- Add /missions/health endpoint handling to MissionManager test mocks
- Add listMissionsWithSummaries to mission-e2e test mock
- Add planState to Slice type and mock factories
- Add stuckKillCount to retry task test assertions
- Update log message for stuck-killed retry
- Added MeshTopology.tsx with SVG visualization
- Nodes arranged in circle with local node at center
- Color-coded by status
- Added CSS for mesh topology styles
- Add reconcileAllMissionFeatures() method to sync task status with mission features
- Refresh selected mission detail when feature:updated SSE event fires
- Add unit tests for reconcileAllMissionFeatures and SSE handler
- Fix async getTask call in reconcileAllMissionFeatures (keep await)
- DELETE /api/nodes/:id now returns 204 No Content
- GET /api/nodes/:id/metrics returns systemMetrics from node
- Added test for GET /api/mesh/state endpoint
- Modified POST /api/nodes to make type optional (defaults to 'remote')
- Changed DELETE /api/nodes/:id to return 204 No Content
- Updated GET /api/nodes/:id/metrics to return SystemMetrics from node.systemMetrics
- Added GET /api/mesh/state route for full mesh topology state
- Add agent rating API routes (GET/PATCH /api/agents/:id/ratings)
- Add API helper functions for fetching and updating agent ratings
- Add Performance tab UI to AgentDetailView with star rating display
- Add CSS styling for the Performance tab and star rating components
- Add comprehensive route tests for agent ratings endpoints
- Add API routes for agent reflections (GET, POST, DELETE)
- Create AgentReflectionsTab component with reflection list and form
- Integrate reflections tab into AgentDetailView
- Add CSS styles for the reflections tab UI
- Write tests for API routes and component
- Remove separate Execution section from settings sidebar
- Move step execution settings (runStepsInNewSessions, maxParallelSteps) into Scheduling section
- Consolidate related runtime behavior settings for better UX
- Update tests to navigate to Scheduling section instead of Execution
- Add chat system type definitions (ChatSession, ChatMessage, ChatMessageRole)
- Add SQLite schema migration for chat_sessions and chat_messages tables
- Implement ChatStore with full CRUD operations for sessions and messages
- Export ChatStore and types from @fusion/core public API
- Add comprehensive test suite for ChatStore with session/message operations
- Update schema version expectations in existing tests
- Add FTS5 virtual table with v21 database migration for task search
- Add searchTasks() method to TaskStore with FTS5 query support
- Add q= search parameter to GET /api/tasks route for server-side search
- Update useTasks hook and frontend API to support searchQuery prop
- Update Board.tsx and App.tsx to pass searchQuery through component hierarchy
- Add comprehensive tests for FTS5 index and searchTasks functionality
- Apply consistent agent-active state styling across ListView and BoardView
- Add .agent-active CSS class to unify visual indicator styling
- Update ListView component to use unified active state classes
- Add ListView test coverage for agent-active state rendering
The cachedMaxConcurrent variable driving the AgentSemaphore was only
updated during unpause transitions and merge retry timers. Changing
maxConcurrent in the dashboard had no effect on the semaphore until one
of those events fired, causing fewer tasks to run than configured.
Also adds HeartbeatMonitor and HeartbeatTriggerScheduler wiring to both
serve and dashboard commands, with tests.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- Add mobile responsive CSS rules for ProjectOverview component
- Add mobile breakpoint styles for card layouts and navigation elements
- Add ProjectOverview mobile structure tests for responsive behavior
- Add ProjectCard mobile visibility tests
- Add ember theme to COLOR_THEMES array in types.ts
- Add ember theme CSS variables in styles.css
- Add ember theme swatch preview in ThemeSelector component
- Add ember to validThemes validation in index.html
- Add ThemeSelector tests for ember theme
- Add 8 new theme IDs to COLOR_THEMES in core types
- Add CSS blocks with color swatches for all 8 new themes
- Update ThemeSelector to display new theme options
- Register themes in index.html validThemes array
- Add comprehensive tests for new theme support
- Include changeset for @gsxdsm/fusion package
- Remove column-header min-height from mobile media block for tighter layout
- Reduce TaskCard touch targets in mobile view for better mobile UX
- Update board-mobile tests to reflect new mobile styling behavior
- Create FusionProvider component that initializes TaskStore and provides it via React context
- Add useFusion() hook for accessing TaskStore and project path from any child component
- Implement detectProjectDir() utility that walks up directory tree looking for .fusion/fusion.db
- Update package entry point with context exports and demo app
- Add @fusion/core as dependency to TUI package
- Add comprehensive tests for FusionProvider, useFusion hook, and detectProjectDir
- Add API documentation to README with usage examples
- Add soul and instructionsText fields to all 20 built-in agent presets
- Enable soul-based agent visualization and instructions preview in NewAgentDialog
- Add comprehensive tests for preset soul and instructionsText rendering flow
- Include changeset for @gsxdsm/fusion minor release
The store's updateTask treats `undefined` as "no change" and `null` as
"clear field". Six places in executor.ts passed `undefined` when intending
to clear worktree/branch, leaving stale references that caused tasks to
get stuck in in-progress with no active session after stuck-kills, pauses,
and transient error retries.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Previously, recoverOrphanedExecutions() skipped tasks whose worktree
directory still existed, assuming an active session. After engine crashes
where resumeOrphaned() failed, these tasks were stuck forever. Now uses
a tiered grace period: 60s for missing worktrees, 5min for existing ones
to avoid racing with startup recovery.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- Remove min-height/min-width 44px enforcement from FileBrowser styles.css that was overriding native sizing
- Update FileBrowser test to match the actual CSS without mobile touch target rules
- Update memory log with the change rationale
- Reduce column count badge height by removing min-height override on mobile
- Add explicit compact sizing for column-header icon buttons (archive, etc.)
- Ensure consistent 28px sizing for tappable elements on mobile viewports
- Add overflow: hidden and text-overflow: ellipsis to mobile modal header to prevent title from overlapping close button
- Add min-width: 0 and flex-shrink: 1 to header title for proper flex truncation
- Add CSS rules for FileBrowserModal breadcrumb and filename truncation on narrow screens
- Add tests for mobile close button visibility and CSS text truncation behavior
- Move model selector from sticky footer into scrollable content area
- Slim the footer to only contain the Start Planning button
- Add .planning-model-select-group CSS for proper flexbox layout
- Center footer button and model-select row for better mobile UX
- Remove inline styles from model selector, use CSS classes instead
- Add expandable scripts submenu inside the More sheet on MobileNavBar
- Include script status indicators, run/stop buttons, and navigation
- Add CSS styles for submenu animation, hover states, and mobile touch targets
- Add comprehensive tests for submenu toggle, script actions, and edge cases
- Reduce mobile media query button heights from 44px to 36px across all dashboard components
- Update mobile CSS tests (nav bar, board, modals, mission planning) to expect 36px touch targets
- Add CSS testing patterns to project memory for consistent test guidelines
- Add isolate: true to vitest configs for safe parallel test execution
- Simplify engine test script from 7 sequential runs to single parallel run
- Fix TypeScript type errors in Column, Board, WorktreeGroup, and App components
- Fix board-mobile test to match current TaskCard tap behavior
- Relax App deep-link test to handle React Strict Mode behavior
- Remove autoFocus attribute from agent name input to prevent unexpected focus behavior
- Make name field non-blocking so preset flow is not interrupted by focus steal
- Add comprehensive tests for NewAgentDialog covering preset flow and name field behavior
- Move settings icon placement from its previous position to directly right of the stop button in the Header component
- Update Header component layout to reflect the new icon ordering
- Update Header tests to match the new element ordering
- Replace hardcoded colors with CSS custom properties (theme tokens) across GitHub import components
- Add accent-color to radio buttons using --in-progress token
- Add explicit light theme overrides via [data-theme="light"] selectors for tabs, toolbar, issues list, preview, labels, and state panels
- Style filter input with theme-aware border, background, focus ring, and placeholder colors
- Fix form-error background to use color-mix instead of hardcoded rgba
- Update useModalManager to accept Task objects for immediate modal display before full detail loads
- Refactor TaskDetailModal to render with basic Task data and load TaskDetail asynchronously
- Update TaskCard click handlers to open modal immediately with optimistic data
- Update ListView click handler for same optimistic opening behavior
- Update AppModals types to support Task | TaskDetail union
- Add CSS for modal loading/skeleton states
- Add comprehensive tests for TaskDetailModal and useModalManager optimistic flow
- Simplify TaskCard and ListView tests to reflect new optimistic opening pattern
When context overflow or compaction causes an agent to lose awareness of
the task_done tool, the executor now checks if all steps are complete
before failing — treating it as an implicit task_done. Also adds
self-healing recovery for tasks that already slipped through as
misclassified failures in in-review.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Reduce min-height and padding on done column cards for mobile breakpoints
- Update board-mobile test expectations to match new compact sizing
- Prevents excessive vertical stretching of completed task cards on small screens