- Update AGENTS.md with authoritative run lifecycle semantics documentation
- Add API-level regression test for repeated manual run prevention
- Update run lifecycle tests to match corrected behavior
- Refactor AgentStore to use structured run records as the authoritative source for run state
- Improve run status queries with better filtering and ordering
store.logEntry, assignTask, and checkoutTask now receive an extra
runContext argument. Updated test assertions in executor.test.ts,
agent-heartbeat.test.ts, and restart.integration.test.ts to match
the current call signatures using expect.objectContaining where
appropriate.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
In step-session mode each step is registered with StuckTaskDetector
under a compound key (e.g. "FN-1452-step-1") rather than the bare
task ID. When stuck detection fired, event.taskId was that compound
key, breaking three things simultaneously:
1. beforeRequeue("FN-1452-step-1") → store.getTask() threw (no such
task) → stuckKillCount never incremented on the real task.
2. markStuckAborted("FN-1452-step-1") → activeStepExecutors.get()
returned undefined (keyed by "FN-1452") → terminateAllSessions()
never called, other step sessions kept running.
3. stuckAborted.set("FN-1452-step-1") → executor checks
stuckAborted.has("FN-1452") → miss → stuckRequeue never set
→ task never moved to todo.
Fix: add a canonicalTaskId field to TrackedTask (defaults to the
tracking key for single-session mode where they are identical).
StepSessionExecutor now passes taskDetail.id as the third arg to
trackTask(). killAndRetry() uses entry.canonicalTaskId for all
external callbacks so they always resolve to the real task ID.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Unify slice activation and auto-triage semantics for mission progression
- Align engine progression with stale recovery logic for active missions
- Fix scheduler delegation check to use feature.missionId instead of deprecated field
- Add integration tests for stale mission recovery scenarios
- Fix mission API recovery gaps for active missions (activate on first non-done slice)
- Update README.md autopilot documentation section
- Add CSS classes in styles.css for planning and breakdown modal styling
- Refactor MissionInterviewModal to use CSS classes instead of inline styles (164 lines reduced)
- Refactor PlanningModeModal to use CSS classes (67 lines reduced)
- Refactor SubtaskBreakdownModal to use CSS classes (26 lines reduced)
- Improve maintainability and consistency of modal styling
- Add run-audit event types and SQLite schema migration
- Implement typed store read/write/query APIs for run audit
- Make audit writes atomic with task updates (transactional writes)
- Fix SQLite parameter type casting in db layer
- Add comprehensive tests for RunMutationContext and audit functionality
- Replace text input with always-open dropdown in MessageComposer component
- Update MessageComposer tests to cover dropdown behavior
- Remove .message-composer-input from shared CSS rule
- Add semantic CSS custom property tokens to :root for theme-aware colors
- Replace hardcoded colors in MailboxModal with semantic tokens
- Replace hardcoded colors in mission components with semantic tokens
- Replace hardcoded colors in terminal and model combobox with semantic tokens
- Add light theme overrides for new semantic tokens
- Add MailboxModal CSS regression tests for theme awareness
- Document semantic token naming convention in memory
- Add MemoryInsights class in @fusion/core for AI-powered memory audit generation
- Add post-run hook to CronRunner for triggering memory summarization after scheduled tasks
- Wire memory background processing in both dashboard and serve commands
- Add memoryAuditEnabled and memoryAuditSchedule settings for configurable automation
- Fix startup ordering: sync automation before cronRunner.start() to prevent race conditions
- Add comprehensive tests for memory-insights and dashboard/serve integration
- Update contributing.md and settings-reference.md with documentation
- Fix mission autopilot slice activation when tasks complete
- Fix state refresh in MissionManager to properly reflect autopilot status
- Fix scheduler to correctly trigger autopilot progression events
- Add unit tests for mission store autopilot state transitions
- Add component tests for MissionManager autopilot toggle and state display
- Add e2e tests for mission autopilot lifecycle (enable → task completion → slice progression)
- Update documentation with autopilot state machine details
- Add rust, copper, foundry, carbon theme IDs to core types
- Register themes in ThemeSelector component and flash-prevention logic
- Add CSS theme blocks and swatches for all four new themes
- Add ThemeSelector unit tests for the new themes
- Update memory documentation with new theme count
- Convert header search box to floating dropdown positioned below header
- Add mobile header controls test for floating search behavior
- Add Header component tests for search box rendering
- Update Header.tsx to render search as floating element with absolute positioning
- Add styles for floating search box with backdrop blur and shadow
- Add regression tests for TaskForm focus management
- Test that description textarea receives focus on mount when empty
- Test that focus stays in place when toggling controls panel
- Test that Escape key collapses expanded panel without losing focus
- Test focus behavior with pre-filled description
- Reduce terminal tab heights from 48px to 36px on mobile
- Adjust terminal button sizes from 36px to 32px
- Optimize padding and font sizes for mobile viewports
- Refactor beginBreakdown() to accept explicit description parameter
- Update auto-start effect to pass current initialDescription directly
- Pass description as broadcastTitle to connectToSubtaskStream for correct background activity title
- Preserves existing send-to-background and cancel semantics
- Wrap .btn-icon hover styles in @media (hover: hover) to prevent touch activation on mobile/touch devices
- Add .btn-icon:active styles for proper pressed state feedback
- Fixes unintended hover state activation on touch interactions
- Add getAppVersion() utility that walks up directories to find package.json
- Add parseSemver() for semver string parsing with major/minor/patch components
- Add CentralCore version sync methods (getAppVersion, syncVersion, getLastSyncTime, getSyncStatus)
- Add CentralCore events for version sync lifecycle (version-sync-started, version-sync-completed, version-sync-failed)
- Add schema v4 migration with appVersion and lastSyncTime columns to projects table
- Export new types (CentralSyncStatus, SyncResult) and utilities via @fusion/core
- Update memory.md documentation with new types
- Fix TypeScript error in app-version.ts (return pkg.version directly instead of cached variable)
- Fix mobile layout issues in activity log modal
- Update theme colors for better contrast and consistency
- Improve responsive behavior on smaller screens
Squash merge of fusion/fn-1431. The changes were already integrated into main through other commits; this commit formally records the merge.
- Add robust timestamp parsing in getLastSessionResetTimestamp()
- Support multiple date formats (ISO-8601, Unix timestamps in seconds and milliseconds)
- Add fallback chain: environment variable → Claude config file → current date
- Parse Unix timestamps (both seconds and milliseconds) from config file
- Add comprehensive tests for timestamp parsing edge cases
- Add Budget Settings section to agent ConfigTab in AgentDetailView
- Section includes Token Budget, Usage Threshold, Budget Period, and Reset Day controls
- Budget settings are stored in agent.runtimeConfig.budgetConfig and persisted via PATCH /api/agents/:id
- Add comprehensive tests for Budget Settings UI interactions
- Document Budget Settings in AGENTS.md
- Increase heartbeat test timeout from 150ms to 250ms for CI environments
- Increase in-process-runtime test timeouts to handle slower CI execution
- Prevents flaky test failures in slow CI runners
- Remove unnecessary autopilotLoading state variable from MissionManager
- Simplify autopilot toggle by relying directly on autopilotEnabled prop
- Eliminate intermediate loading state that added complexity without benefit
- Replace hardcoded purple-tinted --bg-secondary (#1a1a2e) and --bg-tertiary (#252536) with dynamic color-mix values
- Use --surface and --card blending for better visual consistency with the dark theme palette
- Ensure background layers harmonize with the existing surface/card design tokens
- Add Kimi logo to ProviderIcon component for visual identification
- Add Kimi provider to usage mapping for cost tracking
- Add Kimi usage fetcher to fetch cost and usage data from Kimi API
- Add ProviderIcon tests for Kimi logo verification
- Add UsageIndicator tests for Kimi provider
- Add comprehensive usage tests for Kimi fetcher
- Remove internal search filter from ListView, use searchQuery prop instead
- Show header search bar in list view for consistent filtering
- Remove obsolete list-filter styles (63 lines removed)
- Update tests to reflect new searchQuery prop behavior
- Add RunMutationContext type to track which agent run caused a mutation
- Thread runContext through TaskStore.logEntry, addComment, addSteeringComment, and pauseTask
- Propagate runContext from HeartbeatMonitor.executeHeartbeat to task store operations
- Propagate runContext from TaskExecutor.execute to task store operations
- Add GET /api/agents/:id/runs/:runId/mutations endpoint to query mutations by runId
- Add createTaskLogToolWithContext for heartbeat tools with run context support
- Add comprehensive tests for RunMutationContext across store and heartbeat modules
- Update memory.md with RunMutationContext usage convention
- Deprecate autoAdvance field in favor of autopilotEnabled as the sole control
- Remove autoAdvance guard logic from MissionAutopilot engine class
- Simplify MissionManager UI to use single autopilot toggle with visual state indicator
- Update MissionAutopilot tests to use autopilotEnabled instead of autoAdvance
- Update MissionManager component tests for simplified UI
- Update AGENTS.md documentation to reflect the simplified autopilot model
- Fix usage dialog to show Claude session reset time when session is paused
- Add unit tests for reset time display logic in usage calculations
- Update usage summary and history components to include reset time
- Add height: 100dvh and max-height: 100dvh to new-task-modal for mobile viewport support
- Fixes dialog not filling the mobile viewport height on smaller screens
- Replace 🌐/📁 emoji in SettingsModal sidebar with Globe/FileText Lucide icons
- Update SettingsModal tests to expect icon-based scope markers instead of emoji
- Use consistent iconography for visual hierarchy between global and project settings
- Add Maximize2/Minimize2 toggle button to QuickEntryBox for fullscreen textarea expansion
- Add Maximize2/Minimize2 toggle button to InlineCreateCard for fullscreen textarea expansion
- Implement single-textarea pattern with positioned container for expand button overlay
- Add description--fullscreen CSS class for fullscreen textarea styling
- Add focus handoff when entering fullscreen mode
- Add comprehensive unit tests for fullscreen expansion in both components
- Tests cover collapsed/expanded states, toggle button visibility, and class application
- Add core proxy API infrastructure (api-node.ts) with remote node communication
- Create NodeContext for centralized node state management
- Implement useNodeProxy hook for node operations (metrics, events, health)
- Implement useRemoteNodeData hook with 10-second polling for live metrics
- Implement useRemoteNodeEvents hook for real-time SSE event subscription
- Add NodeStatusIndicator component with animated status indicators
- Add comprehensive tests for all new hooks and components
- Update dashboard styles for node status visualization
