Surface workflow-diagnostic session events (warnings and errors from
Agent Framework workflows) in the turn-event log of the Activity Panel.
- Add workflow-diagnostic case to formatTurnEventEntry with label/detail
formatting that includes executor ID, subworkflow ID, exception type,
and diagnostic message when present
- Add AlertTriangle icon for diagnostic events with error/warning color
- Add 5 tests covering executor-failed, workflow-warning, subworkflow-error,
workflow-error, and missing diagnosticKind fallback
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Replace the single aggregate thinking tile with per-turn thinking groups.
Instead of collecting all thinking messages into one flat array and
rendering a single ThinkingProcess before the last assistant message,
process session.messages in chronological order to produce interleaved
display items that naturally group consecutive thinking messages by turn.
Each turn now shows its own collapsible thinking tile placed inline
before its assistant response, with correct per-turn isActive state
and turnStartedAt from the matching run.
Also fixes a latent index-mismatch bug in getAssistantMessagePhase where
the visible-messages index was compared against the full session.messages
index, potentially hiding the Final badge when thinking messages existed.
Switched to ID-based comparison.
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Wire message-reclassified sidecar events through the main process to
the renderer, where reclassified messages are filtered out of the main
transcript and collected into a collapsible ThinkingProcess component.
Main process changes:
- Route message-reclassified via dedicated onMessageReclassified callback
- Add applyMessageReclassified handler that sets messageKind and emits
the session event for the renderer
- Forward assistant-intent and reasoning-delta as turn-scoped events
Renderer changes:
- Split session.messages into visibleMessages and thinkingMessages
- Render ThinkingProcess above the last assistant message
- ThinkingProcess auto-expands during active turns, collapses on finish
- Update messagePhase to skip thinking-kind messages for final detection
Tests:
- 3 new sessionWorkspace tests for reclassification apply/dedup/ignore
- 2 new messagePhase tests for thinking-kind handling
- Updated runTurnPending test fixtures for new callback
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Add three layers of usage visibility:
- ChatPane footer: premium request count, AIU consumed, and quota
remaining below the existing context-window bar
- Settings / CopilotStatusCard: on-demand account quota section with
progress bars, overage indicators, and reset dates fetched via
the new get-quota sidecar command
- Activity Panel: per-agent token/cost/duration totals on each agent
row and a Session Usage summary section between agents and timeline
Backend (sidecar):
- New get-quota command using SDK account.getQuota RPC
- New assistant-usage turn-scoped event from SDK assistant.usage
- QuotaSnapshotMapper for both typed and untyped SDK quota payloads
- DTOs: GetQuotaCommandDto, QuotaSnapshotDto, AccountQuotaResultEventDto,
AssistantUsageEventDto
Frontend:
- Shared types: AssistantUsageEvent, QuotaSnapshot, GetQuotaCommand
- IPC bridge: getQuota channel, assistant-usage event dispatch
- State: SessionRequestUsageMap accumulator with per-agent breakdown
- 8 new tests for accumulator logic and formatting helpers
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
The CodeHighlightPlugin's selection restoration could create an invalid
Lexical selection targeting a LineBreakNode with type 'element'. Since
LineBreakNode is not an ElementNode, Lexical threw during reconciliation
and the LexicalErrorBoundary replaced the editor with an error state.
The fix ensures findPoint never targets a LineBreakNode directly.
Instead it falls back to an element-level selection on the parent
CodeNode using the child index, which is always a valid target.
Extracted the selection helpers (getCodeNodeAbsoluteOffset,
findCodeNodeSelectionPoint, restoreCodeNodeSelection) into
markdownEditor.ts for testability and added regression tests.
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Display compact status cards in the chat transcript showing each
sub-agent's name, current activity (Thinking, Using grep, etc.),
and elapsed time. Cards appear while sub-agents are running and
clear when the session goes idle.
- Extend SessionEventRecord with description, error, toolCallId,
and model fields for subagent events
- Forward additional subagent fields in handleTurnScopedEvent
- Add subagentTracker reducer to derive active subagent state from
session events, including agent-activity correlation
- Create SubagentActivityCard component with spinner, status icon,
agent name, activity label, and live elapsed timer
- Wire activeSubagents state in App.tsx and pass to ChatPane
- Add 16 unit tests for the subagent tracker
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
When the agent produces multiple intermediate responses during a turn,
each gets a unique messageId but all stay pending until finalizeTurn().
This caused multiple stacked 'Thinking' bubbles in the chat transcript.
Now, when a new messageId arrives in applyTurnDelta, any previously
pending assistant messages are marked complete and message-complete
events are emitted before the new message-delta. The renderer reducer
mirrors this logic for defensive consistency.
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Replace the old disconnected-node addAgentNodeToGraph with the
mode-aware addAgentToGraph so new agents are automatically wired
into the graph topology. Replace the graph-nuking removeAgent
path with removeAgentFromGraph for surgical node removal. Split
emitChange into graph-preserving and mode-change paths so editing
agent properties no longer discards custom positions and edges.
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Clear only active agent statuses when a session reaches idle, and keep the terminal cancelled/error run cleanup in place. Add regression coverage for idle cleanup after cancellation-like flows.
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Clear only still-active side-panel agent states when a run ends as cancelled or error, while preserving completed agent statuses. Add a renderer regression test for cancelled runs.
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Resolve chat tooling from the selected session's project instead of the
separately tracked selectedProjectId so project-scoped discovered MCPs do
not appear for the wrong session. Also synchronize selectedProjectId when
a session is selected and add regression tests for both behaviors.
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
- Add CSS styling for selected edges (indigo highlight with glow)
- Enable orchestrator↔agent connections in group-chat mode
- Downgrade disconnected-agent validation from error to warning
- Rename addHandoffEdge to addEdge (generic for all modes)
- Add tests for group-chat connection rules
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Extend edge deletion to group-chat orchestrator↔agent edges in addition
to handoff agent↔agent edges. Structural edges (user-input/output,
distributor/collector) remain non-deletable across all modes.
Sequential, concurrent, and single modes keep fully structural topologies
with no user-deletable edges.
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
- Color-code edges by semantic role: structural edges (system↔node) in
muted zinc, agent-to-agent edges in indigo for clear visual separation
- Switch from smoothstep to bezier curves for smoother, more natural
edge routing that reduces right-angle overlaps
- Stagger agent nodes at alternating X offsets in concurrent, handoff,
and group-chat layouts so parallel fan-out/fan-in edges take distinct
bezier paths instead of stacking
- Increase spacing in all layouts to give edges more room
- Add dagre-powered auto-layout button (top-right of canvas) that
computes optimal left-to-right hierarchical node positions
- Add 2 new tests for auto-layout across all orchestration modes
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
- Add ArrowClosed markers to all graph edges for clear directional flow
- Replace generic Bot icon with AI provider logos (OpenAI, Anthropic,
Google) on agent nodes, inferred from the agent's model id
- Show model display name as subtitle on agent nodes (e.g. 'GPT-5.4',
'Claude Opus 4.5')
- Thread availableModels catalog from PatternEditor through canvas to
resolve friendly model names; falls back to raw model id
- Add 3 new tests for arrow markers, provider icons, and model labels
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
- Hide input handle on User Input nodes and output handle on User Output
nodes (separate userInputNode/userOutputNode node types)
- Enable edge deletion in handoff mode only (Delete key or React Flow UI);
only agent-to-agent edges are deletable, structural edges are protected
- Block all edge mutations (add/delete) in concurrent and group chat modes
- Add agent as disconnected node without auto-rebuilding the graph
- Add sequential reorder controls (↑↓) in the inspector panel with
position/order swap and automatic edge rebuilding
- Replace circular group chat layout with vertical column to prevent
bidirectional edge crossings
- Tighten handoff layout spacing so edges between triage and specialists
don't overlap nodes
- Use smoothstep edge type for cleaner edge routing across all modes
- Add 6 new tests covering reorder, disconnected add, edge deletion rules,
and node type assertions
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Replace the flat agent list and static flow diagrams in PatternEditor with
an interactive React Flow canvas backed by the authoritative PatternGraph
model. Users can now see and manipulate orchestration topology visually.
Changes:
- Add @xyflow/react (React Flow) as a canvas library
- Create patternGraph view-model layer for graph-to-canvas projection
- Create GraphNodes with custom system/agent node components
- Create PatternGraphCanvas with drag, connect, and selection support
- Create PatternGraphInspector for selected node/agent details
- Rewrite PatternEditor with split layout: canvas + inspector sidebar
- Add dark theme CSS overrides for React Flow
- Add 11 tests covering view-model projection, connection rules, and
graph mutation helpers
- Update ARCHITECTURE.md to reflect the new renderer graph boundary
The canvas projects PatternGraph into React Flow nodes/edges. Handoff mode
supports drawing new agent-to-agent edges interactively. Node positions
persist to the graph on drag completion. Agent editing moves from inline
forms to the inspector panel on node selection.
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
- Add RunTimeline component with vertical event timeline, collapsible run
cards, per-kind icons, status animations, and agent lane badges
- Add runTimelineFormatting helpers for timestamps, durations, labels,
content truncation, and consecutive thinking-event collapsing
- Integrate Timeline section into ActivityPanel between Agents and Tools
- Wire jump-to-message in App.tsx using DOM data-message-id attributes
on ChatPane messages with smooth scroll and temporary highlight ring
- Add comprehensive unit tests for all formatting helpers
- Update HANDOVER.md to document completed frontend implementation
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
