Team leadership, task prioritization, strategy selection, and fleet coordination for multi-agent systems.
- Priority queues — weighted task prioritization with deadline tracking and dynamic reordering
- Strategy engine — select and apply coordination strategies (sequential, parallel, adaptive) based on task requirements
- Fleet coordination — assign vessels to tasks, track availability, balance workloads
- Leadership styles — configurable leadership modes (directive, collaborative, delegative) that affect task distribution
- Status reports — structured fleet health reports with metrics, health status, and trend data
pip install captainfromcaptainimportCaptain, LeadershipStyle, Task, Priority# Create a captain with a leadership stylecaptain=Captain(name="fleet-captain", style=LeadershipStyle.ADAPTIVE)
# Add fleet vesselsfromcaptainimportFleetVessel, FleetCoordinationcoord=FleetCoordination()
coord.register(FleetVessel(id="agent-1", capabilities=["rust", "benchmarking"]))
coord.register(FleetVessel(id="agent-2", capabilities=["python", "docs"]))
# Create and prioritize taskstask=Task(title="Rewrite README", priority=Priority.HIGH, deadline="2026-06-01")
captain.assign(task, fleet=coord)
# Run strategyfromcaptainimportStrategyEngine, Strategy, ResourceBudgetengine=StrategyEngine()
plan=engine.plan(tasks=captain.pending_tasks, fleet=coord, budget=ResourceBudget(max_agents=5))
print(plan)
# Get status reportreport=captain.report()
print(f"Fleet health: {report.health}")
print(f"Tasks in progress: {report.metrics.tasks_in_progress}")- OpenConstruct Documentation — ecosystem-wide docs and guides
The commanding vessel of the SuperInstance fleet. Receives dispatched work from Co-Captain and distributes it across fleet agents.
- co-captain-git-agent — Human liaison (dispatches to captain)
- cocapn-com — Inter-agent messaging
- cluster-orchestrator — Cluster-level orchestration
- fleet-health-monitor — Fleet health daemon
pytest tests/pip install captainPython 3.10+. MIT license.