Merged four step commits bringing AgentDetailView improvements with model favoritism UI, bundled plugin install wiring for the droid runtime, aligned documentation between settings reference and the plugin README, plus mobile layout refinements with safe-area padding test updates.
Fusion-Task-Id: FN-3261
Merges FN-3306 agent detail favorites wiring and test coverage, FN-3244 mobile FAB drag path hardening, FN-3303 mobile split layout updates, FN-3366 CI shard contract documentation with a new `scripts/ci-test-shard.mjs`, and FN-3014 research recovery semantics — plus expanded research extension tool
Fusion-Task-Id: FN-3306
Merges FN-3244's three-step hardening of the mobile FAB drag path, adding regression tests and finalizing TypeScript typing. Includes FN-3296's tokenized typography for agent health reasons, FN-3014's documented recovery semantics for the research system, and related CLI/dashboard work that improves
Fusion-Task-Id: FN-3244
Documents research recovery semantics in the CLI reference and settings reference, adding 9 lines of documentation across both files.
Fusion-Task-Id: FN-3014
Fixed health reason display in the Agent Detail view by tokenizing its typography to align with the design system's font token standards, updating the component's CSS and TypeScript files.
Fusion-Task-Id: FN-3296
Removed fallback logic for task status colors in TaskDetailModal, simplifying the component by relying on proper CSS token fallbacks instead of inline JavaScript defaults.
Fusion-Task-Id: FN-3270
This merge consolidates several improvements across the dashboard, engine, and CLI: mobile nav now extends into iOS safe area, done columns sort by recency with merge-active state preserved, agent views gain collapsible error displays, per-project HeartbeatMonitor isolation is fixed for multi-projec
Fusion-Task-Id: FN-3196
- Add a collapsible agent error display in agent views, including modal integration, styling updates, and larger touch targets for action controls
- Expand dashboard test coverage for agent error rendering/collapse behavior plus board/list/task card sorting and merge-active state regressions
- Fix skill name/path resolution and align Fusion two-segment skill naming with bare pi-coding-agent skill names
- Harden runtime and route behavior across engine/dashboard/droid flows with added regression and e2e test coverage
Fusion-Task-Id: FN-3342
Icons in the installed PWA were overlapped by the iOS home indicator and a
body-colored gap appeared between the nav bar and the executor status bar
(whose `bottom` already includes safe-area-inset-bottom + standalone-bottom-gap).
Adding matching `padding-bottom` to `.mobile-nav-bar` lifts icons above the
indicator and closes the gap.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
This merge adds done-column sorting by most recent completion while preserving merge-active state on verification bounces, introduces a comprehensive droid runtime regression test suite covering engine delegation, auth routes, and model routing, hardens the CLI native bundle externalization, and fix
Fusion-Task-Id: FN-3074
- Sort Done column and list views by most recent completion timestamp to keep recent work visible first
- Update dashboard and task management docs to document done-column recency ordering behavior
- Preserve merge-active task state across verification bounce paths and board/list rendering updates
- Forward removeDependencyReferences in deleteTask flows and treat finish_reason=repeat as a soft-stop in engine executor
Fusion-Task-Id: FN-3364
This merge lands FN-3231 across two steps: it preserves a merge-active fix when verification bounces occur (step 1) and ensures the fix is retained during board routing transitions (step 2). Changes span the dashboard Board routing logic and the engine executor, with corresponding test coverage adde
Fusion-Task-Id: FN-3231
useTasks.deleteTask had a single-arg signature and dropped the options
object, so the modal's "remove dependency references" confirmation
flag never reached the API and deletion still failed with
TASK_HAS_DEPENDENTS. Forward options through the hook and widen the
matching taskOperations.deleteTask type in AppModals.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Moonshot/Kimi returns a non-standard finish_reason of "repeat" when its
server-side repetition detector trips. pi-ai surfaces this as a fatal
session state error, killing the agent heartbeat. Filter that case in
promptSessionAndCheck so the truncated turn is logged and the heartbeat
continues on the next tick.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
The merge adds Step 2 of the native externalization hardening for the CLI package, including a minor changeset for `@runfusion/fusion` and tests validating bundle output and package configuration behavior. The changes strengthen the contract governing how native modules are handled during the build
Fusion-Task-Id: FN-3262
Fixes FN-3112: Updates NodesView CSS to use tokenized border-radius and transition values from the design system instead of hardcoded pixel values, ensuring consistency with the dashboard's established token variables.
Fusion-Task-Id: FN-3112
Merges three commits: tokenizes MissionManager CSS spacing and colors to align with the design system, fixes a mobile viewport teardown bug when closing planning mode, and resolves token and focus regressions in the task detail modal. Also adds a new test for the planning mode viewport fix.
Fusion-Task-Id: FN-3073
Merges FN-3237 to stabilize the mobile viewport teardown flow when closing the planning modal, ensuring the viewport state resets correctly. Adds test coverage for the mobile planning close path and includes a changeset for the patch release.
Fusion-Task-Id: FN-3237
The merge completes the CSS refactoring of MissionManager by replacing hardcoded color literals with CSS custom properties and spacing values with design tokens, bringing the component in line with the dashboard's design system. A test was also updated to stabilize the mission activity load-more ass
Fusion-Task-Id: FN-3138
This merge addresses FN-3072 design feedback by updating CSS design tokens and mobile touch targets in ModelOnboardingModal, PluginSlot, and SettingsModal components, while also adding tests for droid settings plugin integration states. FN-3137 refines MissionManager's back-button navigation test as
Fusion-Task-Id: FN-3072
Refined the back-button navigation tests in MissionManager by consolidating assertions and removing redundant test cases, reducing the test file by 19 lines while maintaining coverage of the navigation behavior.
Fusion-Task-Id: FN-3137
This merge introduces a droid runtime plugin system with event bridge and process management (FN-3228), stabilizes plugin workspace verification gates and suppresses placeholder plugin shells in onboarding/settings flows (FN-3071), tokenizes danger/error styling in task detail components (FN-3350),
Fusion-Task-Id: FN-3071
This merge brings three major improvements: a new `fusion-plugin-droid-runtime` plugin that provides event bridging, process management, MCP schema serving, and tool mapping for spawning AI agents; comprehensive UX polish across the dashboard including tokenized danger/error styling in task details,
Fusion-Task-Id: FN-3228
Fixes task detail modal danger/error styling by switching to token-based CSS variables for consistent theming across dark and light modes.
Fusion-Task-Id: FN-3350
The merge delivers a new `fusion-plugin-droid-runtime` plugin providing a full MCP server and runtime adapter for droid-based agents, with event bridging, process management, tool mapping, and a prompt builder. It also adds agent delegation and org hierarchy tools to the pi extension, while fixing a
Fusion-Task-Id: FN-3332
This merge adds agent delegation tools (list_agents, delegate_task) and org hierarchy tools to the pi extension, along with corresponding skill documentation and tests; it also cleans up AgentDetailView by removing its dedicated CSS and reducing the component, collapses agent tool output in logs by
Fusion-Task-Id: FN-3355
Fixes deep-link close to preserve browser history state and adds comprehensive integration tests for the navigation history feature, covering App component and useDeepLink hook behavior.
Fusion-Task-Id: FN-3344
Adds `list_agents` and `delegate_task` tools to the pi extension for inter-agent delegation, along with supporting reference docs and 248 new tests. The sync script was also updated to keep the skill tools in sync with the extension implementation.
Fusion-Task-Id: FN-3327
Merged fix for FN-3354 completing Step 6 of verification gate improvements. Changes adjust styling in the Agent Detail view CSS and update a corresponding test assertion in the Model Onboarding modal, suggesting the verification gate behavior or visual presentation was refined.
Fusion-Task-Id: FN-3354
This merge delivers v0.17.2 and ships FN-3343 keyboard handling across four dashboard modals (NewTaskModal, SubtaskBreakdownModal, MilestoneSliceInterviewModal, SettingsModal) with mobile CSS overrides, plus keyboard style regression tests for each modal and a minor test alignment for the onboarding
Fusion-Task-Id: FN-3343
Merges v0.17.2 with two features: updated board ordering comparator (FN-3257) now covers legacy priority normalization with new tests, and an in-review retry mechanism (FN-3349) with delivery docs. All package changelogs updated and version-bumped.
Fusion-Task-Id: FN-3257
This merge adds a changeset for feature FN-3349 (in-review retry behavior) and delivery documentation to the `@runfusion/fusion` package.
Fusion-Task-Id: FN-3349
When running multiple projects (e.g., primary + fitness-app), on-demand
agent heartbeat triggers from the dashboard silently created zombie run
records that never executed. The agent appeared "running" but produced
no logs, no system prompt, no execution prompt.
Root cause: the dashboard routes used a single HeartbeatMonitor bound to
whichever project initialized first. For agents in other projects,
isHeartbeatMonitorForProject() returned false, causing either a 400
error or a fallback to agentStore.startHeartbeatRun() which only creates
a record without executing.
Fix: add resolveHeartbeatMonitor() which looks up the correct engine's
HeartbeatMonitor via engineManager by matching the scoped store's root
directory. Update POST /agents/:id/runs, POST /agents/:id/heartbeat,
POST /agents/:id/runs/stop, and triggerCommentWakeForAssignedAgent to
use the resolved monitor instead of the single static one.
ClosesRunfusion/Fusion#31
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Co-authored-by: Orca <help@stably.ai>
Replaced the board's task ordering from time-based (`columnMovedAt`/`createdAt`) to a priority-then-task-number sort, inlining previously external priority helpers to fix Vite bundling. Tests were updated in `Board.test.tsx` to match the new sort semantics and `ChatView.test.tsx` keyboard tests were
Fusion-Task-Id: FN-3257
- Throw from init() when integrity check fails and recovery doesn't help,
preventing writes to a known-corrupt database
- Only drain buffer on successful flush; requeue valid entries on transient
failures (busy/IO) so they aren't silently lost
- Add spy on flushAgentLogBuffer in deleteTask test to prove flush-before-delete
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Merged test(FN-3344): added comprehensive unit and integration tests for the navigation history hook, covering 470 lines of test coverage across navigation state management, history tracking, and edge cases.
Fusion-Task-Id: FN-3344
Cleans up lint and type errors in the engine by removing unused imports from executor and merger, and replacing `any` types with proper type annotations in verification-utils.
Fusion-Task-Id: FN-3345
When App-level activation triggers the lock from an input gaining focus,
iOS may have already scrolled the document by the time the lock effect
runs. Capturing that already-shifted scrollY and restoring it on release
left the dashboard pushed up after the keyboard dismissed (the inline-
edit case). Always snap to 0 instead — the dashboard has body{overflow:
hidden} so user-initiated scroll is always 0 anyway, and any non-zero
value at lock time is iOS-forced and should be reset.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Inline editing on the dashboard (e.g. TaskCard inline edit) doesn't go
through a modal, so it didn't get useMobileScrollLock applied. iOS would
shift the visualViewport / window scroll for the focused input and
sometimes leave the dashboard pushed up after the keyboard dismissed.
App.tsx now activates useMobileScrollLock whenever mobileKeyboardOpen is
true (mobile + keyboard up + no modal open). Modals continue to manage
their own lock; the hook is reference-counted so they don't conflict.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
