **Paragraph scenario selection visibility** — Target states now correctly show selection range for `]p`/`[p` commands
- **Cursor rendering at end-of-line** — Cursor now displays correctly when positioned at the end of a line

### Quality

- **Tests**: 1239+ (up from 1116)
- **Clippy**: Zero warnings
- **Scenarios**: 136 total (realistic Rust code content)

## [0.5.2] - 2026-01-07

### Added

- **141 new training scenarios** covering all command categories:
  - Movement: basic, word, line, paragraph, search, text objects
  - Editing: insert, delete, change, replace, case toggle, surround
  - Selection: expand, shrink, line selection, text object selection
  - Clipboard: yank, paste, registers
  - Advanced: macros, multiple cursors, search & replace

- **Quest system** with 5 quest types:
  - Command practice, scenario completion, speed runs
  - Time invested, exploration quests

### Fixed

- **Helix behavior accuracy improvements**
  - `~` (switch_case) now keeps cursor in place and toggles entire selection
  - `p`/`P` (paste) now keeps cursor on last pasted character, not after it
  - `i` (insert mode) now collapses selection when entering insert mode
  - Word movements (`w`, `b`) now correctly create selections in target state

- **XP breakdown display**
  - Separated mastery factor and repeat penalty in results screen
  - Fixed misleading "Learning: -30%" display (was showing combined reduction)

- **Scenario fixes**
  - Fixed scenario sorting to sort by difficulty first
  - Added missing selection fields to 13 movement scenarios
  - Fixed switch_case scenarios to match new cursor behavior
  - Fixed clipboard scenarios cursor positions

### Changed

- CI: Switched to `Swatinem/rust-cache@v2` for proper target directory caching (faster builds)

### Security

- Updated `lru` crate to 0.16.3 (RUSTSEC-2026-0002 soundness fix)
- Added ignore for RUSTSEC-2025-0141 (bincode unmaintained, transitive dep)

## [0.5.1] - 2026-01-07

### Changed

- **Refactored to helix-core primitives** (#110)
  - Migrated text objects (`miw`, `maw`, `mip`, `map`) to `helix_core::textobject`
  - Migrated character search (`f`, `F`, `t`, `T`) to `helix_core::search`
  - Migrated surround helpers to `helix_core::surround::find_nth_pairs_pos`
  - Migrated bracket matching (`mm`) to `helix_core::match_brackets`
  - Migrated paragraph movement (`{`, `}`) to `helix_core::movement`
  - Migrated toggle comments (`Ctrl-c`) to `helix_core::comment`
  - Migrated split selection on newlines (`Alt-s`) to `helix_core::selection`

- **Code cleanup and deduplication** (#111)
  - Extracted `extract_word_at_cursor` helper for search commands
  - Removed ~400 lines of excessive comments and documentation
  - Added 14 new unit tests for helper functions

### Fixed

- `join_selections_space` (`Alt-J`) now correctly selects inserted space (#110)

## [0.5.0] - 2026-01-07

### Added

**Major Feature: 45+ New Commands with Typestate Input Architecture**

This release significantly expands command coverage with a complete typestate-based input system for handling multi-key sequences.

- **Selection Commands (12 new)** (#100)
  - <kbd>s</kbd> select_regex: Select regex matches in selection
  - <kbd>S</kbd> split_selection: Split selection on regex
  - <kbd>Alt</kbd>-<kbd>s</kbd> split_selection_newlines: Split on newlines
  - <kbd>&</kbd> align_selections: Align selections to columns
  - <kbd>_</kbd> trim_selections: Trim whitespace from selections
  - <kbd>Alt</kbd>-<kbd>-</kbd> merge_selections: Merge all selections
  - <kbd>Alt</kbd>-<kbd>_</kbd> merge_consecutive: Merge adjacent selections
  - <kbd>C</kbd> copy_selection_next_line: Copy selection down
  - <kbd>Alt</kbd>-<kbd>C</kbd> copy_selection_prev_line: Copy selection up
  - <kbd>K</kbd> keep_selections_matching: Keep matching selections
  - <kbd>Alt</kbd>-<kbd>K</kbd> remove_selections_matching: Remove matching
  - <kbd>Ctrl</kbd>-<kbd>c</kbd> toggle_comments: Toggle line comments

- **Search Commands (6 new)** (#100)
  - <kbd>/</kbd> search_forward: Search forward with regex
  - <kbd>?</kbd> search_backward: Search backward with regex
  - <kbd>n</kbd> search_next: Jump to next match
  - <kbd>N</kbd> search_prev: Jump to previous match
  - <kbd>*</kbd> search_word_under_cursor: Search word forward
  - <kbd>Alt</kbd>-<kbd>*</kbd> search_selection: Search selection text

- **View Commands (6 new)** (#100)
  - <kbd>z</kbd>/<kbd>zz</kbd> view_center: Center view on cursor
  - <kbd>zt</kbd> view_top: Scroll cursor to top
  - <kbd>zb</kbd> view_bottom: Scroll cursor to bottom
  - <kbd>zm</kbd> view_center_horizontal: Center horizontally
  - <kbd>zj</kbd> scroll_down: Scroll view down
  - <kbd>zk</kbd> scroll_up: Scroll view up

- **Movement Commands (8 new)** (#105, #106)
  - <kbd>{</kbd> goto_prev_paragraph: Jump to previous paragraph
  - <kbd>}</kbd> goto_next_paragraph: Jump to next paragraph
  - <kbd>^</kbd> goto_first_nonblank: Alias for <kbd>gs</kbd>
  - <kbd>Ctrl</kbd>-<kbd>b</kbd> page_up: Scroll page up
  - <kbd>Ctrl</kbd>-<kbd>f</kbd> page_down: Scroll page down
  - <kbd>Ctrl</kbd>-<kbd>u</kbd> half_page_up: Scroll half page up
  - <kbd>Ctrl</kbd>-<kbd>d</kbd> half_page_down: Scroll half page down
  - <kbd>Alt</kbd>-<kbd>.</kbd> repeat_last_motion: Repeat last f/F/t/T

- **Editing Commands (3 new)** (#106)
  - <kbd>Alt</kbd>-<kbd>`</kbd> switch_to_uppercase: Convert to uppercase
  - <kbd>R</kbd> replace_with_yanked: Replace selection with yanked text
  - <kbd>Alt</kbd>-<kbd>J</kbd> join_selections_space: Join with spaces

- **Selection Management (3 new)** (#106)
  - <kbd>,</kbd> keep_primary_selection: Keep only primary selection
  - <kbd>Alt</kbd>-<kbd>,</kbd> remove_primary_selection: Remove primary selection
  - <kbd>Alt</kbd>-<kbd>x</kbd> shrink_to_line_bounds: Shrink selection to line bounds

- **Surround Commands (Match Mode)** (#107)
  - <kbd>ms</kbd> + char: Add surround (wrap selection with brackets/quotes)
  - <kbd>md</kbd> + char: Delete surround (remove enclosing pair)
  - <kbd>mr</kbd> + char + char: Replace surround (change pair type)

- **Text Object Selection** (#108)
  - <kbd>ma</kbd> + object: Select around (includes delimiters)
  - <kbd>mi</kbd> + object: Select inside (excludes delimiters)
  - Supported objects: `w`, `W`, `(`, `[`, `{`, `<`, `"`, `'`, `` ` ``, `p`

- **Lesson Navigation** (#103)
  - Next/Previous scenario navigation after completion
  - <kbd>n</kbd> to go to next lesson from Results screen
  - <kbd>p</kbd> to go to previous lesson
  - <kbd>l</kbd> to return to scenario list

### Changed

- **Typestate Input Architecture** (#100, #102)
  - Replaced `command_buffer: String` with `InputStateMachine`
  - Added `InputStateAccess` trait for uniform state access
  - 8 pending states: `FindCharPending`, `TillCharPending`, `SurroundAddPending`, `SurroundDeletePending`, `SurroundReplacePending`, `TextObjectAroundPending`, `TextObjectInsidePending`, `ViewPending`
  - Type-safe state transitions at compile time
  - `FindState` module to track last find/till motion for repeat

- **Command System**
  - Extended command constants for all new commands
  - Extended `cmd_to_key_events()` for multi-key command sequences
  - Unified handler pattern across all pending states

### Fixed

- **Alt-, Command** (#106)
  - Fixed `Alt-,` to correctly call `remove_primary_selection`
  - Was incorrectly mapped to `repeat_last_motion_reverse` (Vim feature, not Helix)

- **Empty Document Handling** (#108)
  - Fixed panic in `find_surrounding_pair()` when document is empty
  - Fixed panic in `select_around_paragraph()` / `select_inside_paragraph()` for empty documents
  - Added early returns with safe defaults for edge cases

### Quality

- **Tests**: 1116 (was 845, +271 new tests)
- **Clippy**: Zero warnings
- **Commands**: 45+ new commands (total ~90 commands supported)
- **Architecture**: Complete typestate input system for all multi-key sequences

## [0.4.8] - 2026-01-06

### Added

- **Code Coverage Integration** — Codecov workflow with coverage reports (#93)
- **UI Screenshots** — Added visual examples to README (#88)
- **Comprehensive Test Coverage** — 845 tests (up from ~760)
  - Handler tests: minigame.rs, scenario.rs, navigation.rs, profile.rs
  - Render module tests: helpers.rs, editor.rs, screen integration tests
  - Edge case coverage: Unicode, empty content, small/large terminals

### Changed

- **Architecture Refactoring** — 5-phase cleanup for better maintainability
  - Phase 1: Eliminate `Rc<RefCell<>>` from ProgressState (#91)
  - Phase 2: Remove GameState.session duplication (#92)
  - Phase 3: Standardize error handling with `UserError` (#94)
  - Phase 4: Introduce Game Service Layer (#95)
  - Phase 5: Migrate minigame handlers to HandlerContext pattern (#96)

- **CI/CD Improvements**
  - Removed sccache, kept cargo cache for simplicity (#98)
  - Merged coverage.yml into main ci.yml workflow
  - Bumped actions: upload-artifact v6, download-artifact v7, cache v5
  - Increased Miri job timeout to 90 minutes

- **Documentation**
  - Improved README structure and content
  - Reformatted keybindings with `<kbd>` tags
  - Simplified installation section (removed hardcoded version links)

### Dependencies

- `tokio`: 1.48.0 → 1.49.0
- `tui-big-text`: 0.7.3 → 0.8.1
- `toml`: 0.9.8 → 0.9.10
- `tempfile`: 3.23.0 → 3.24.0
- `serde_json`: 1.0.145 → 1.0.148
- `tracing`: 0.1.43 → 0.1.44
- `criterion`: 0.8.0 → 0.8.1

### Quality

- **Tests**: 845 (was ~760, +85 new tests)
- **Clippy**: Zero warnings
- **Architecture**: Cleaner separation of concerns with HandlerContext pattern

## [0.4.7] - 2025-12-03

### Changed

- **Constants Module** — Extracted hardcoded values to `src/constants.rs` (#77)
  - Centralized timing constants (tick rate, animation delays)
  - UI dimension constants (key history size, popup dimensions)
  - Game balance constants (XP values, streak multipliers)
  - Improved maintainability and configurability

- **Unified Command Handling** — Refactored command parsing between modes (#76)
  - Shared `CommandContext` trait for training and arcade modes
  - Eliminated code duplication in key handling
  - Consistent behavior across all game modes

### Fixed

- **Match Mode Multi-Key Handling** — Added <kbd>m</kbd> prefix to multi-key commands (#75)
  - <kbd>mm</kbd> now correctly detected as partial command (was incorrectly treating <kbd>m</kbd> as complete)
  - Fixes issue where Match Mode was not accessible in certain contexts

## [0.4.6] - 2025-12-02

### Added

- **Helix-style Menu Navigation** — Navigate scenario list with vim-like commands
  - <kbd>j</kbd>/<kbd>k</kbd> for up/down movement
  - <kbd>gg</kbd> to jump to first item, <kbd>G</kbd> to jump to last
  - Count prefixes: <kbd>5j</kbd> moves 5 items down, <kbd>10k</kbd> moves 10 up
  - <kbd>15G</kbd> or <kbd>15gg</kbd> jumps directly to item 15

- **Numeric Count Prefixes for Commands** — Execute commands multiple times
  - <kbd>3h</kbd> moves left 3 times, <kbd>5j</kbd> moves down 5 times
  - <kbd>2w</kbd> moves forward 2 words
  - Counts as single action for scoring (not N separate actions)

### Fixed

- **Match Mode Implementation** — <kbd>m</kbd> now correctly enters Match Mode
  - <kbd>mm</kbd> jumps to matching bracket (was incorrectly just <kbd>m</kbd>)
  - Matches official Helix keymap documentation

- **Count Prefix Scoring** — Commands like <kbd>3w</kbd> now count as 1 action, not 3
  - Added `record_action_with_count` method for proper scoring

- **Results Screen Progression Panel** — Improved XP and stats display
  - Added "Your Stats" section with Level, Total XP, Scenarios, Streak
  - Shows earned XP from scenario (+N) next to total
  - Fixed data source (was reading cleared state)

- **Unified Exit Keymap** — <kbd>Ctrl</kbd>-<kbd>Q</kbd> now works consistently across all screens
  - Menu/ModeSelection: exits application
  - Task/Results/Profile/Stats/Review/MiniGame: returns to previous screen

## [0.4.5] - 2025-12-02

### Added

- **Command Registry Architecture** — Type-safe O(1) command dispatch system
  - `CommandRegistry<M>` with mode-specific command registration
  - `CommandMetadata` for rich command documentation (name, key, description, category)
  - `KeyTrie` for efficient multi-key command resolution (<kbd>gg</kbd>, <kbd>ge</kbd>, <kbd>gh</kbd>, <kbd>gl</kbd>, <kbd>gs</kbd>)
  - Category-based organization (Movement, Editing, Selection, Clipboard)
  - Compile-time mode safety via PhantomData markers

### Fixed

- **Cursor display on empty lines** — Use block character (█) instead of space to prevent visual line duplication when cursor is on empty line
- **Append after word movement** — Fix `append()` to handle forward selections correctly (<kbd>e</kbd> + <kbd>a</kbd> now works properly)
- **Scenario corrections**:
  - `append_mode_001`: Fixed expected content from "hello !world" to "hello! world"
  - Removed non-existent <kbd>G</kbd> command (Helix uses <kbd>ge</kbd> for goto last line)
  - Removed duplicate `document_end_001` scenario (use `goto_last_line_001` with <kbd>ge</kbd>)
  - Updated hints to reference correct <kbd>ge</kbd> command instead of <kbd>G</kbd>

### Changed

- Simplified command dispatch from O(n) string matching to O(1) registry lookup
- Reduced cyclomatic complexity in command handling code
- 86 scenarios (removed 1 duplicate)

## [0.4.4] - 2025-12-01

### Added

- **Notification System** — Real-time notifications for level-ups, quest completions, and achievements
  - Auto-dismissing popups (3 seconds) in top-right corner
  - Color-coded by type (yellow/magenta/green/cyan)
  - Queue system for multiple simultaneous notifications

- **Filtering System** — Full implementation of scenario filters
  - Category filtering (toggle Movement, Editing, Clipboard, Advanced)
  - Difficulty filtering (toggle Beginner, Intermediate, Advanced)
  - Completion filtering (cycle: All → Completed → Not Completed)

- **Learning Scheduler Improvements** — Mix new content with reviews
  - 50/50 split between review items and new content in practice sessions
  - Weak command detection (low mastery, high difficulty, many lapses)
  - Untried essential command suggestions

- **Repeat Command Enhancement** — Full support for insert mode entry commands
  - Correctly replay <kbd>a</kbd>, <kbd>A</kbd>, <kbd>I</kbd>, <kbd>o</kbd>, <kbd>O</kbd> with proper entry point
  - `entry_command` field in `RepeatableAction::InsertSequence`

### Changed

- **Type-Safe Handler Architecture** — Compile-time guarantees for screen handlers
  - `HandlerContext<'a>` struct for borrowing non-screen state
  - `HandlerOutcome` enum (Stay/Transition) for explicit screen changes
  - `extract_screen!` macro for type-safe screen data extraction
  - 30+ handlers refactored to use new patterns

- **Unified Command Parsing** — Extracted shared logic into `CommandContext` trait
  - `ParsedCommand` enum (Complete/Partial/Invalid)
  - Single source of truth for command buffer parsing
  - Reduced code duplication between training and arcade modes

### Removed

- 23 redundant tests (3.2% reduction, coverage maintained)
- Phase/Stage/Iteration comments from documentation
- Invalid <kbd>dd</kbd> command references (replaced with <kbd>x</kbd> + <kbd>d</kbd> pattern)

### Fixed

- Unused import warnings after test cleanup
- Damaged "Add" comments from bulk replacement

### Quality

- **Tests**: 685 (was 708, -23 redundant)
- **Clippy**: Zero warnings
- **Code**: Cleaner documentation without development phase references

## [0.4.3] - 2025-12-01

### Added

- **MSRV Check** — CI now verifies build with minimum supported Rust version (1.89)
- **Miri Workflow** — Scheduled weekly undefined behavior detection (Sundays)
- **`#![forbid(unsafe_code)]`** — Enforced in lib.rs and main.rs

### Changed

- **MSRV bumped to 1.89** — Required by dependencies (bytemuck avx512_simd)
- **Float rounding** — XP calculations now use `.round()` for cross-platform consistency
- **Modern Rust syntax** — Refactored to use `if let && condition` chains

### Fixed

- Float-to-int conversion differences between native and Miri execution
- Async tests excluded from Miri (too slow, 300+ seconds each)

## [0.4.2] - 2025-11-30

### Added

**Expanded Commands** (#52)

- **14 New Commands**: Expanded from 31 to 45+ supported commands
  - Find/till character: <kbd>f</kbd>, <kbd>F</kbd>, <kbd>t</kbd>, <kbd>T</kbd> (jump to/before character)
  - Match brackets: <kbd>m</kbd> (jump to matching bracket)
  - Goto commands: <kbd>gh</kbd> (line start), <kbd>gl</kbd> (line end), <kbd>gs</kbd> (first non-whitespace), <kbd>ge</kbd> (last line)
  - Selection: <kbd>x</kbd> (select line), <kbd>X</kbd> (extend to line), <kbd>v</kbd> (select mode), <kbd>;</kbd> (collapse selection)
  - Case switching: <kbd>~</kbd> (toggle case)
  - Delete selection: <kbd>d</kbd> (delete current selection)

- **New Scenario Files**:
  - `find-till.toml`: <kbd>f</kbd>/<kbd>F</kbd>/<kbd>t</kbd>/<kbd>T</kbd> character search (5 scenarios)
  - `goto-commands.toml`: <kbd>gh</kbd>/<kbd>gl</kbd>/<kbd>gs</kbd>/<kbd>ge</kbd> navigation (4 scenarios)
  - `match-brackets.toml`: bracket matching (4 scenarios)
  - `line-selection.toml`: <kbd>x</kbd>/<kbd>X</kbd> line selection (3 scenarios)

### Changed

- **Helix-correct command behavior**:
  - <kbd>d</kbd> now executes immediately as delete selection (not waiting for <kbd>dd</kbd>)
  - <kbd>x</kbd> selects current line only (not including next line)
  - Idiomatic Helix: use <kbd>xd</kbd> to delete line (select + delete)

- **Selection visualization**: Selected text now highlighted with blue background

### Fixed

- Selection not being displayed visually after selection commands
- <kbd>d</kbd> command waiting for second <kbd>d</kbd> instead of executing immediately
- <kbd>x</kbd> command selecting two lines instead of one
- Selection end boundary including next line when it shouldn't

## [0.4.1] - 2025-11-30

### Added

**Expanded Content** (#51)

- **Training Scenarios**: Expanded from 25 to 78 scenarios (3.1x increase)
  - `basic-movement.toml`: <kbd>h</kbd>, <kbd>j</kbd>, <kbd>k</kbd>, <kbd>l</kbd> character navigation (9 scenarios)
  - `word-basics.toml`: <kbd>w</kbd>, <kbd>b</kbd> word navigation (7 scenarios)
  - `line-navigation.toml`: <kbd>0</kbd>, <kbd>gg</kbd> line/document navigation (7 scenarios)
  - `combined.toml`: multi-key navigation challenges (8 scenarios)
  - `precision.toml`: single-step precision movements (8 scenarios)
  - `deletion.toml`: <kbd>x</kbd>, <kbd>dd</kbd> deletion operations (8 scenarios)
  - `delete-advanced.toml`: advanced deletion patterns (6 scenarios)

- **Daily Quests**: Expanded from 12 to 55 quest templates (4.6x increase)
  - Easy (14): Movement commands (<kbd>h</kbd>, <kbd>j</kbd>, <kbd>k</kbd>, <kbd>l</kbd>, <kbd>w</kbd>, <kbd>b</kbd>, <kbd>0</kbd>, <kbd>gg</kbd>, <kbd>G</kbd>), editing (<kbd>x</kbd>, <kbd>dd</kbd>, <kbd>yy</kbd>)
  - Medium (23): Word navigation, insert/append modes, clipboard, undo/redo
  - Hard (18): Marathons, speed runs, time challenges, exploration quests

## [0.4.0] - 2025-11-30

### 🎮 Phase 2.1: Mini-Games Mode (Arcade Training)

This release introduces Mini-Games mode - an arcade-style training experience with time pressure and gamification mechanics designed to build muscle memory through fast-paced repetition.

### Added

**Mini-Games Mode** (#50)

- **Arcade Training Experience**
  - Mode selection screen (Training vs Arcade)
  - 60-second timed sessions with countdown
  - Automatic scenario progression with 2-second transitions
  - 5-10 second time limits per scenario based on difficulty

- **Scoring & Progression**
  - Real-time score tracking with streak multiplier (1.0x → 5.0x)
  - XP awards per scenario (15 base + streak bonus)
  - "+X XP" notification popup during transitions
  - High score tracking per session

- **Lives System**
  - Start with 3 lives
  - Lose life on timeout or excessive actions
  - Lives displayed with ❤️ indicators
  - Bonus life at 1000, 2500, 5000 points

- **UI/UX Improvements**
  - Key history display (last 5 keys, big text)
  - Pause menu (Esc) with profile/stats navigation
  - Game over screen with final score and stats
  - Countdown animation (3... 2... 1... GO!)

- **Integration**
  - Full XP/quest/FSRS integration
  - Profile and statistics accessible from pause menu
  - Proper navigation flow (arcade → pause → profile → arcade)

**Technical Implementation**:

- New `MiniGameSession` struct with state machine (Countdown → Playing → Transition → Paused → GameOver)
- `MiniGameData` screen state with command buffer and key history
- `ReturnDestination` enum for context-aware back navigation
- Unified `render_key_history_popup` (DRY refactoring)
- 82 new tests for mini-game functionality

### Changed

- BackToMenu now returns to ModeSelection (main mode menu)
- XP display fixed to show progress within current level (not total XP)
- Refactored render functions for code reuse between training and arcade modes

### Quality

- All 645 tests passing
- Zero clippy warnings
- Code formatted with rustfmt nightly
- Security audit passed (cargo deny check)

## [0.3.0] - 2025-11-30

### 🚀 Async Non-Blocking Architecture

This release introduces a complete async architecture overhaul, making the application instantly responsive with background data loading.

### Added

**Async Architecture** (#49)

- Tokio async runtime with non-blocking event loop
- `AsyncState<T>` enum for type-safe loading states (Loading/Ready/Failed)
- `DataLoadMessage` enum for channel-based communication
- Background data loaders using `tokio::task::spawn_blocking`
- Parallel loading of scenarios and profile at startup
- Biased `tokio::select!` for responsive keyboard handling

**New Modules**:
- `src/async_state.rs` (179 lines) - Type-safe async state management
- `src/data_loader.rs` (239 lines) - Background data loading functions

**Quest TOML System** (#48)

- Quest templates extracted to TOML files (`quests/en/daily.toml`)
- `QuestLoader` with security validation and limits
- `QuestTemplateRegistry` for template management
- 12 quest templates across 3 difficulty levels
- Strict TOML parsing with `#[serde(deny_unknown_fields)]`
- Validation for IDs, versions, locales, and level ranges

### Changed

**Performance Optimizations**:
- Runtime limited to 2 worker threads (sufficient for TUI workload)
- Instant UI display (~10-50ms to first render)
- Scenarios and profile load in parallel background tasks
- Keyboard events prioritized with biased select

**Code Quality** (#47, #45):
- Large functions split into focused helpers
- Hardcoded commands replaced with constants
- CI/CD improvements with cargo doc tests and benchmark checks

### Performance

- **Cold start**: 50-80ms (UI displays immediately)
- **Background loading**: Scenarios + profile load in parallel
- **Event loop**: Non-blocking, remains responsive during all operations
- **Memory overhead**: 1-2 MB for async runtime

### Quality

- All 563 tests passing (9 new async tests)
- Zero clippy warnings
- Security audit passed (cargo deny check)
- Comprehensive test coverage for `handle_data_message()`

### Security

- Bounded channels (32 messages) prevent memory exhaustion
- Error sanitization chain intact
- No race conditions (single-threaded event loop)
- Path validation layer for quest templates
- All input bounded by strict limits

## [0.2.1] - 2025-11-29

### Fixed

**Esc Key Conflict** (#41)
- Fixed critical conflict where <kbd>Esc</kbd> key was used for both abandoning scenarios and exiting Helix insert mode
- Changed scenario abandonment from <kbd>Esc</kbd> to <kbd>Ctrl</kbd>+<kbd>Q</kbd>
- Now users can properly use <kbd>Esc</kbd> to exit insert mode (native Helix behavior)
- Updated UI instructions to show <kbd>Ctrl</kbd>-<kbd>Q</kbd> instead of <kbd>Esc</kbd>

**FSRS Command Tracking** (#40)
- Fixed Review Session being non-functional due to missing command tracking
- Commands from completed scenarios now properly recorded in FSRS scheduler
- First reviews are immediately available (not delayed until next day)
- Review Commands menu now shows badge [N] when reviews are due

### Changed

**Major Code Quality Improvements** (#41)
- Refactored 6 complex functions into 20+ focused, maintainable functions
- Improved code readability and testability throughout codebase

**execute_command** (src/helix/simulator/commands/mod.rs):
- Split 193-line function into 4 focused functions (26+9+56+95+28 lines)
- Separated Insert mode and Normal mode logic
- Isolated repeat command recording
- Reduced cyclomatic complexity significantly

**handle_task_keys** (src/main.rs):
- Split 112-line function into 5 focused functions
- Clear separation: special keys, Insert mode input, Normal mode mapping
- Better maintainability for keyboard handling

**Configuration & Learning**:
- visit_toml_files: Functional style, no side effects (38 lines)
- update_fsrs_state: Separated FSRS calculation from state mutation (3 functions)
- generate_quests: Data-driven design with QuestDistribution struct
- validate_scenario: DRY principle with reusable validators

**State Handlers** (from v0.2.0, documented):
- Modularized state.rs into focused handler modules
- 8 handler modules: navigation, menu, scenario, gameplay, profile, quests, review, filters
- Reduced main state file from 2,084 to 1,486 lines

### Performance

**CI/CD Improvements**:
- Added cargo registry caching (crates.io index, .crate files, git deps)
- Combined with sccache: full build cache (registry + compilation)
- Dependency downloads: ~10-30s faster on cache hits
- Fixed sccache configuration for 88-100% cache hit rate

### Quality

- All 352 tests passing
- Zero clippy warnings
- Code formatted with rustfmt nightly
- Single Responsibility Principle applied throughout
- Improved testability and maintainability

## [0.2.0] - 2025-11-28

### 🎉 Phase 1 COMPLETE - Smart Learning System

This release completes Phase 1 of the helix-trainer roadmap by implementing the Interactive Review Session UI, the final missing piece of the FSRS spaced repetition system.

### Added

**Interactive Review Session UI** (#39)

- Review session screen with progress tracking ("Reviewing 3/5 commands")
- Command mastery display (Beginner → Intermediate → Advanced → Master)
- Next review date indicator based on FSRS scheduling
- Simple MVP interaction: <kbd>s</kbd> (success), <kbd>f</kbd> (failed), <kbd>Esc</kbd> (abandon)
- Menu integration with yellow badge `[N]` showing count of due reviews
- XP rewards system:
  - Base: 10 XP per command reviewed
  - Success rate bonus: 0-20 XP (Example: 5 reviews at 80% = 66 XP total)
- Keyboard shortcuts: Press <kbd>r</kbd> from menu to start review session

**Technical Implementation**:

- New `Screen::Review` state and `ReviewSessionState` structure
- Review screen rendering module (`src/ui/render/review.rs`)
- Message handlers for review session workflow
- Event handling for review interactions
- 33 comprehensive tests (3 unit tests passing, 30 integration tests documented)

### Performance

All performance targets exceeded:
- **Rendering**: 0.5ms per frame (60+ FPS, 32x faster than 16ms target)
- **Memory**: 440 bytes review state (56% of 1KB budget)
- **State updates**: O(1) complexity
- **Event handling**: <0.1ms (10x faster than 1ms target)

### Security

- **Security Rating**: A+ (rust-security-maintenance audit)
- Zero unsafe blocks
- All input validated through safe Rust enums
- Bounds checking on all array access
- Tamper-proof XP calculations
- Zero vulnerabilities found

### Quality

- All 352 library tests passing
- Zero clippy warnings (`clippy --all-targets --all-features -- -D warnings`)
- Code review approved by rust-code-reviewer agent
- Performance reviewed by rust-performance-engineer agent
- Security audited by rust-security-maintenance agent

### Changed

- Updated README to highlight Phase 1 completion
- Reorganized features section to emphasize Smart Learning System
- Menu now shows "Review Commands (r)" with due count badge

### Phase 1 Status: 100% Complete ✅

With this release, all Phase 1 components are fully implemented:
- ✅ FSRS spaced repetition backend
- ✅ Interactive review sessions (NEW in this release)
- ✅ Daily quest system
- ✅ XP/leveling progression
- ✅ Profile & Statistics screens
- ✅ Scenario mastery tracking
- ✅ Anti-farming protection

## [0.1.5] - 2025-11-28

### Added

**Phase B: Profile & Statistics Menu Integration**

- Main menu integration for Profile and Statistics screens
  - Added "View Profile (<kbd>p</kbd>)" menu item with keyboard shortcut
  - Added "Statistics (<kbd>s</kbd>)" menu item with keyboard shortcut
  - Profile and Statistics now accessible from main menu with arrow navigation
  - Visual separator grouping system options (Profile, Statistics, Quit)
  - Keyboard hints displayed in menu items

- Enhanced navigation
  - Press <kbd>p</kbd> for instant Profile screen access (1-keypress shortcut)
  - Press <kbd>s</kbd> for instant Statistics screen access (1-keypress shortcut)
  - Arrow keys / <kbd>j</kbd>/<kbd>k</kbd> navigation through all menu items
  - Number keys (<kbd>1</kbd>-<kbd>9</kbd>) still work for scenario shortcuts

### Changed

- Menu structure: Scenarios → Separator → Profile → Statistics → Quit
- Updated menu navigation logic to support +3 items (Profile, Statistics, Quit)
- Quit option moved to bottom after Profile and Statistics

### Fixed

- Profile and Statistics screens now discoverable (previously only accessible from results screen)
- Menu scrolling works correctly with new items

## [0.1.4] - 2025-11-28

### Added

**Phase 1.5: Scenario Metadata & Filtering System**

- Rich scenario metadata system
  - `ScenarioMetadata` with category, difficulty, tags, and taught commands
  - 25 scenarios fully categorized: Movement (5), Editing (11), Clipboard (3), Advanced (6)
  - Difficulty levels: Beginner (20), Intermediate (5)
  - Each scenario tagged with taught commands and practice focus areas

- Flexible filtering and sorting system
  - `ScenarioCollection` with six sort modes (alphabetical, difficulty, category, completion, recent, random)
  - Filter by category (Movement, Editing, Clipboard, Advanced)
  - Filter by difficulty (Beginner, Intermediate, Advanced)
  - Filter by taught commands (e.g., show all scenarios teaching <kbd>w</kbd> or <kbd>dd</kbd>)
  - Filter by completion status (completed vs. uncompleted)
  - Chainable filters for complex queries
  - 100% backward compatible with existing scenarios (metadata optional)

- Enhanced UI indicators
  - Difficulty badges: 🟢 Beginner / 🟡 Intermediate / 🔴 Advanced
  - Completion status: ✅ for completed scenarios
  - Visual feedback in scenario menu

- Automated scenario validation tests
  - `test_all_scenarios_load_successfully`: validates all 25 scenarios load without errors
  - `test_all_scenarios_execute_solution`: executes solution commands and verifies target state
  - Catches cursor position errors, invalid command formats, and other issues
  - Runs during CI to prevent scenario regressions

- Performance benchmarks
  - Criterion benchmarks in `benches/filtering_sorting.rs`
  - <1ms filtering/sorting for collections of 1000 scenarios
  - Memory-efficient metadata storage

### Changed

- Integrated `ScenarioCollection` into `AppState` (replacing raw `Vec<Scenario>`)
- Menu now displays difficulty and completion indicators for each scenario
- Updated all 25 scenarios with complete metadata (category, difficulty, tags, taught commands)

### Fixed

- Corrected cursor positions in multiple scenarios:
  - `repeat_insert_001`: Fixed target cursor [1, 17] → [1, 16]
  - `paste_before_001`: Fixed target cursor [0, 1] → [0, 2]
  - `delete_line_001`: Start cursor on line 1 instead of line 0
  - `document_end_001`: Cursor at [4, 9] (end of document)
  - `select_word_001`: Cursor at [0, 9] (after word end)
  - `repeat_indent_001`: Cursor [2, 2] → [2, 4] (after second indent)

- Fixed command format in scenarios:
  - Changed multi-key commands from arrays to strings (e.g., `["d", "d"]` → `["dd"]`)
  - Applies to: <kbd>dd</kbd>, <kbd>gg</kbd>, <kbd>r</kbd>_, and other multi-key sequences

- Fixed repeat command repeatability:
  - Made all printable ASCII characters and space repeatable in `is_repeatable_command()`
  - Allows replace commands like <kbd>r</kbd>_ to work with repeat (<kbd>.</kbd>)

- Fixed indentation in `repeat_indent_001`:
  - Changed from 4-space to 2-space indentation to match simulator behavior

- Fixed append mode behavior in `append_mode_001`:
  - Adjusted target content to match actual <kbd>e</kbd> command behavior (cursor after word, not on last char)

### Performance

- Filtering 1000 scenarios: ~50-200 µs (well under 1ms budget)
- Sorting 1000 scenarios: ~100-400 µs depending on sort mode
- Zero overhead for scenarios without metadata (backward compatible)

### Security

- Added `walkdir = "2.5"` dev-dependency for safe directory traversal in tests
- All scenario validation tests run with bounds checking

### Tests

- Added 16 comprehensive unit tests for `ScenarioCollection`
- Added 2 integration tests for scenario validation
- Total test count: 392 (all passing)
- Zero clippy warnings

## [0.1.3] - 2025-11-28

### Fixed

- **Hint system completely non-functional**: Hints were stored in wrong TOML section and never loaded
  - Moved `hints` arrays from `[scenarios.solution]` to `[[scenarios]]` level in all 11 scenario files
  - Fixed 24 scenarios containing 55 total hints
  - Hints now properly deserialize and display when requested
- **Hint key conflict**: UI showed `[h: Show Hint]` but <kbd>h</kbd> is Helix left movement command
  - Added <kbd>?</kbd> as primary hint key (intuitive, no conflicts)
  - Kept <kbd>F1</kbd> as alternative for accessibility
  - Handles both `Char('?')` and `Char('/')` + <kbd>Shift</kbd> for cross-platform support
  - Updated UI to show `[?: Hint | F1]`

### Added

- **Hint toggle behavior**: Press <kbd>?</kbd> to open hint, press again to close (improved UX)
- **Cross-platform hint key support**: Properly handles different keyboard layouts and modifier keys

## [0.1.2] - 2025-11-28

### Fixed

- Indent/dedent commands (<kbd>></kbd>, <kbd><</kbd>) not working in TUI (#35)

## [0.1.1] - 2025-11-28

### Added

- Scenario mastery tracking system to prevent XP farming
  - Three-tier mastery levels: Learning → Proficient → Mastered
  - Graduated XP scaling: 100% → 50% → 20%
  - Session spam protection (same-day repeat penalties: 100% → 70% → 30%)
  - Prevents farming by reducing XP for mastered scenarios by 80%
  - New module: `src/learning/scenario_history.rs` (600+ lines)
  - 23 comprehensive tests for mastery tracking

- Bounded scenario tracking (MAX_SCENARIOS_TRACKED = 10,000)
  - Prevents unbounded memory growth
  - Defense in depth against DoS attacks

- Scenario ID validation at storage boundary
  - Validates alphanumeric + underscore + hyphen only
  - MAX_SCENARIO_ID_LENGTH = 100 characters
  - Blocks path traversal and injection attempts

- Performance benchmarks and profiling
  - Criterion benchmarks: `benches/xp_scaling.rs`
  - DHAT memory profiling: `examples/memory_profile.rs`

### Changed

- Replaced `String` dates with `chrono::NaiveDate` for type safety
- Consistent `.round()` usage in XP calculations (prevents truncation)
- Mastery level displayed in results screen with emoji indicators

### Fixed

- Floating-point precision issues in XP calculations
- Potential memory exhaustion from unbounded HashMap growth

### Performance

- XP calculation overhead: 260-460 ns (2000x under 1ms budget)
- Memory per scenario: ~288 bytes
- Profile serialization: ~40 µs for 100 scenarios

### Security

- Added MAX_SCENARIOS_TRACKED limit (defense in depth)
- Added scenario ID validation (alphanumeric + `_` + `-` only)
- Type-safe date handling with NaiveDate
- All arithmetic uses saturating operations

## [0.1.0] - 2025-11-28

### Added

**Phase 1: Smart Learning & Gamification System**

- FSRS-based spaced repetition learning system for scientifically-optimized practice
  - Card state management (New, Learning, Review, Relearning)
  - Performance tracking per scenario (Again, Hard, Good, Easy ratings)
  - Intelligent review scheduling (20-30% fewer reviews than traditional methods)
  - Review analytics and statistics
  - Persistent storage in `~/.config/helix-trainer/learning.json`

- Daily quests gamification system with Duolingo-style mechanics
  - Three quest types: Practice (3 scenarios), Learning (2 new), Review (5 cards)
  - Daily quest rotation at midnight local time
  - Streak tracking with loss aversion mechanics
  - Quest completion rewards (XP bonuses)
  - Persistent storage in `~/.config/helix-trainer/quests.json`

- XP and leveling progression system
  - Experience points for scenario completion
  - Performance-based XP multipliers (1.0x to 2.0x)
  - Quest completion bonuses (50-150 XP)
  - Level progression with exponential scaling
  - Profile screen with detailed statistics

- User profile and statistics tracking
  - Total scenarios completed
  - Average performance score
  - Current streak and longest streak
  - Total XP and current level
  - Review history and analytics
  - Persistent storage in `~/.config/helix-trainer/profile.json`

- Enhanced UI screens
  - Profile screen with level, XP, and statistics
  - Statistics screen with review calendar and performance graphs
  - Quest tracker in main menu
  - XP breakdown in results screen
  - Success animations with XP rewards

**Phase A: Foundation (Complete)**

- Core TUI framework with ratatui 0.29
  - Main menu with scenario selection
  - Task screen with editor simulation
  - Results screen with performance rating
  - Keyboard-driven navigation (Vim-style)

- Helix simulator using official helix-core library (v25.07.1)
  - 30+ commands: <kbd>h</kbd>,<kbd>j</kbd>,<kbd>k</kbd>,<kbd>l</kbd>,<kbd>w</kbd>,<kbd>b</kbd>,<kbd>e</kbd>,<kbd>0</kbd>,<kbd>$</kbd>,<kbd>x</kbd>,<kbd>dd</kbd>,<kbd>i</kbd>,<kbd>a</kbd>,<kbd>I</kbd>,<kbd>A</kbd>,<kbd>o</kbd>,<kbd>O</kbd>,<kbd>r</kbd>,<kbd>c</kbd>,<kbd>y</kbd>,<kbd>p</kbd>,<kbd>P</kbd>,<kbd>u</kbd>,<kbd>U</kbd>,<kbd>gg</kbd>,<kbd>G</kbd>,<kbd>J</kbd>,<kbd>></kbd>,<kbd><</kbd>
  - Repeat command (<kbd>.</kbd>) for efficient editing workflows
  - Insert mode with text input, <kbd>Backspace</kbd>, arrow keys
  - Multi-key command buffer (<kbd>dd</kbd>, <kbd>gg</kbd>)
  - Yank/paste clipboard support
  - Automatic completion detection
  - Cursor and selection visualization

- Beautiful UI components
  - Large key history display (tui-big-text, 8-line tall characters)
  - Success popup with 1.5s delay
  - Diff highlighting (red/green) in results
  - Action count indicators
  - Performance rating with emoji (Perfect/Excellent/Good/Fair/Poor)
  - Hint system (<kbd>F1</kbd> key)

- TOML scenario system with security validation
  - 20 training scenarios (basic to intermediate)
  - Configurable setup, target, solution, hints
  - Performance scoring (optimal count, max points, tolerance)
  - Input validation and sanitization
  - Path traversal prevention
  - Content size limits

- Testing and quality infrastructure
  - 164 passing tests (unit, integration, property-based)
  - Zero clippy warnings policy
  - cargo-nextest for fast test execution
  - Proptest for property-based testing
  - 100% test coverage for core modules

- Cross-platform CI/CD pipeline
  - GitHub Actions workflows for Linux, macOS, Windows
  - Automated testing with cargo-nextest
  - Rustfmt (nightly) formatting checks
  - Clippy lints with -D warnings
  - Security audit with cargo-deny
  - Rust 2024 edition support
  - MSRV: 1.85

- Internationalization (i18n) support
  - rust-i18n integration for multi-language support
  - English locale (default)
  - Infrastructure for future language additions

- Build optimization
  - sccache integration for fast incremental builds
  - Cargo cache optimization
  - 5x faster rebuilds (10s vs 54s)

- Release automation
  - Cross-platform binary builds (9 platforms)
  - Linux x86_64 (GNU and musl)
  - Linux ARM64 (aarch64 GNU and musl)
  - macOS x86_64 (Intel)
  - macOS ARM64 (Apple Silicon M1/M2/M3)
  - Windows x86_64
  - Windows ARM64
  - Automatic GitHub releases on tags
  - Changelog generation and extraction
  - SHA256 checksums for all binaries

- Documentation improvements
  - Problem-solution framing in README
  - Quick Links navigation section
  - Consolidated Features section
  - Code examples with real output
  - CONTRIBUTING.md extraction
  - CI status badge
  - Visual TUI example

### Changed

- Upgraded to Rust 2024 edition (latest stable)
- Migrated to helix-core 25.07.1 (from 24.07)
- Optimized dependencies (removed tokio, unnecessary features)
- Refactored codebase into 31 modular files
- Enhanced error messages with context and suggestions
- Improved scoring algorithm for better feedback

### Fixed

- Replace command cursor initialization bugs
- Multi-key command handling edge cases
- Insert mode arrow key navigation
- Yank/paste clipboard state management
- Undo/redo transaction handling

### Security

- All input validation with bounds checking
- Path traversal prevention in scenario loading
- Content size limits (1MB scenarios, 100KB files)
- Safe arithmetic with checked operations
- No unsafe blocks in codebase
- Dependency security audits with cargo-deny

## [0.0.1] - 2024-10-01

### Added

- Initial project structure
- Basic TUI skeleton with ratatui
- Prototype scenario loader
- Proof of concept demo

---

[Unreleased]: https://github.com/bug-ops/helix-trainer/compare/v0.5.12...HEAD
[0.5.12]: https://github.com/bug-ops/helix-trainer/compare/v0.5.11...v0.5.12
[0.5.11]: https://github.com/bug-ops/helix-trainer/compare/v0.5.10...v0.5.11
[0.5.10]: https://github.com/bug-ops/helix-trainer/compare/v0.5.9...v0.5.10
[0.5.9]: https://github.com/bug-ops/helix-trainer/compare/v0.5.8...v0.5.9
[0.5.8]: https://github.com/bug-ops/helix-trainer/compare/v0.5.7...v0.5.8
[0.5.7]: https://github.com/bug-ops/helix-trainer/compare/v0.5.6...v0.5.7
[0.5.6]: https://github.com/bug-ops/helix-trainer/compare/v0.5.5...v0.5.6
[0.5.5]: https://github.com/bug-ops/helix-trainer/compare/v0.5.4...v0.5.5
[0.5.4]: https://github.com/bug-ops/helix-trainer/compare/v0.5.3...v0.5.4
[0.5.3]: https://github.com/bug-ops/helix-trainer/compare/v0.5.2...v0.5.3
[0.5.2]: https://github.com/bug-ops/helix-trainer/compare/v0.5.1...v0.5.2
[0.5.1]: https://github.com/bug-ops/helix-trainer/compare/v0.5.0...v0.5.1
[0.5.0]: https://github.com/bug-ops/helix-trainer/compare/v0.4.8...v0.5.0
[0.4.8]: https://github.com/bug-ops/helix-trainer/compare/v0.4.7...v0.4.8
[0.4.7]: https://github.com/bug-ops/helix-trainer/compare/v0.4.6...v0.4.7
[0.4.6]: https://github.com/bug-ops/helix-trainer/compare/v0.4.5...v0.4.6
[0.4.5]: https://github.com/bug-ops/helix-trainer/compare/v0.4.4...v0.4.5
[0.4.4]: https://github.com/bug-ops/helix-trainer/compare/v0.4.3...v0.4.4
[0.4.3]: https://github.com/bug-ops/helix-trainer/compare/v0.4.2...v0.4.3
[0.4.2]: https://github.com/bug-ops/helix-trainer/compare/v0.4.1...v0.4.2
[0.4.1]: https://github.com/bug-ops/helix-trainer/compare/v0.4.0...v0.4.1
[0.4.0]: https://github.com/bug-ops/helix-trainer/compare/v0.3.0...v0.4.0
[0.3.0]: https://github.com/bug-ops/helix-trainer/compare/v0.2.1...v0.3.0
[0.2.1]: https://github.com/bug-ops/helix-trainer/compare/v0.2.0...v0.2.1
[0.2.0]: https://github.com/bug-ops/helix-trainer/compare/v0.1.5...v0.2.0
[0.1.5]: https://github.com/bug-ops/helix-trainer/compare/v0.1.4...v0.1.5
[0.1.4]: https://github.com/bug-ops/helix-trainer/compare/v0.1.3...v0.1.4
[0.1.3]: https://github.com/bug-ops/helix-trainer/compare/v0.1.2...v0.1.3
[0.1.2]: https://github.com/bug-ops/helix-trainer/compare/v0.1.1...v0.1.2
[0.1.1]: https://github.com/bug-ops/helix-trainer/compare/v0.1.0...v0.1.1
[0.1.0]: https://github.com/bug-ops/helix-trainer/releases/tag/v0.1.0
[0.0.1]: https://github.com/bug-ops/helix-trainer/releases/tag/v0.0.1