feat: wire MultiAgentCoordinator into runtime

CLAUDE.md

@@ -98,13 +98,13 @@ curl http://localhost:3000/api/v1/health   # backend (via web proxy)
 
 ```text
 src/ai_company/
-  api/            # Litestar REST + WebSocket API (controllers, guards, channels, JWT + API key auth, approval gate integration)
+  api/            # Litestar REST + WebSocket API (controllers, guards, channels, JWT + API key auth, approval gate integration, coordination endpoint)
   budget/         # Cost tracking, budget enforcement (pre-flight/in-flight checks, auto-downgrade), billing periods, cost tiers, quota/subscription tracking, CFO cost optimization (anomaly detection, efficiency analysis, downgrade recommendations, approval decisions), spending reports, budget errors (BudgetExhaustedError, DailyLimitExceededError, QuotaExhaustedError)
   cli/            # CLI interface (future — thin API wrapper if needed)
   communication/  # Message bus, dispatcher, messenger, channels, delegation, loop prevention, conflict resolution, meeting protocol
   config/         # YAML company config loading and validation
   core/           # Shared domain models, base classes, and resilience config (RetryConfig, RateLimiterConfig)
-  engine/         # Agent orchestration, execution loops, parallel execution, task decomposition, routing, task assignment, centralized single-writer task state engine (TaskEngine), task lifecycle, recovery, shutdown, workspace isolation, coordination (multi-agent pipeline: TopologyDispatcher protocol, 4 dispatchers — SAS/centralized/decentralized/context-dependent, wave execution, workspace lifecycle integration), coordination error classification, prompt policy validation, checkpoint recovery (checkpoint/, per-turn persistence, heartbeat detection, CheckpointRecoveryStrategy), approval gate (escalation detection, context parking/resume, EscalationInfo/ResumePayload models)
+  engine/         # Agent orchestration, execution loops, parallel execution, task decomposition, routing, task assignment, centralized single-writer task state engine (TaskEngine), task lifecycle, recovery, shutdown, workspace isolation, coordination (multi-agent pipeline: TopologyDispatcher protocol, 4 dispatchers — SAS/centralized/decentralized/context-dependent, wave execution, workspace lifecycle integration, CoordinationSectionConfig company config bridge, build_coordinator factory), coordination error classification, prompt policy validation, checkpoint recovery (checkpoint/, per-turn persistence, heartbeat detection, CheckpointRecoveryStrategy), approval gate (escalation detection, context parking/resume, EscalationInfo/ResumePayload models)
   hr/             # HR engine: hiring, firing, onboarding, offboarding, agent registry, performance tracking (task metrics, collaboration scoring, trend detection), promotion/demotion (criteria evaluation, approval strategies, model mapping)
   memory/         # Persistent agent memory (pluggable MemoryBackend protocol), backends/ (Mem0 adapter: backends/mem0/), retrieval pipeline (ranking, injection, context formatting, non-inferable filtering), shared org memory (org/), consolidation/archival (consolidation/)
   persistence/    # Operational data persistence — pluggable PersistenceBackend protocol, SQLite initial (see Memory & Persistence design page)
@@ -151,7 +151,7 @@ web/              # Vue 3 + PrimeVue + Tailwind CSS dashboard
 - **Every module** with business logic MUST have: `from ai_company.observability import get_logger` then `logger = get_logger(__name__)`
 - **Never** use `import logging` / `logging.getLogger()` / `print()` in application code
 - **Variable name**: always `logger` (not `_logger`, not `log`)
-- **Event names**: always use constants from the domain-specific module under `ai_company.observability.events` (e.g. `PROVIDER_CALL_START` from `events.provider`, `BUDGET_RECORD_ADDED` from `events.budget`, `CFO_ANOMALY_DETECTED` from `events.cfo`, `CONFLICT_DETECTED` from `events.conflict`, `MEETING_STARTED` from `events.meeting`, `CLASSIFICATION_START` from `events.classification`, `CONSOLIDATION_START` from `events.consolidation`, `ORG_MEMORY_QUERY_START` from `events.org_memory`, `API_REQUEST_STARTED` from `events.api`, `API_ROUTE_NOT_FOUND` from `events.api`, `CODE_RUNNER_EXECUTE_START` from `events.code_runner`, `DOCKER_EXECUTE_START` from `events.docker`, `MCP_INVOKE_START` from `events.mcp`, `SECURITY_EVALUATE_START` from `events.security`, `HR_HIRING_REQUEST_CREATED` from `events.hr`, `PERF_METRIC_RECORDED` from `events.performance`, `TRUST_EVALUATE_START` from `events.trust`, `PROMOTION_EVALUATE_START` from `events.promotion`, `PROMPT_BUILD_START` from `events.prompt`, `MEMORY_RETRIEVAL_START` from `events.memory`, `MEMORY_BACKEND_CONNECTED` from `events.memory`, `MEMORY_ENTRY_STORED` from `events.memory`, `MEMORY_BACKEND_SYSTEM_ERROR` from `events.memory`, `AUTONOMY_ACTION_AUTO_APPROVED` from `events.autonomy`, `TIMEOUT_POLICY_EVALUATED` from `events.timeout`, `PERSISTENCE_AUDIT_ENTRY_SAVED` from `events.persistence`, `TASK_ENGINE_STARTED` from `events.task_engine`, `COORDINATION_STARTED` from `events.coordination`, `COMMUNICATION_DISPATCH_START` from `events.communication`, `COMPANY_STARTED` from `events.company`, `CONFIG_LOADED` from `events.config`, `CORRELATION_ID_CREATED` from `events.correlation`, `DECOMPOSITION_STARTED` from `events.decomposition`, `DELEGATION_STARTED` from `events.delegation`, `EXECUTION_LOOP_START` from `events.execution`, `CHECKPOINT_SAVED` from `events.checkpoint`, `PERSISTENCE_CHECKPOINT_SAVED` from `events.persistence`, `GIT_OPERATION_START` from `events.git`, `PARALLEL_GROUP_START` from `events.parallel`, `PERSONALITY_LOADED` from `events.personality`, `QUOTA_CHECKED` from `events.quota`, `ROLE_ASSIGNED` from `events.role`, `ROUTING_STARTED` from `events.routing`, `SANDBOX_EXECUTE_START` from `events.sandbox`, `TASK_CREATED` from `events.task`, `TASK_ASSIGNMENT_STARTED` from `events.task_assignment`, `TASK_ROUTING_STARTED` from `events.task_routing`, `TEMPLATE_LOADED` from `events.template`, `TOOL_INVOKE_START` from `events.tool`, `TOOL_OUTPUT_WITHHELD` from `events.tool`, `WORKSPACE_CREATED` from `events.workspace`, `APPROVAL_GATE_ESCALATION_DETECTED` from `events.approval_gate`, `APPROVAL_GATE_ESCALATION_FAILED` from `events.approval_gate`, `APPROVAL_GATE_INITIALIZED` from `events.approval_gate`, `APPROVAL_GATE_RISK_CLASSIFIED` from `events.approval_gate`, `APPROVAL_GATE_RISK_CLASSIFY_FAILED` from `events.approval_gate`, `APPROVAL_GATE_CONTEXT_PARKED` from `events.approval_gate`, `APPROVAL_GATE_CONTEXT_PARK_FAILED` from `events.approval_gate`, `APPROVAL_GATE_PARK_TASKLESS` from `events.approval_gate`, `APPROVAL_GATE_RESUME_STARTED` from `events.approval_gate`, `APPROVAL_GATE_CONTEXT_RESUMED` from `events.approval_gate`, `APPROVAL_GATE_RESUME_FAILED` from `events.approval_gate`, `APPROVAL_GATE_RESUME_DELETE_FAILED` from `events.approval_gate`, `APPROVAL_GATE_RESUME_TRIGGERED` from `events.approval_gate`, `APPROVAL_GATE_NO_PARKED_CONTEXT` from `events.approval_gate`, `APPROVAL_GATE_LOOP_WIRING_WARNING` from `events.approval_gate`). Import directly: `from ai_company.observability.events.<domain> import EVENT_CONSTANT`
+- **Event names**: always use constants from the domain-specific module under `ai_company.observability.events` (e.g., `PROVIDER_CALL_START` from `events.provider`, `BUDGET_RECORD_ADDED` from `events.budget`, `CFO_ANOMALY_DETECTED` from `events.cfo`, `CONFLICT_DETECTED` from `events.conflict`, `MEETING_STARTED` from `events.meeting`, `CLASSIFICATION_START` from `events.classification`, `CONSOLIDATION_START` from `events.consolidation`, `ORG_MEMORY_QUERY_START` from `events.org_memory`, `API_REQUEST_STARTED` from `events.api`, `API_ROUTE_NOT_FOUND` from `events.api`, `API_COORDINATION_STARTED` from `events.api`, `API_COORDINATION_COMPLETED` from `events.api`, `API_COORDINATION_FAILED` from `events.api`, `API_COORDINATION_AGENT_RESOLVE_FAILED` from `events.api`, `CODE_RUNNER_EXECUTE_START` from `events.code_runner`, `DOCKER_EXECUTE_START` from `events.docker`, `MCP_INVOKE_START` from `events.mcp`, `SECURITY_EVALUATE_START` from `events.security`, `HR_HIRING_REQUEST_CREATED` from `events.hr`, `PERF_METRIC_RECORDED` from `events.performance`, `TRUST_EVALUATE_START` from `events.trust`, `PROMOTION_EVALUATE_START` from `events.promotion`, `PROMPT_BUILD_START` from `events.prompt`, `MEMORY_RETRIEVAL_START` from `events.memory`, `MEMORY_BACKEND_CONNECTED` from `events.memory`, `MEMORY_ENTRY_STORED` from `events.memory`, `MEMORY_BACKEND_SYSTEM_ERROR` from `events.memory`, `AUTONOMY_ACTION_AUTO_APPROVED` from `events.autonomy`, `TIMEOUT_POLICY_EVALUATED` from `events.timeout`, `PERSISTENCE_AUDIT_ENTRY_SAVED` from `events.persistence`, `TASK_ENGINE_STARTED` from `events.task_engine`, `COORDINATION_STARTED` from `events.coordination`, `COORDINATION_FACTORY_BUILT` from `events.coordination`, `COMMUNICATION_DISPATCH_START` from `events.communication`, `COMPANY_STARTED` from `events.company`, `CONFIG_LOADED` from `events.config`, `CORRELATION_ID_CREATED` from `events.correlation`, `DECOMPOSITION_STARTED` from `events.decomposition`, `DELEGATION_STARTED` from `events.delegation`, `EXECUTION_LOOP_START` from `events.execution`, `CHECKPOINT_SAVED` from `events.checkpoint`, `PERSISTENCE_CHECKPOINT_SAVED` from `events.persistence`, `GIT_OPERATION_START` from `events.git`, `PARALLEL_GROUP_START` from `events.parallel`, `PERSONALITY_LOADED` from `events.personality`, `QUOTA_CHECKED` from `events.quota`, `ROLE_ASSIGNED` from `events.role`, `ROUTING_STARTED` from `events.routing`, `SANDBOX_EXECUTE_START` from `events.sandbox`, `TASK_CREATED` from `events.task`, `TASK_ASSIGNMENT_STARTED` from `events.task_assignment`, `TASK_ROUTING_STARTED` from `events.task_routing`, `TEMPLATE_LOADED` from `events.template`, `TOOL_INVOKE_START` from `events.tool`, `TOOL_OUTPUT_WITHHELD` from `events.tool`, `WORKSPACE_CREATED` from `events.workspace`, `APPROVAL_GATE_ESCALATION_DETECTED` from `events.approval_gate`, `APPROVAL_GATE_ESCALATION_FAILED` from `events.approval_gate`, `APPROVAL_GATE_INITIALIZED` from `events.approval_gate`, `APPROVAL_GATE_RISK_CLASSIFIED` from `events.approval_gate`, `APPROVAL_GATE_RISK_CLASSIFY_FAILED` from `events.approval_gate`, `APPROVAL_GATE_CONTEXT_PARKED` from `events.approval_gate`, `APPROVAL_GATE_CONTEXT_PARK_FAILED` from `events.approval_gate`, `APPROVAL_GATE_PARK_TASKLESS` from `events.approval_gate`, `APPROVAL_GATE_RESUME_STARTED` from `events.approval_gate`, `APPROVAL_GATE_CONTEXT_RESUMED` from `events.approval_gate`, `APPROVAL_GATE_RESUME_FAILED` from `events.approval_gate`, `APPROVAL_GATE_RESUME_DELETE_FAILED` from `events.approval_gate`, `APPROVAL_GATE_RESUME_TRIGGERED` from `events.approval_gate`, `APPROVAL_GATE_NO_PARKED_CONTEXT` from `events.approval_gate`, `APPROVAL_GATE_LOOP_WIRING_WARNING` from `events.approval_gate`). Import directly: `from ai_company.observability.events.<domain> import EVENT_CONSTANT`
 - **Structured kwargs**: always `logger.info(EVENT, key=value)` — never `logger.info("msg %s", val)`
 - **All error paths** must log at WARNING or ERROR with context before raising
 - **All state transitions** must log at INFO

docs/design/engine.md

@@ -408,6 +408,9 @@ All loop implementations satisfy the `ExecutionLoop` runtime-checkable protocol:
 composes the execution loop with prompt construction, context management, tool
 invocation, and cost tracking into a single `run()` call.
 
+The engine also exposes an optional ``coordinate()`` method that delegates to a
+``MultiAgentCoordinator`` when one is configured (see :doc:`coordination`).
+
 **Signature:**
 
 ```python
@@ -787,6 +790,10 @@ coordination:
     parallel_default: "centralized"
     # mixed tasks -> SAS backbone for sequential phases, delegates parallel sub-tasks
     mixed_default: "context_dependent"  # hybrid: not a single topology -- engine selects per-phase
+  max_concurrency_per_wave: null        # None = unlimited
+  fail_fast: false
+  enable_workspace_isolation: true
+  base_branch: main
 ```
 
 The auto-selector uses task structure, artifact count, and (when available from

docs/design/operations.md

@@ -964,6 +964,7 @@ future CLI tool are thin clients that call the API -- they contain no business l
 | `/api/v1/departments` | Department management |
 | `/api/v1/projects` | Project CRUD |
 | `/api/v1/tasks` | Task management |
+| `POST /api/v1/tasks/{task_id}/coordinate` | Trigger multi-agent coordination |
 | `/api/v1/messages` | Communication log |
 | `/api/v1/meetings` | Schedule, view meeting outputs |
 | `/api/v1/artifacts` | Browse produced artifacts (code, docs, etc.) |

docs/roadmap/index.md

@@ -6,7 +6,7 @@ The SynthOrg core framework is complete. The following subsystems are built and
 
 - Provider abstraction layer (LiteLLM adapter, routing, resilience)
 - Budget and cost management (tracking, enforcement, CFO optimization, quotas)
-- Agent engine (execution loops, parallel execution, task decomposition, routing, assignment, recovery, shutdown)
+- Agent engine (execution loops, parallel execution, task decomposition, routing, assignment, recovery, shutdown, multi-agent coordination)
 - Communication layer (message bus, delegation, loop prevention, conflict resolution, meeting protocol)
 - Memory system (pluggable backend protocol, Mem0 adapter, retrieval pipeline, shared org memory, consolidation)
 - Security and approval system (rule engine, output scanning, progressive trust, autonomy levels, timeout policies)

src/ai_company/api/app.py

@@ -35,7 +35,9 @@
 from ai_company.communication.bus_protocol import MessageBus  # noqa: TC001
 from ai_company.config.schema import RootConfig
 from ai_company.core.approval import ApprovalItem  # noqa: TC001
+from ai_company.engine.coordination.service import MultiAgentCoordinator  # noqa: TC001
 from ai_company.engine.task_engine import TaskEngine  # noqa: TC001
+from ai_company.hr.registry import AgentRegistryService  # noqa: TC001
 from ai_company.observability import get_logger
 from ai_company.observability.events.api import (
     API_APP_SHUTDOWN,
@@ -355,6 +357,8 @@ def create_app(  # noqa: PLR0913
     approval_store: ApprovalStore | None = None,
     auth_service: AuthService | None = None,
     task_engine: TaskEngine | None = None,
+    coordinator: MultiAgentCoordinator | None = None,
+    agent_registry: AgentRegistryService | None = None,
 ) -> Litestar:
     """Create and configure the Litestar application.
 
@@ -369,6 +373,8 @@ def create_app(  # noqa: PLR0913
         approval_store: Approval queue store.
         auth_service: Pre-built auth service (for testing).
         task_engine: Centralized task state engine.
+        coordinator: Multi-agent coordinator.
+        agent_registry: Agent registry service.
 
     Returns:
         Configured Litestar application.
@@ -403,6 +409,8 @@ def create_app(  # noqa: PLR0913
         approval_store=effective_approval_store,
         auth_service=auth_service,
         task_engine=task_engine,
+        coordinator=coordinator,
+        agent_registry=agent_registry,
         startup_time=time.monotonic(),
     )
 

src/ai_company/api/channels.py

@@ -4,6 +4,8 @@
 creates the Litestar ``ChannelsPlugin`` with an in-memory backend.
 """
 
+from typing import TYPE_CHECKING, Any
+
 from litestar.channels import ChannelsPlugin
 from litestar.channels.backends.memory import MemoryChannelsBackend
 
@@ -24,6 +26,27 @@
 )
 
 
+if TYPE_CHECKING:
+    from litestar import Request
+
+
+def get_channels_plugin(
+    request: Request[Any, Any, Any],
+) -> ChannelsPlugin | None:
+    """Extract the ``ChannelsPlugin`` from the application, or ``None``.
+
+    Args:
+        request: The incoming Litestar request.
+
+    Returns:
+        The ``ChannelsPlugin`` instance if registered, otherwise ``None``.
+    """
+    for plugin in request.app.plugins:
+        if isinstance(plugin, ChannelsPlugin):
+            return plugin
+    return None
+
+
 def create_channels_plugin() -> ChannelsPlugin:
     """Create the channels plugin with in-memory backend.
 

src/ai_company/api/controllers/__init__.py

@@ -10,6 +10,7 @@
 from ai_company.api.controllers.autonomy import AutonomyController
 from ai_company.api.controllers.budget import BudgetController
 from ai_company.api.controllers.company import CompanyController
+from ai_company.api.controllers.coordination import CoordinationController
 from ai_company.api.controllers.departments import DepartmentController
 from ai_company.api.controllers.health import HealthController
 from ai_company.api.controllers.meetings import MeetingController
@@ -35,6 +36,7 @@
     ApprovalsController,
     AutonomyController,
     AuthController,
+    CoordinationController,
 )
 
 __all__ = [
@@ -48,6 +50,7 @@
     "BudgetController",
     "CompanyController",
     "Controller",
+    "CoordinationController",
     "DepartmentController",
     "HealthController",
     "MeetingController",

src/ai_company/api/controllers/approvals.py

@@ -5,11 +5,11 @@
 from uuid import uuid4
 
 from litestar import Controller, Request, get, post
-from litestar.channels import ChannelsPlugin
+from litestar.channels import ChannelsPlugin  # noqa: TC002
 from litestar.datastructures import State  # noqa: TC002
 
 from ai_company.api.auth.models import AuthenticatedUser
-from ai_company.api.channels import CHANNEL_APPROVALS
+from ai_company.api.channels import CHANNEL_APPROVALS, get_channels_plugin
 from ai_company.api.dto import (
     ApiResponse,
     ApproveRequest,
@@ -44,7 +44,7 @@
 logger = get_logger(__name__)
 
 
-def _get_channels_plugin(
+def _require_channels_plugin(
     request: Request[Any, Any, Any],
 ) -> ChannelsPlugin:
     """Extract the ChannelsPlugin from the application.
@@ -58,12 +58,12 @@ def _get_channels_plugin(
     Raises:
         RuntimeError: If no ChannelsPlugin is registered on the app.
     """
-    for plugin in request.app.plugins:
-        if isinstance(plugin, ChannelsPlugin):
-            return plugin
-    msg = "ChannelsPlugin not registered"
-    logger.error(API_APPROVAL_PUBLISH_FAILED, error=msg)
-    raise RuntimeError(msg)
+    plugin = get_channels_plugin(request)
+    if plugin is None:
+        msg = "ChannelsPlugin not registered"
+        logger.error(API_APPROVAL_PUBLISH_FAILED, error=msg)
+        raise RuntimeError(msg)
+    return plugin
 
 
 def _publish_approval_event(
@@ -93,7 +93,7 @@ def _publish_approval_event(
         },
     )
     try:
-        channels_plugin = _get_channels_plugin(request)
+        channels_plugin = _require_channels_plugin(request)
         channels_plugin.publish(
             event.model_dump_json(),
             channels=[CHANNEL_APPROVALS],