aitbc/.windsurf/skills/openclaw-coordination-orchestrator.md

description, title, version

description	title	version
Atomic OpenClaw multi-agent workflow coordination with deterministic outputs	openclaw-coordination-orchestrator	1.0

OpenClaw Coordination Orchestrator

Purpose

Coordinate multi-agent workflows, manage agent task distribution, and orchestrate complex operations across multiple OpenClaw agents.

Activation

Trigger when user requests multi-agent coordination: task distribution, workflow orchestration, agent collaboration, or parallel execution management.

Input

{
  "operation": "distribute|orchestrate|collaborate|monitor",
  "agents": ["agent1", "agent2", "..."],
  "task_type": "analysis|execution|validation|testing",
  "workflow": "string (optional for orchestrate)",
  "parallel": "boolean (optional, default: true)"
}

Output

{
  "summary": "Multi-agent coordination completed successfully",
  "operation": "distribute|orchestrate|collaborate|monitor",
  "agents_assigned": ["agent1", "agent2", "..."],
  "task_distribution": {
    "agent1": "task_description",
    "agent2": "task_description"
  },
  "workflow_status": "active|completed|failed",
  "collaboration_results": {},
  "issues": [],
  "recommendations": [],
  "confidence": 1.0,
  "execution_time": "number",
  "validation_status": "success|partial|failed"
}

Process

1. Analyze

• Validate agent availability
• Check agent connectivity
• Assess task complexity
• Determine optimal distribution strategy

2. Plan

• Select coordination approach
• Define task allocation
• Set execution order
• Plan fallback mechanisms

3. Execute

• Distribute tasks to agents
• Monitor agent progress
• Coordinate inter-agent communication
• Aggregate results

4. Validate

• Verify task completion
• Check result consistency
• Validate workflow integrity
• Confirm agent satisfaction

Constraints

• MUST NOT modify agent configurations without approval
• MUST NOT exceed 120 seconds for complex workflows
• MUST validate agent availability before distribution
• MUST handle agent failures gracefully
• MUST respect agent capacity limits

Environment Assumptions

• OpenClaw agents operational and accessible
• Agent communication channels available
• Task queue system functional
• Agent status monitoring active
• Collaboration protocol established

Error Handling

• Agent offline → Reassign task to available agent
• Task timeout → Retry with different agent
• Communication failure → Use fallback coordination
• Agent capacity exceeded → Queue task for later execution

Example Usage Prompt

Orchestrate parallel analysis workflow across main and trading agents

Expected Output Example

{
  "summary": "Multi-agent workflow orchestrated successfully across 2 agents",
  "operation": "orchestrate",
  "agents_assigned": ["main", "trading"],
  "task_distribution": {
    "main": "Analyze blockchain state and transaction patterns",
    "trading": "Analyze marketplace pricing and order flow"
  },
  "workflow_status": "completed",
  "collaboration_results": {
    "main": {"status": "completed", "result": "analysis_complete"},
    "trading": {"status": "completed", "result": "analysis_complete"}
  },
  "issues": [],
  "recommendations": ["Consider adding GPU agent for compute-intensive analysis"],
  "confidence": 1.0,
  "execution_time": 45.2,
  "validation_status": "success"
}

Model Routing Suggestion

Reasoning Model (Claude Sonnet, GPT-4)

• Complex workflow orchestration
• Task distribution strategy
• Agent capacity planning
• Collaboration protocol management

Performance Notes

• Execution Time: 10-60 seconds for distribution, 30-120 seconds for complex workflows
• Memory Usage: <200MB for coordination operations
• Network Requirements: Agent communication channels
• Concurrency: Safe for multiple parallel workflows