- Add MessageSquare icon to GraphNodeCard; prop-thread onConversationOpen
and selectedTaskId through WorkflowGraph node data (no useUrlState
inside ReactFlow nodes — avoids context/timing issues)
- Fix ContextualRightPanel: check taskId before epicId so clicking the
conversation icon always opens ThreadDrawer even when an epic filter
is active
- setEpicId now clears task from URL so selecting an epic resets any
open conversation thread
- handleGraphSelect toggles: second click on same node calls setTaskId(null)
closing the right panel
- Add onSelect to WorkflowGraph flowModel deps to prevent stale callbacks
- Fix ContextualRightPanel onClose no-ops: wired to setTaskId(null) /
setSwarmId(null) so back button works
- Right panel always visible (removed panel==='open' gate in UnifiedShell)
- SmartDag task grid: horizontal scroll, fixed-width cards, hideClosed=true
- Add <Suspense> in page.tsx for useSearchParams compatibility
- Enable dolt auto-start in .beads/config.yaml
- Add 14 static analysis tests (graph-node-conversation.test.tsx)
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
## Design Decision
Per bd (bead) system design, a task should have only ONE agent archetype
assigned at a time. This provides clear ownership and simpler mental model.
## What Changed
When assigning a new archetype:
1. Remove any existing agent: labels first (DELETE API)
2. Then add the new agent: label (POST API)
3. Optimistic UI updates to match
## Why This Makes Sense
- Clear ownership: 'Who's working on this?'
- Simpler coordination between tasks
- Matches how bd/agent orchestration is intended to work
- Reassigning is still possible (just click a different archetype)
## UI Behavior
- If task has 'coder' assigned, clicking 'architect' will:
1. Remove 'coder' label
2. Add 'architect' label
- Dropdown shows 'Assigned' badge on current archetype
- X button still available to unassign completely
## Test Coverage
Added graph-node-single-archetype.test.tsx with 5 tests:
- Removes existing labels before adding new
- Calls DELETE before POST
- Only allows one archetype per task
- Preserves non-agent labels
- Returns early if same archetype clicked
## The Bug
User reported: 'An archetype can only exist on one task at a time - when
I try to make the next task have the same arch, it deleted the one I
added prior.'
## Root Cause
The SSE subscription (useBeadsSubscription) refreshes data from the server
whenever ANY change happens. When user assigns an archetype:
1. User clicks assign -> optimistic update adds label locally
2. SSE fires (from previous operation or heartbeat) -> fetches fresh data
3. useEffect syncs localLabels with data.labels (which doesn't have the
new label yet)
4. Label disappears from UI
5. Eventually API completes and another SSE refresh brings it back
This race condition causes labels to flicker or disappear entirely.
## The Fix
Track pending optimistic labels in a useRef Set, and merge them with
incoming server data during sync:
1. pendingOptimisticLabels = useRef<Set<string>>(new Set())
2. When optimistically adding: add to pending set
3. useEffect merge: combine server labels + pending labels
4. After API completes: remove from pending set
This ensures optimistic labels survive SSE refreshes.
## Test Coverage
Added graph-node-labels-optimistic.test.tsx with 10 tests:
- Uses useRef for tracking
- Tracks in a Set
- Preserves labels during sync
- Adds/removes from pending set
- Handles multiple concurrent operations
- Per-node state isolation
## Verification
- typecheck: pass
- lint: pass (0 errors)
- test: all pass
Various supporting changes made during the assign archetypes feature development:
- Added contextual-right-panel.tsx and swarm-command-feed.tsx
- Updated activity-panel.tsx with new features
- UI improvements to left-panel, mobile-nav
- Test updates for url-state-integration, mobile-nav, top-bar
- Package.json updates for dependencies
- Global CSS refinements
These changes support the main assign archetypes feature but are
not directly part of its core functionality.
## Context
This commit adds the supporting infrastructure that makes the assign
feature work end-to-end.
## Components Added/Modified
### SmartDag
- Main view component for graph-based task management
- Integrates TaskCardGrid and WorkflowGraph
- Has 'Assign' mode toggle button
- Passes archetypes and assignMode to WorkflowGraph
- Manages filter state (hideClosed, sortReadyFirst, etc.)
### useGraphAnalysis Hook
- Extracted graph analysis logic for reuse
- Returns: actionableNodeIds, cycleNodeIdSet, blockerTooltipMap, etc.
- Used by both SmartDag and AssignmentPanel
- Ensures consistent 'actionable' definition across components
### UnifiedShell
- Added assignMode state
- Added selectedAssignIssue state
- Renders AssignmentPanel when in graph view + assign mode
- Wires up onAssignModeChange and onSelectedIssueChange callbacks
## Design Philosophy
- Shared hook means single source of truth for 'actionable'
- Clean separation between view (SmartDag) and sidebar (AssignmentPanel)
- URL state preserved for navigation
## Test Coverage
- SmartDag tests: 12 tests covering buttons, callbacks, imports
- useGraphAnalysis tests: 6 tests covering cycle detection, blockers
- UnifiedShell tests: 9 tests covering state and rendering
## The Feature Request
User wanted an enhanced sidebar panel showing:
- Tasks needing agents (ready but unassigned)
- Pre-assigned tasks waiting to start
- Active workers on current epic
## Design Collaboration
We discussed what each section should show:
1. **Needs Agent**: Actionable tasks (no blockers) without agent: label
2. **Pre-assigned**: Tasks with agent: label, not yet in_progress
3. **Squad Roster**: in_progress tasks with assignee
## Technical Implementation
- Uses useGraphAnalysis hook for actionableNodeIds
- Helper functions: hasAgentLabel(), getAgentLabels(), extractArchetypeIdFromLabel()
- Quick assign dropdown on each 'Needs Agent' item
- Archetype badges shown on 'Pre-assigned' items
## UI/UX Decisions
- Each section has count badge in header
- Max-height with scroll for each section
- Consistent styling with existing panel patterns
- Uses CSS variables for theming
## Test Coverage
- Added assignment-panel-sections.test.tsx with 5 TDD tests
- Tests verify: useGraphAnalysis import, section headers, filtering logic
## Beads: beadboard-b7t (closed)
## The Collaboration Story
User requested ability to assign agent archetypes to tasks directly from
graph nodes. This was the core UI feature request.
## Design Decisions Made Together
1. **Placement**: We decided to put the assign UI at the bottom of the card
to not interfere with existing status/badges display
2. **Pattern**: Used Radix dropdown-menu (already in project) for consistency
3. **Visual feedback**: Added loading spinner during API calls, success/error
messages that auto-dismiss
4. **Closed tasks**: Excluded closed tasks from assignment (logical constraint)
## Issues We Encountered
- Initially only added POST handler for assignment
- User pointed out we needed DELETE for unassign - added that
- The unassign button (X) needed to call DELETE not POST
## API Changes
- Added DELETE handler to /api/swarm/prep for removing agent assignments
- Uses 'bd label remove' under the hood
## What the UI Shows
- Dropdown with available archetypes
- Badge showing assigned archetype name with color
- X button to unassign (on each badge)
- 'Assigned' label on already-assigned archetypes in dropdown
## Test Coverage
- Added graph-node-assign.test.tsx with 6 TDD tests
## Beads: beadboard-brq (closed)
## Context
This is the foundation commit for the 'Assign Archetypes to Tasks' feature.
We needed a way to display which agents are assigned to tasks directly on
the graph nodes.
## Decision Process
- User wanted to see agent assignments on DAG nodes
- We discovered that labels (including 'agent:archetype-id' format) weren't
being passed through the WorkflowGraph component
- Added 'labels' and 'archetypes' to GraphNodeData interface
## What Changed
- WorkflowGraph now passes issue.labels to each node's data
- GraphNodeData interface updated to include labels: string[]
- Added archetypes prop for dropdown population
## Test Coverage
- Added graph-node-labels.test.tsx with 4 passing tests
## Beads: beadboard-yo5 (closed)
This commit includes the new SwarmWorkspace with its 3 sub-tabs, the LeftPanel mission picker, and the comprehensive Operations Command Dashboard featuring the live interactive DAG telemetry and task assignment prep flow.
STORY:
With the shell layout complete, we needed the actual content for each view.
Three agents worked in parallel on the card components that would populate
the Social and Swarm views, plus integrating the existing graph into the shell.
COLLABORATION:
Agent bb-98c (social-card-builder) created SocialCard:
- Task ID with teal styling
- UNLOCKS section (green) showing what this task unblocks
- BLOCKS section (amber) showing what's blocking this task
- Agent avatars with liveness glow
- View-jump icons for quick navigation
Agent bb-nuy (swarm-card-builder) created SwarmCard:
- Agent roster with liveness indicators
- Progress bar (ASCII block format: ████████░░░░)
- Attention items with warning styling
- View-jump icons
Agent bb-54x (graph-integrator) integrated WorkflowGraph:
- Created GraphView wrapper with Flow/Overview tabs
- Wired into UnifiedShell when view=graph
- Connected taskId to selectedId for URL sync
- Connected graphTab to URL state
DELIVERABLES:
- src/components/social/social-card.tsx: Task card for activity feed
- src/components/swarm/swarm-card.tsx: Swarm health card
- src/components/graph/graph-view.tsx: Graph wrapper with tabs
- src/components/shared/mobile-nav.tsx: Bottom tab bar
- Tests for all components
VERIFICATION:
- npm run typecheck: PASS
- npm run lint: PASS
- npm run test: PASS
CLOSES: bb-ui2.11, bb-ui2.16, bb-ui2.20
STORY:
Phase 1 of the Unified UX epic required a complete 3-panel shell layout
with responsive behavior across mobile, tablet, and desktop breakpoints.
The existing page structure was fragmented - we needed a cohesive shell.
COLLABORATION:
Three agents (bb-5am, bb-dwz, bb-3dv) worked in parallel on:
- TopBar: View tabs (Social/Graph/Swarm) with active states, filter input
- LeftPanel: Channel tree navigation with epic filtering, responsive collapse
- RightPanel: Detail strip with sidebar (desktop) / drawer (tablet/mobile) modes
We encountered a hydration mismatch error on mobile/tablet because
useResponsive was returning different values on server vs client.
Fixed by defaulting to desktop on server and only updating after mount.
Mobile navigation (bb-ui2.27) added:
- Hamburger menu for left panel access on mobile/tablet
- Bottom tab bar for thumb-friendly view switching
DELIVERABLES:
- src/components/shared/top-bar.tsx: TopBar with view tabs + hamburger
- src/components/shared/left-panel.tsx: Epic tree with expand/collapse
- src/components/shared/right-panel.tsx: Responsive sidebar/drawer
- src/components/shared/unified-shell.tsx: Main 3-panel grid layout
- src/components/shared/mobile-nav.tsx: Bottom tab bar for mobile
- src/hooks/use-responsive.ts: Breakpoint detection (mobile/tablet/desktop)
- Tests for all components
VERIFICATION:
- npm run typecheck: PASS
- npm run lint: PASS
- npm run test: PASS
CLOSES: bb-ui2.6, bb-ui2.7, bb-ui2.8, bb-ui2.9, bb-ui2.27
STORY:
Development requires supporting scripts and comprehensive test coverage.
This commit adds utility scripts and extends test suites.
COLLABORATION:
- scripts/capture-sessions.mjs: Screenshot capture for sessions hub
- scripts/capture-timeline.mjs: Timeline view capture utility
- sessions-header-logic.ts: Session header business logic
- Additional test files for sessions, hooks, and snapshot diffing
STORY:
Agent identities were stored in a local JSON registry, but they should
be first-class beads visible in the BeadBoard system. This consolidates
agent identity to bd CLI as the source of truth.
COLLABORATION:
Replaced local JSON registry with bd CLI wrapper in agent-registry.ts:
- All agent operations now delegate to bd CLI
- Agents appear as team-visible beads with gt:agent label
- Identity isolation prevents agent beads from polluting mission lists
The consolidation makes agents visible to the entire team and ensures
consistent identity management across all tools.
DELIVERABLES:
- src/lib/agent-registry.ts refactored to bd CLI wrapper
- tests/lib/agent-registry-bd.test.ts for bd integration
- tools/bb.ts updated for consolidated identity ops
VERIFICATION:
- All registry tests PASS
- Agents appear on agent page but NOT in task lists
- Quality gates (typecheck, lint) GREEN
CLOSES: bb-1y7
EPIC: bb-u6f
STORY:
The Swarm view needs to show epics as "swarms" - groups of agents
working together on a mission. This requires aggregating bead data
by swarm/epic and computing health statistics.
COLLABORATION:
Created buildSwarmCards function that transforms epic + agent data:
SwarmCard interface:
- id, title, status
- stats: { completed, active, ready, blocked, total }
- agents: AgentRoster[] with liveness
- attention: blocked tasks + stuck agents needing attention
- lastActivity
We also created swarm-molecules.ts for molecular composition patterns
used by the swarm orchestration layer.
DELIVERABLES:
- src/lib/swarm-cards.ts with SwarmCard, AgentRoster types
- src/lib/swarm-molecules.ts for molecular composition
- tests/lib/swarm-cards.test.ts
- tests/lib/swarm-molecules.test.ts
VERIFICATION:
- npm run typecheck: PASS
- npm run lint: PASS
- npm run test: PASS
CLOSES: bb-ui2.15
BLOCKS: bb-ui2.16, bb-ui2.17
STORY:
The Social view needs to transform raw BeadIssue data into renderable
SocialCard objects. This includes computing blocked/blocking relationships
from dependencies and extracting agent assignments.
COLLABORATION:
Created buildSocialCards function that transforms BeadIssue → SocialCard:
SocialCard interface:
- id, title, status
- blockedBy: tasks this task depends on
- blocking: tasks that depend on this task
- agents: assigned agents with liveness
- lastActivity: most recent event
The function derives blockedBy from depends_on dependencies and blocking
from blocked_by reverse dependencies, creating a complete picture of
task relationships for the activity feed.
DELIVERABLES:
- src/lib/social-cards.ts with SocialCard interface and builder
- tests/lib/social-cards.test.ts
VERIFICATION:
- npm run typecheck: PASS
- npm run lint: PASS
- npm run test: PASS
CLOSES: bb-ui2.10
BLOCKS: bb-ui2.11, bb-ui2.12
STORY:
The Unified UX needs URL as the single source of truth for view state.
This enables deep linking, bookmarking, and shareable URLs that
preserve exactly what the user was looking at.
COLLABORATION:
Created useUrlState hook using Next.js useSearchParams and useRouter:
Interface:
- view: 'social' | 'graph' | 'swarm' (default: social)
- taskId: selected task ID (for detail panel)
- swarmId: selected swarm ID
- panel: 'open' | 'closed' (detail panel state)
- graphTab: 'flow' | 'overview' (graph view mode)
URL Patterns:
- /?view=social
- /?view=social&task=bb-buff.1&panel=open
- /?view=swarm&swarm=bb-buff
- /?view=graph&task=bb-buff.1&graphTab=flow
The hook uses router.push for URL updates, ensuring no local state
drift from the URL source of truth.
DELIVERABLES:
- src/hooks/use-url-state.ts with parseUrlState, buildUrlParams, useUrlState
- tests/hooks/use-url-state.test.ts with 18+ tests
VERIFICATION:
- npm run typecheck: PASS
- npm run lint: PASS
- npm run test: PASS (all tests including new ones)
CLOSES: bb-ui2.4
BLOCKS: bb-ui2.5
STORY:
In a multi-agent control center, operators need to quickly identify
what TYPE of agent they're looking at - UI agents, graph agents,
orchestrators, or general workers.
COLLABORATION:
We implemented role-based border colors on agent avatars:
- ui agents: blue border
- graph agents: green border
- orchestrators: purple border
- default/other: gray border
The agent-station component now displays these colors, making it
instantly visible what role each agent plays in the swarm.
DELIVERABLES:
- AgentStation component with role-based styling
- agent-station-logic.ts with role color derivation
- Tests: agent-station-logic.test.ts updated and passing
VERIFICATION:
- typecheck: PASS
- lint: PASS
- test: PASS
CLOSES: bb-buff.3.4
STORY:
The Sessions Hub needs to show which task each agent is working on,
creating a visual "mission link" between agents and their active work.
COLLABORATION:
We implemented the data layer for mission pathing:
- getAgentActiveMissions() returns tasks owned by an agent
- getActiveMissionCount() for badge indicators
- getMissionsByAgent() groups all missions for batch rendering
DELIVERABLES:
- src/lib/agent-sessions.ts extended with mission functions
- Tests: 8/8 PASS in tests/lib/mission-pathing.test.ts
STATUS: in_progress (visual rendering layer still pending)
Visual path lines would require SVG overlay + position tracking.
CLOSES: partial bb-buff.3.3
STORY:
The Sessions Hub needed clear visual distinction between healthy
agents and those in trouble. Users couldn't quickly identify stuck
or dead agents in the control center view.
COLLABORATION:
We added 'stuck' and 'dead' states to the AgentSessionState type,
created deriveSessionState() with Zero-Failure-Check priority,
and implemented restrained visual treatments:
- stuck: pulsing red border (ring-2 ring-red-500 animate-pulse)
- dead: strong ghosting (opacity-40 grayscale)
- evicted: milder ghosting (opacity-60 grayscale-[0.5])
Session cards now display STUCK/OFFLINE badges with aria-labels
for accessibility.
DELIVERABLES:
- AgentSessionState extended with stuck/dead states
- deriveSessionState() derives from ZFC state priority
- Visual treatments for stuck/dead/evicted
- Accessible badges with aria-label
TESTS:
- tests/lib/agent-sessions-state.test.ts: 6/6 PASS
- tests/components/shared/status-utils-visual.test.ts: 4/4 PASS
- tests/components/sessions/session-feed-card-state.test.tsx: 4/4 PASS
CLOSES: bb-buff.3.2
BLOCKS: bb-buff.3.3
STORY:
The session backend needed to aggregate agent health from a live
telemetry stream rather than static bead metadata. This refactor
makes liveness signals real-time and accurate.
COLLABORATION:
We extended the ActivityEvent model with a native 'heartbeat' kind,
updated extendActivityLease() to emit through the activity bus, and
refactored getAgentLivenessMap() to prioritize heartbeat activity
history over stale bead metadata.
DELIVERABLES:
- ActivityEvent extended with 'heartbeat' kind
- extendActivityLease() emits heartbeats through activity bus
- getAgentLivenessMap() prefers telemetry over static metadata
- Registry APIs support projectRoot injection for testing
- Tests verify preference logic via TDD
VERIFICATION:
- 93/93 tests PASSING
- Heartbeat override verified in isolated temp projects
CLOSES: bb-buff.1.3
BLOCKS: bb-buff.3.2, bb-buff.3.3, bb-buff.2.1
Research revealed that agent identities (consolidated to bd beads) were appearing in standard task lists because the data-access layer lacked identity-awareness.
- Refactored read-issues.ts and parser.ts to explicitly exclude beads labeled 'gt:agent' from standard mission flows.
- Verified that agent personas remain targetable by the registry but are invisible to Kanban/Graph/Sessions.
- Added Characterization Test: identity-isolation.test.ts.
This restores the 'War Room' clarity by separating Operatives from Missions.
OPERATIVE: silver-castle
SESSION: 2026-02-14-1630
We've transformed the Social-Dense Hub into a high-fidelity operational surface.
- BACKEND: Implemented Global Incursion Engine in agent-sessions.ts (N^2 overlap detection) and added the 60m 'Idle' state.
- API: Enriched the sessions payload with full metadata and active conflict arrays.
- HEADER: Delivered 4-state agent stations (Active/Stale/Evicted/Idle) with real-time 'time-ago' timers.
- FEED: Implemented the 'Fire Map' visuals:
* Global Incursion Ticker: High-visibility alerts for agent collisions.
* Local Conflict Badges: Pulsing pills on affected task cards.
- Refactored components for React-static compliance and strict TypeScript safety.
This commit completes the visibility track, allowing the human supervisor to monitor agent presence and friction in real-time.
OPERATIVE: silver-castle
SESSION: 2026-02-14-1430
Following a critical collaboration to resolve Windows terminal pop-ups, we've delivered a more robust 'Passive Activity' architecture:
- Terminology Pivot: Renamed 'Heartbeat' to 'Activity Lease' (Parking Permit model).
- Side-Effect Extension: tools/bb.ts now automatically extends the agent's lease whenever they perform real work (any CLI command).
- Passive Handshake: bb-init.mjs now only performs an initial registration/lease start, with no background loops.
- 100% Silence: Removed all background process spawning, ensuring zero terminal disruption on Windows.
- High Observability: Liveness is still tracked via the 15m threshold, but relies on activity rather than periodic pings.
OPERATIVE: silver-castle
SESSION: 2026-02-14-1330
The previous background loop approach was disruptive on Windows (terminal pop-ups).
We collaborated to find a more robust, silent alternative:
- Removed all background heartbeat worker logic and PID management.
- Implemented 'Passive Heartbeat' in tools/bb.ts: every agent command now refreshes liveness as a side-effect.
- Updated bb-init.mjs to use explicit heartbeat calls for adoption/registration.
- Result is 100% silent observability: if an agent is working, they are Active. If they stop, they drift to Stale.
OPERATIVE: silver-castle
SESSION: 2026-02-14-1300
Finalizing the backend engine for the Operative Protocol.
- Updated agent-sessions.ts to use the deriveLiveness logic and the 15m protocol threshold.
- Integrated the agentLivenessMap into the session aggregation to provide real-time status in the Hub.
- Updated the GET /api/sessions endpoint to fetch and serve liveness metadata.
- Fixed linting warnings (unused imports) in reservations, sessions, and test files.
This commit completes the backend contract for bb-u6f.6.2, providing the data layer necessary for the upcoming 'War Room' UI enhancements.
OPERATIVE: silver-castle
SESSION: 2026-02-14-1145
Our collaboration led to a rigorous 'Session Constitution' where we prioritized observability and concurrency safety.
I've delivered the first four pillars of the backend engine:
1. Liveness Registry: Heartbeat logic and derivation of active/stale/evicted states based on the 15m threshold.
2. Overlap Classifier: Canonical path normalization (Windows-aware) and exact/partial overlap detection.
3. Takeover Rules: Enforced discipline where active agents are protected, while stale/evicted ones can be overtaken via --takeover-stale.
4. Protocol Schema: Establishing the v1 envelope for high-fidelity agent signaling.
TDD was applied throughout, with 100% pass rate on the new liveness, overlap, takeover, and protocol tests.
We completed the 'Deep Metadata Etch' today, transforming our Beads issues from simple trackers into a permanent narrative of our collaboration.
Triumphs:
- Exhaustively updated all epic and sub-task descriptions with technical implementation reports and 'Execution Tales'.
- Finalized the 'bb' agent CLI skill (bb.ps1), providing a reliable, path-safe interface for cross-agent communication.
- Published ADR-001 and RFC-001 to document our coordination protocols.
- Fixed the 'missing closed issues' bug across all pages by enforcing --all and --limit 0 in read-issues.ts.
Raw Honest Moment:
We realized our 'Memory Bank' was initially too shallow. We went back and re-wrote descriptions for over 15 beads to ensure that future AI agents (and human maintainers) understand not just *what* we built, but *why* we chose specific architectural trade-offs. This commit represents our commitment to documentation as a first-class citizen of engineering.
We added the third major surface to the BeadBoard workspace: the Chronological Timeline. This provides the 'Audit' layer of our operational hierarchy.
Triumphs:
- Built the /timeline route with sticky date grouping and polymorphic EventCards.
- Integrated the ActivityPersistence library to bridge the gap between ephemeral SSE events and persistent project history.
- Implemented real-time Agent Stats endpoints (/api/agents/[id]/stats) that derive throughput and 'Wins' from the project stream.
Raw Honest Moment:
We almost shipped this without persistence, which would have meant the project history would disappear every time the server restarted. Realizing that 'Observability' requires 'Survivability' led us to build the .beadboard/activity.json buffer, a small but vital piece of engineering that makes the timeline actually useful.
Today we reached a major architectural conclusion: project history shouldn't be stored, it should be derived. We rejected the overhead of a separate SQLite event store in favor of an O(N) snapshot-diffing engine that computes human-readable narratives directly from the issues.jsonl source of truth.
Key Triumphs:
- Implemented O(N) diffing algorithm in src/lib/snapshot-differ.ts that transforms raw JSONL into 16 distinct social event types.
- Engineered a file-based persistence layer (src/lib/activity-persistence.ts) to solve the 'Next.js HMR Wiped My Memory' bug, ensuring project heartbeat survives server restarts.
- Developed the agent-session data model that unifies Beads, Activity, and Cross-Agent Mail into a single 'Mission' context.
Raw Honest Moment:
We struggled for over an hour with 'missing history' before realizing that development-mode reloads were purging our in-memory buffers. The shift to a file-backed ring buffer was a reactive pivot that became a core project strength.
