claudedesk 4.4.1 → 4.6.0

package/CHANGELOG.md

@@ -8,13 +8,90 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 ## [Unreleased]
 
 ### Planned
-- Screenshot capture for README documentation
-- Automated tests with Jest/Vitest
 - Homebrew/Chocolatey packaging for easier installation
 - Development dependency security updates (electron v40+, vite v7+)
 
 ---
 
+## [4.6.0] - 2026-02-19
+
+### Added
+- **LaunchTunnel integration** — Expose local ports to the internet via LaunchTunnel (14th domain)
+  - `TunnelManager` with hybrid REST API + CLI (`lt preview`) process management
+  - API key management with validation, stored in `~/.claudedesk/tunnel-settings.json`
+  - Tunnel list with 30s cache, status mapping (API snake_case → camelCase)
+  - Account info section (email, plan badge, status)
+  - Request log viewer with method/path/status/duration/size columns
+  - TunnelPanel (4 views: setup/main/settings/logs), TunnelCreateDialog, TunnelRequestLogs
+  - CLI auto-detection via `where`/`which`, `shell: true` for cross-platform spawn
+  - Ctrl+Shift+U keyboard shortcut, ToolsDropdown entry with active count badge
+  - 17 IPC methods (`tunnel:*`): 13 invoke + 4 events
+
+### Fixed
+- **Tunnel spawn ENOENT** — Use `shell: true` in spawn for cross-platform `.cmd` shim compatibility
+- **Tunnel status mapping** — API returns `"active"` not `"connected"`; added to `mapApiStatus`
+- **Tunnel CLI subcommand** — Fixed `lt preview --port` (was missing `preview` subcommand)
+- **Tunnel URL parsing** — Match `URL:` output format from LaunchTunnel CLI (not `your url is:`)
+- **API snake_case mapping** — Map `created_at`, `status_code`, `duration_ms` etc. to camelCase
+- **Account response unwrapping** — API returns `{"user": {...}}` wrapper; now unwrapped correctly
+- **Request logs endpoint** — Reverted from `/requests` back to correct `/logs` path
+
+### Changed
+- **IPC contract** — Expanded from 149 to ~166 methods (added 17 tunnel methods)
+- **Project scale** — ~150 source files, ~49,000 LOC, 14 domains, 14 managers
+
+---
+
+## [4.5.0] - 2026-02-17
+
+### Added
+- **Testing infrastructure** — Vitest 4 + @testing-library/react + Playwright for Electron
+  - 250 tests across 20 test files, 3 workspace projects (shared/main/renderer)
+  - Auto-derived electronAPI mock from IPC contract
+  - E2E tests with Playwright for Electron (app launch, sessions, split view, keyboard shortcuts)
+  - CI workflow with coverage artifacts
+- **Git integration** — Full Git panel with staging, commits, branches, and real-time status
+  - `GitManager` with `child_process.execFile` (shell injection safe), per-directory mutex, `.git` fs.watch()
+  - File staging/unstaging (individual + bulk), branch display, commit history log
+  - AI commit message generation (heuristic-based conventional commits format)
+  - Optional checkpoint creation on commit
+  - Ctrl+Shift+G keyboard shortcut, ToolsDropdown entry with staged count badge
+  - 30 IPC methods (`git:*`): 26 invoke + 4 events
+- **Diff viewer** — Full-screen diff overlay with syntax-highlighted unified diffs
+  - Categorized file navigation (staged/unstaged/untracked/conflicted)
+  - Dual gutter line numbers, colored add/remove/context lines
+  - Keyboard navigation (J/K between files), stage/unstage/discard actions from diff view
+  - Unified diff parser (`diff-parser.ts`) with old/new line number tracking
+- **Git worktrees** — Worktree management panel with create, remove, and prune operations
+  - `WorktreePanel` for listing and managing worktrees
+  - `WorktreeCleanupDialog` for cleanup prompts when closing managed worktree sessions
+- **Session playbooks** — Automated multi-step prompt sequences (13th domain)
+  - `PlaybookManager` (CRUD + persistence), `PlaybookExecutor` (execution engine)
+  - 5 built-in playbooks: API endpoint, bug investigation, code review, component creation, refactor
+  - Silence-based step completion (3s no output = done), confirmation gates between steps
+  - Dynamic parameter forms (text/multiline/select/filepath), variable interpolation
+  - PlaybookPicker (fuzzy search), PlaybookEditor (3-tab slide-in), PlaybookProgressPanel (bottom-docked)
+  - Import/export playbooks as JSON, library browser for built-in + custom playbooks
+  - Persistence: `~/.claudedesk/playbooks.json`
+  - Ctrl+Shift+B keyboard shortcut, ToolsDropdown entry
+  - 15 IPC methods (`playbook:*`): 12 invoke + 3 events
+- **Repository Atlas Engine (RAE)** — Automated CLAUDE.md + repo-index.md generation
+  - File enumeration via `git ls-files`, regex import analysis, 3-tier domain inference
+  - AtlasPanel UI with idle/scanning/preview states
+  - 6 IPC methods (`atlas:*`)
+
+### Fixed
+- **Git generate button** — Action bar Generate now auto-opens CommitDialog with the generated message
+- **CommitDialog generate button** — Directly sets title from return value instead of relying on fragile useEffect prop chain
+- **Silent null guard in useGit** — `generateMessage()` now shows error toast when no project directory is available instead of silently returning
+- **Stale generated message** — Generated commit message state is cleared after successful commit
+
+### Changed
+- **IPC contract** — Expanded from 102 to 149 methods (118 invoke + 8 send + 23 event)
+- **Project scale** — 138 source files, ~45,800 LOC, 13 domains, 13 managers
+
+---
+
 ## [4.4.1] - 2026-02-13
 
 ### Fixed
@@ -203,7 +280,9 @@ See [CONTRIBUTING.md](CONTRIBUTING.md) for guidelines on suggesting changes and
 
 ---
 
-[Unreleased]: https://github.com/carloluisito/claudedesk/compare/v4.4.1...HEAD
+[Unreleased]: https://github.com/carloluisito/claudedesk/compare/v4.6.0...HEAD
+[4.6.0]: https://github.com/carloluisito/claudedesk/compare/v4.5.0...v4.6.0
+[4.5.0]: https://github.com/carloluisito/claudedesk/compare/v4.4.1...v4.5.0
 [4.4.1]: https://github.com/carloluisito/claudedesk/compare/v4.3.1...v4.4.1
 [4.3.1]: https://github.com/carloluisito/claudedesk/compare/v4.3.0...v4.3.1
 [4.3.0]: https://github.com/carloluisito/claudedesk/compare/v4.1.1...v4.3.0

package/CLAUDE.md

@@ -2,6 +2,10 @@
 
 Electron desktop app wrapping Claude Code CLI with multi-session terminals, split views, and agent team visualization.
 
+## Workflow Rule
+
+**Always run the `requirements-clarifier` agent (via the Task tool) on the user's initial prompt before planning or implementing.** This ensures requirements are analyzed, ambiguities are surfaced, and acceptance criteria are established before any work begins. Skip only for trivial tasks (typo fixes, single-line changes, or purely informational questions).
+
 ## Tech Stack
 
 Electron 28 | React 18 | TypeScript | xterm.js | node-pty | Tailwind CSS | reactflow
@@ -11,9 +15,9 @@ Electron 28 | React 18 | TypeScript | xterm.js | node-pty | Tailwind CSS | react
 ```
 ┌─────────────────────────────────────────────┐
 │  Main Process (Node.js)                     │
-│  9 managers + IPC handlers + session pool   │
+│  14 managers + IPC handlers + session pool  │
 └──────────────────┬──────────────────────────┘
-                   │ IPC (102 methods)
+                   │ IPC (~166 methods)
 ┌──────────────────┴──────────────────────────┐
 │  Preload (auto-derived context bridge)      │
 └──────────────────┬──────────────────────────┘
@@ -26,7 +30,7 @@ Electron 28 | React 18 | TypeScript | xterm.js | node-pty | Tailwind CSS | react
 
 **3-layer pattern per domain:** Manager (main) → Hook (renderer) → Components (renderer)
 
-**IPC contract** (`src/shared/ipc-contract.ts`) is the single source of truth — 102 methods. The preload bridge and `ElectronAPI` type are auto-derived from it.
+**IPC contract** (`src/shared/ipc-contract.ts`) is the single source of truth — ~166 methods. The preload bridge and `ElectronAPI` type are auto-derived from it.
 
 ## Domain Map
 
@@ -42,7 +46,9 @@ Electron 28 | React 18 | TypeScript | xterm.js | node-pty | Tailwind CSS | react
 | Drag-Drop | file-dragdrop-handler, file-utils | useDragDrop, DragDropOverlay, DragDropContextMenu, DragDropSettings | ipc-types.ts | `dragdrop:*` |
 | Workspaces | settings-persistence | SettingsDialog | ipc-types.ts | `workspace:*` |
 | Atlas | atlas-manager | useAtlas, AtlasPanel | types/atlas-types.ts | `atlas:*` |
-| Git | git-manager | useGit, GitPanel, CommitDialog | types/git-types.ts | `git:*` |
+| Git | git-manager | useGit, useDiffViewer, GitPanel, DiffViewer, DiffFileNav, DiffViewerHeader, DiffContentArea, CommitDialog, WorktreePanel, WorktreeCleanupDialog, diff-parser | types/git-types.ts | `git:*` |
+| Playbooks | playbook-manager, playbook-executor, built-in-playbooks | usePlaybooks, PlaybookPicker, PlaybookParameterDialog, PlaybookProgressPanel, PlaybookPanel, PlaybookEditor | types/playbook-types.ts | `playbook:*` |
+| Tunnels | tunnel-manager | useTunnel, TunnelPanel, TunnelCreateDialog, TunnelRequestLogs | types/tunnel-types.ts | `tunnel:*` |
 | Window | index.ts | ConfirmDialog, SettingsDialog, AboutDialog, TitleBarBranding | ipc-types.ts | `window:*`, `dialog:*` |
 
 ## Adding a New IPC Method
@@ -72,6 +78,7 @@ That's it. The preload bridge and types auto-derive.
 - **Split view**: `useSplitView` manages a tree of leaf/branch nodes. Max 4 panes. State persisted in settings.
 - **Agent team detection**: `AgentTeamManager` watches `~/.claude/teams/` and `~/.claude/tasks/` via `fs.watch()`. Auto-links sessions within 30s of team creation.
 - **Git integration**: `GitManager` uses `child_process.execFile` (not `exec` — prevents shell injection). Per-directory mutex serializes operations. `.git` directory watching with 500ms debounce for real-time status. Heuristic-based AI commit message generation (conventional commits format).
+- **Playbook execution**: `PlaybookExecutor` writes prompts to PTY and uses silence-based detection (3s no output = step done) for step completion. Never sends Ctrl+C on cancel (just stops sending further steps). One execution per session, confirmation gates pause between steps.
 
 ## Testing
 

package/dist/main/built-in-playbooks.js

@@ -0,0 +1,231 @@
+"use strict";
+Object.defineProperty(exports, "__esModule", { value: true });
+exports.BUILT_IN_PLAYBOOKS = void 0;
+const DEFAULT_EXECUTION = {
+    silenceThresholdMs: 3000,
+    interStepDelayMs: 1000,
+    stepTimeoutMs: 300000,
+    stepTimeoutPolicy: 'pause',
+    createCheckpointBeforeRun: false,
+};
+let stepCounter = 0;
+function stepId() {
+    return `builtin-step-${++stepCounter}`;
+}
+exports.BUILT_IN_PLAYBOOKS = [
+    {
+        id: 'builtin-add-api-endpoint',
+        type: 'built-in',
+        name: 'Add API Endpoint',
+        description: 'Scaffold a new REST API endpoint with route, handler, validation, and tests.',
+        icon: '\u{1F310}',
+        category: 'Backend',
+        keywords: ['api', 'endpoint', 'route', 'rest', 'handler', 'scaffold'],
+        variables: [
+            { name: 'method', label: 'HTTP Method', type: 'select', required: true, options: ['GET', 'POST', 'PUT', 'PATCH', 'DELETE'] },
+            { name: 'path', label: 'Route Path', type: 'text', required: true, placeholder: '/api/users/:id' },
+            { name: 'description', label: 'Endpoint Description', type: 'text', required: true, placeholder: 'Fetch user by ID' },
+            { name: 'auth', label: 'Authentication', type: 'select', required: true, default: 'required', options: ['required', 'optional', 'none'] },
+        ],
+        steps: [
+            {
+                id: stepId(),
+                name: 'Analyze existing patterns',
+                prompt: 'Look at the existing API routes and handlers in this project. Identify the patterns used for routing, request validation, error handling, and response formatting. Summarize the conventions you find.',
+                requireConfirmation: false,
+            },
+            {
+                id: stepId(),
+                name: 'Create route and handler',
+                prompt: 'Create a new {{method}} endpoint at {{path}} — {{description}}. Authentication: {{auth}}. Follow the existing project patterns you identified. Include request validation and proper error handling.',
+                requireConfirmation: true,
+            },
+            {
+                id: stepId(),
+                name: 'Add tests',
+                prompt: 'Write tests for the new {{method}} {{path}} endpoint. Cover: successful request, validation errors, authentication (if {{auth}} is not "none"), and edge cases. Follow the existing test patterns.',
+                requireConfirmation: false,
+            },
+            {
+                id: stepId(),
+                name: 'Verify and document',
+                prompt: 'Run the tests you just wrote. If any fail, fix them. Then add a brief description of the new endpoint to any existing API documentation files if present.',
+                requireConfirmation: false,
+            },
+        ],
+        execution: DEFAULT_EXECUTION,
+        createdAt: 0,
+        updatedAt: 0,
+    },
+    {
+        id: 'builtin-bug-investigation',
+        type: 'built-in',
+        name: 'Bug Investigation',
+        description: 'Systematic bug investigation: reproduce, find root cause, fix, and verify.',
+        icon: '\u{1F41B}',
+        category: 'Debugging',
+        keywords: ['bug', 'debug', 'fix', 'investigate', 'error', 'issue'],
+        variables: [
+            { name: 'bug_description', label: 'Bug Description', type: 'multiline', required: true, placeholder: 'Describe the bug behavior...' },
+            { name: 'expected_behavior', label: 'Expected Behavior', type: 'text', required: true, placeholder: 'What should happen instead?' },
+            { name: 'affected_area', label: 'Affected Area', type: 'text', required: false, placeholder: 'e.g., login page, API, database' },
+        ],
+        steps: [
+            {
+                id: stepId(),
+                name: 'Understand and locate',
+                prompt: 'I need to investigate this bug:\n\n{{bug_description}}\n\nExpected behavior: {{expected_behavior}}\nAffected area: {{affected_area}}\n\nSearch the codebase to find the relevant files and understand the current code flow. Identify where the bug likely originates.',
+                requireConfirmation: false,
+            },
+            {
+                id: stepId(),
+                name: 'Root cause analysis',
+                prompt: 'Based on what you found, identify the root cause of the bug. Explain the exact code path that leads to the incorrect behavior. Show the specific lines of code involved.',
+                requireConfirmation: true,
+            },
+            {
+                id: stepId(),
+                name: 'Implement fix',
+                prompt: 'Implement the minimal fix for this bug. Change only what is necessary to correct the behavior. Do not refactor surrounding code.',
+                requireConfirmation: true,
+            },
+            {
+                id: stepId(),
+                name: 'Verify fix',
+                prompt: 'Run any existing tests related to the fix. If tests fail, fix them. If no tests cover this area, write a targeted test that would have caught this bug.',
+                requireConfirmation: false,
+            },
+        ],
+        execution: DEFAULT_EXECUTION,
+        createdAt: 0,
+        updatedAt: 0,
+    },
+    {
+        id: 'builtin-code-review',
+        type: 'built-in',
+        name: 'Code Review',
+        description: 'Comprehensive code review: security, correctness, performance, and style.',
+        icon: '\u{1F50D}',
+        category: 'Quality',
+        keywords: ['review', 'audit', 'quality', 'security', 'style', 'lint'],
+        variables: [
+            { name: 'target', label: 'Files/Directories to Review', type: 'text', required: true, placeholder: 'src/api/ or specific file path' },
+            { name: 'focus', label: 'Review Focus', type: 'select', required: true, default: 'all', options: ['all', 'security', 'performance', 'correctness', 'style'] },
+        ],
+        steps: [
+            {
+                id: stepId(),
+                name: 'Read and understand',
+                prompt: 'Read through all the code in {{target}}. Understand its purpose, dependencies, and how it fits into the larger codebase.',
+                requireConfirmation: false,
+            },
+            {
+                id: stepId(),
+                name: 'Analyze issues',
+                prompt: 'Review the code in {{target}} with focus on: {{focus}}. Produce a structured review with categories: Critical (must fix), Important (should fix), Minor (nice to fix). For each issue, cite the exact file and line.',
+                requireConfirmation: true,
+            },
+            {
+                id: stepId(),
+                name: 'Fix critical issues',
+                prompt: 'Fix all Critical issues from your review. Make the minimal changes needed. Do not fix Important or Minor issues in this step.',
+                requireConfirmation: true,
+            },
+            {
+                id: stepId(),
+                name: 'Verify fixes',
+                prompt: 'Run tests to verify your fixes did not break anything. Summarize what was changed and what remains as Important/Minor issues for future work.',
+                requireConfirmation: false,
+            },
+        ],
+        execution: DEFAULT_EXECUTION,
+        createdAt: 0,
+        updatedAt: 0,
+    },
+    {
+        id: 'builtin-new-component',
+        type: 'built-in',
+        name: 'New Component',
+        description: 'Create a new UI component with props, styling, and tests following project patterns.',
+        icon: '\u{1F9E9}',
+        category: 'Frontend',
+        keywords: ['component', 'ui', 'react', 'frontend', 'widget', 'create'],
+        variables: [
+            { name: 'component_name', label: 'Component Name', type: 'text', required: true, placeholder: 'UserProfileCard' },
+            { name: 'purpose', label: 'Component Purpose', type: 'text', required: true, placeholder: 'Display user profile with avatar and bio' },
+            { name: 'location', label: 'Directory', type: 'text', required: false, default: 'src/components', placeholder: 'src/components' },
+        ],
+        steps: [
+            {
+                id: stepId(),
+                name: 'Analyze existing components',
+                prompt: 'Look at the existing components in {{location}} and nearby directories. Identify the patterns used for: component structure, props typing, styling approach, and file organization.',
+                requireConfirmation: false,
+            },
+            {
+                id: stepId(),
+                name: 'Create component',
+                prompt: 'Create a new {{component_name}} component in {{location}}. Purpose: {{purpose}}. Follow the existing component patterns. Include: TypeScript props interface, proper styling, and appropriate state management.',
+                requireConfirmation: true,
+            },
+            {
+                id: stepId(),
+                name: 'Add tests',
+                prompt: 'Write tests for the {{component_name}} component. Test: rendering with different props, user interactions, edge cases (empty data, long strings, etc.). Follow existing test patterns.',
+                requireConfirmation: false,
+            },
+            {
+                id: stepId(),
+                name: 'Export and integrate',
+                prompt: 'Add {{component_name}} to the appropriate barrel export file (index.ts) if one exists. Show an example of how to use the component.',
+                requireConfirmation: false,
+            },
+        ],
+        execution: DEFAULT_EXECUTION,
+        createdAt: 0,
+        updatedAt: 0,
+    },
+    {
+        id: 'builtin-refactor-pattern',
+        type: 'built-in',
+        name: 'Refactor Pattern',
+        description: 'Systematically refactor a code pattern across the codebase.',
+        icon: '\u{1F504}',
+        category: 'Refactoring',
+        keywords: ['refactor', 'rename', 'pattern', 'cleanup', 'migrate', 'modernize'],
+        variables: [
+            { name: 'current_pattern', label: 'Current Pattern', type: 'multiline', required: true, placeholder: 'Describe the pattern to replace...' },
+            { name: 'new_pattern', label: 'New Pattern', type: 'multiline', required: true, placeholder: 'Describe the replacement pattern...' },
+            { name: 'scope', label: 'Scope', type: 'text', required: false, default: '.', placeholder: 'Directory or file glob to scope the refactor' },
+        ],
+        steps: [
+            {
+                id: stepId(),
+                name: 'Find all occurrences',
+                prompt: 'Search the codebase within {{scope}} for all occurrences of this pattern:\n\n{{current_pattern}}\n\nList every file and line where this pattern appears. Count total occurrences.',
+                requireConfirmation: false,
+            },
+            {
+                id: stepId(),
+                name: 'Plan refactor',
+                prompt: 'For each occurrence found, plan the transformation to:\n\n{{new_pattern}}\n\nIdentify any edge cases where the transformation needs special handling. List the files that will be changed.',
+                requireConfirmation: true,
+            },
+            {
+                id: stepId(),
+                name: 'Apply refactor',
+                prompt: 'Apply the refactoring to all identified locations. Transform each occurrence of the current pattern to the new pattern. Handle edge cases as planned.',
+                requireConfirmation: true,
+            },
+            {
+                id: stepId(),
+                name: 'Verify refactor',
+                prompt: 'Run all tests. Search for any remaining occurrences of the old pattern to ensure nothing was missed. Fix any test failures caused by the refactor.',
+                requireConfirmation: false,
+            },
+        ],
+        execution: DEFAULT_EXECUTION,
+        createdAt: 0,
+        updatedAt: 0,
+    },
+];

package/dist/main/cli-manager.js

@@ -37,6 +37,15 @@ exports.CLIManager = void 0;
 const pty = __importStar(require("node-pty"));
 const model_detector_1 = require("../shared/model-detector");
 class CLIManager {
+    // Safe log that ignores EPIPE errors (broken pipe when app window closes)
+    safeLog(...args) {
+        try {
+            console.log(...args);
+        }
+        catch {
+            // Ignore EPIPE / broken pipe errors on stdout
+        }
+    }
     constructor(options) {
         this.ptyProcess = null;
         this.outputCallback = null;
@@ -117,17 +126,20 @@ class CLIManager {
                 cleanEnv[key] = value;
             }
         }
+        const ptyEnv = {
+            ...cleanEnv,
+            TERM: 'xterm-256color',
+            COLORTERM: 'truecolor',
+        };
+        if (this.options.enableAgentTeams !== false) {
+            ptyEnv.CLAUDE_CODE_EXPERIMENTAL_AGENT_TEAMS = '1';
+        }
         this.ptyProcess = pty.spawn(shell, [], {
             name: 'xterm-256color',
             cols: 80,
             rows: 24,
             cwd: this.options.workingDirectory,
-            env: {
-                ...cleanEnv,
-                TERM: 'xterm-256color',
-                COLORTERM: 'truecolor',
-                CLAUDE_CODE_EXPERIMENTAL_AGENT_TEAMS: '1',
-            },
+            env: ptyEnv,
         });
         this._isRunning = true;
         this.ptyProcess.onData((data) => {
@@ -161,7 +173,7 @@ class CLIManager {
             this.initialDetectionBuffer += data;
             const result = (0, model_detector_1.detectModelFromOutput)(this.initialDetectionBuffer, true);
             if (result.model) {
-                console.log('[ModelDetect] Phase 1 detected:', result.model, '(bufLen:', this.initialDetectionBuffer.length, ')');
+                this.safeLog('[ModelDetect] Phase 1 detected:', result.model, '(bufLen:', this.initialDetectionBuffer.length, ')');
                 this.currentModel = result.model;
                 if (this.modelChangeCallback) {
                     this.modelChangeCallback(result.model);
@@ -170,7 +182,7 @@ class CLIManager {
                 this.initialDetectionBuffer = '';
             }
             else if (this.initialDetectionBuffer.length > 8192) {
-                console.log('[ModelDetect] Phase 1 gave up after 8KB');
+                this.safeLog('[ModelDetect] Phase 1 gave up after 8KB');
                 this.initialDetectionDone = true;
                 this.initialDetectionBuffer = '';
             }
@@ -184,7 +196,7 @@ class CLIManager {
             }
             const result = (0, model_detector_1.detectModelFromOutput)(this.switchDetectionBuffer, false);
             if (result.model && result.model !== this.currentModel) {
-                console.log('[ModelDetect] Phase 2 detected:', result.model, '(was:', this.currentModel, ')');
+                this.safeLog('[ModelDetect] Phase 2 detected:', result.model, '(was:', this.currentModel, ')');
                 this.currentModel = result.model;
                 this.switchDetectionBuffer = ''; // Reset after successful detection
                 if (this.modelChangeCallback) {