- Add desktop CSS constraints so column headers and titles can shrink without expanding board columns
- Prevent auto-merge toggle controls from shrinking to preserve stable header layout
- Add board CSS regression tests for the 6-column minmax grid template and header truncation rules
Update the version-hash reload loader message to "Updating to a new
frontend version..." so users understand it refers to the dashboard UI,
not a server/runtime upgrade.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Review level select now treats the empty option as undefined (auto — let
triage decide) rather than coercing to 0/None, and relabels the option
accordingly. Widens onReviewLevelChange to accept undefined.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Done task cards now show agent execution time (sum of workflow step
durations) instead of wallclock time from creation to completion, matching
the Workflow runtime metric in the stats tab. Falls back to the previous
wallclock duration when no workflow timing data is available. Formatter
also updated to render sub-minute durations in seconds (e.g., 45.3s, 1m 30s)
to match the stats panel format.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
- Rework TaskDetailModal move controls to use split-button primary/secondary transitions with in-review-specific options
- Add dedicated split-button styling for divider, chevron menu, focus states, and mobile sizing using design tokens
- Keep in-review merge action visible while move options remain accessible through secondary menu actions
- Expand TaskDetailModal tests to cover split-button rendering, primary click behavior, and secondary transition menu contents
- Add panel resize state and pointer-driven drag handling to QuickChatFAB
- Wire resize behavior into component layout so desktop users can adjust panel dimensions
- Style resize handles and interaction affordances in QuickChatFAB.css
- Expand QuickChatFAB tests to cover resize lifecycle, constraints, and persistence behavior
When the dashboard reloads because the build version hash changed, surface that
on the loading screen instead of the generic "Initializing dashboard..." so the
user understands why they bounced.
Carry the signal across the hard reload via a sessionStorage flag set in
versionCheck just before reloadOnce(), and consume-and-clear it once on the
loader's first render.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
- Update .agent-log-return-to-live to anchor on the right edge while staying sticky at the bottom
- Switch the control container to flex with fit-content width to keep compact button sizing
- Add mobile-specific right offset so spacing stays consistent at smaller breakpoints
- Add plugin runtimes endpoint in dashboard routes and legacy API client support
- Add runtime mode controls to NewAgentDialog with plugin runtime selection UX
- Add runtime mode display and editing support in AgentDetailView with matching styles
- Expand dashboard tests to cover new routes, modal behavior, and mobile agent view flows
- Restructure ScheduledTasksModal header markup to support updated title and actions grouping
- Add automation toolbar styling updates in ScriptsModal.css to match the new modal header layout
- Update ScheduledTasksModal tests to assert the revised header structure and action placement
The X button in the footer's Background Tasks popover silently failed when
ai_sessions.lockedByTab pointed to a stale/closed tab: cancelPlanning et al.
returned 409, dismissSession set lockConflict and bailed before calling
deleteAiSession. Always proceed to delete; the dismissal tombstone already
prevents stale SSE/sync from resurrecting the session.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
- Remove the per-row eye toggle for already hidden usage windows so hidden rows no longer show a show/hide button
- Keep hide controls visible for active rows and preserve existing hide interactions
- Update UsageIndicator tests to assert hidden rows omit hide buttons and visible rows still render hide controls
- Drop obsolete per-row unhide test now that restoring hidden windows is handled through the provider-level show hidden flow
- Apply optimistic start/stop state transitions in AgentsView and AgentDetailView action handlers
- Add optimistic behavior to AgentListModal while preserving rollback on API failure
- Expand unit test coverage for optimistic transitions and failure recovery in all three agent views
- Update dashboard TUI test expectation to align with the new agent state update flow
- Move preview panel markup to support a two-column desktop structure
- Add desktop-specific grid styling for configuration and preview panes
- Preserve existing mobile stacking behavior with responsive CSS overrides
- Track and thread usage trigger bounds from Header through modal management to UsageIndicator
- Render desktop usage details as an anchored popover while preserving existing modal wiring behavior
- Extract usage indicator styling into a dedicated UsageIndicator.css and remove planning-mode style duplication
- Expand header and usage indicator tests to cover anchor rect propagation and popover behavior
- Stabilize CLI bundle output sqlite assertion for workspace test reliability
- Replace hover background fallbacks with direct var(--surface-hover) in QuickChatFAB and SettingsModal styles
- Tighten status-colors theme test to forbid var(--surface-hover, ...) fallback syntax
- Preserve token-based hover styling contract while preventing stale fallback patterns from reappearing
- Enforce terminal xterm root height fill so inline fit heights do not collapse the modal terminal area
- Defer font-size-driven fit to the next animation frame and coalesce pending fits via pendingFitRef
- Remove eager refit calls from zoom keyboard/button handlers and rely on shared font-size effect scheduling
- Expand TerminalModal tests to assert font-size controls and keyboard zoom continue to trigger xterm refits
- Keep the view overflow trigger available when Todos is supported so the Todos view remains reachable
- Update TodoView list rows to use an explicit select button and align active-state assertions with the new structure
- Restore keyboard focus styling and mobile action visibility in TodoView CSS for accessibility and usability
- Harden CLI extension test cleanup with retry logic for transient ENOTEMPTY/EBUSY tempdir removal errors
- Replace hardcoded mobile footer offset with --executor-footer-height in MobileNavBar content padding
- Move standalone display-mode behavior to a :root token override and remove global #root standalone padding
- Scope standalone bottom-gap application to mobile layout containers (project content and executor status bar)
- Update mobile nav and PWA CSS tests to assert tokenized spacing and scoped standalone rules
- Render a compact token usage indicator in TaskCard footer with accessible labeling and token-aware styling
- Track token usage fields in the TaskCard memo comparator and expose a comparator test helper for regression coverage
- Add TaskCard tests for token usage rendering behavior and comparator invalidation on token usage updates
- Configure runtime plugin Vitest setups with an @fusion/engine source alias for reliable workspace test resolution
- Keep restart integration child_process spawn mocking aligned with execSync-driven merge verification behavior
Extends the inMemoryDb opt-in established in 6e4797ff1 to single-instance
TaskStore/AgentStore/RoutineStore/PluginStore tests that were still
opening disk-backed fusion.db files. Cross-instance persistence tests
(open store A, close, open store B on same dir) and migration tests that
seed via a sibling Database instance keep their disk-backed stores —
swapping those would silently drop data between instances.
Sites flipped:
- core: store.test.ts (RunMutationContext, memory-toggle, diagnostics
blocks), store-sort, settings-export, backup, plugin-loader,
agent-instructions, agent-instructions-bundle, mission-store (all 10
triage subtests), mission-planning-context.integration
- dashboard: routes.test.ts (Messaging routes block),
session-reconnect, session-cross-tab, planning
Sites left disk-backed (cross-instance or sibling-Database dependency):
- run-audit*, task-documents, fts5-guard (sibling Database at same dir)
- mission-integration, mission-factory-parity (taskStore2 reopens)
- agent-store checkout-leasing (TaskStore + AgentStore at same dir)
- routes.test.ts AgentStore seed pattern (route handler opens its own)
- cli/extension.test.ts (makeCtx opens its own TaskStore)
All disk-backed tests continue to write to mkdtemp temp dirs — no live-db
risk introduced. Wall-clock impact: core 149s → 19s.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
The auto-resolved FN-2614 merge swapped `import * as engine` for a narrow
named import but left dangling `engine.X` references, breaking the
dashboard build. Restore the namespace import and drop the now-orphaned
engineExports/ntfy helper consts.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Manual version bump to 0.5.0 (changeset version produced 1.0.0 from a
single minor changeset against 0.4.1; that release was rolled back and
the 1.0.0 npm version deprecated).
Aggregates: status terminology refresh (planning/replan), Reviewer
rename, in-review pause behavior, dashboard-tui resize hardening,
dev-server experimental toggle fix, version reporting fix.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Move the xterm.onData -> sendInput wiring (and the window resize listener)
from a separate post-init effect into initTerminal itself, so they share
the xterm instance's lifetime. Under StrictMode + Vite Fast Refresh the
separate effect could re-run and attach a second listener to the same
live xterm instance, producing per-character input doubling (every
keystroke -> two pty.write calls -> shell echoes "aabbcc"). The handler
now reads sendInput via a ref, so function-identity changes no longer
require re-binding. Resize listener is removed at every xterm disposal
site (tab switch, modal close, session-invalid replace, reinitialize).
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
The 5-minute exec timeout only killed the immediate shell, leaving
vitest/pnpm worker trees alive. Across retries these accumulated and
thrashed the host, starving the engine and TUI. Switch verification to
spawn-based runner with detached process group so timeouts SIGTERM the
whole tree (SIGKILL after 5s grace), and bump the wallclock to 10m for
larger workspaces. Stream-truncate output instead of relying on ENOBUFS.
Also fix two flaky/race-prone dashboard tests that were red on main and
blocking every in-review task at merge verification.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
- fix(FN-2613): restore green CI after static engine import changes
- test(FN-2613): complete Step 3 — update mobile and standalone CSS assertions
- feat(FN-2613): complete Step 2 — move standalone token override to global styles
Resolves "Failed to load PTY module" install errors on Linux/macOS by
aliasing node-pty to the homebridge fork, which ships prebuilds for
linux x64/arm64/arm/ia32 across many Node ABIs and uses prebuild-install
for darwin/windows binaries on install.
- Aliased dep so all "node-pty" import specifiers (and vi.mock calls)
keep working unchanged.
- Removed darwin chmod postinstall hack (fork handles permissions).
- Updated cli/build.ts to dynamically resolve node-pty install root and
pick prebuilds by ABI for Linux cross-compile; warn-and-skip for
darwin/windows cross-compile (host-only there, as before).
- Added type shim because the fork's bundled typings declare module
'@homebridge/node-pty-prebuilt-multiarch', not 'node-pty'.
Verified: pnpm install clean, dashboard typecheck clean, native module
loads and spawns shell via fork, host + linux-x64 cross-compile staging
both produce dist/runtime/<plat>/pty.node.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
The dashboard modules used a variable-specifier dynamic import
(`const m = "@fusion/engine"; await import(m)`) to defeat bundler static
analysis. tsup honored that and left the dynamic import in dist/bin.js,
so the published `@runfusion/fusion` package failed at runtime with
"createFnAgent2 is not a function" — `@fusion/engine` isn't on npm and
the silent catch set the binding to undefined. Replaces the trick with
static imports across planning, chat, subtask-breakdown, mission-interview,
agent-generation, ai-refine, roadmap-suggestions, milestone-slice-interview,
and routes. Core can't statically import engine (cycle), so it now exposes
setCreateFnAgent and engine wires itself in at module load. Documents the
pattern in AGENTS.md.
FixesRunfusion/Fusion#9.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
