Orchestrated Task Execution

---

description: Execute task through the Autonomous Orchestration Ecosystem

Orchestrated Task Execution

You MUST follow the orchestrator workflow before executing this task.

Step 1: Read Orchestrator

Read `.claude/agents/orchestrator.md` completely to understand the orchestration logic.

Step 2: Scan Agent Pool

Scan all existing agents:

• .claude/agents/pool/specialized/ - Task-specific agents
• .claude/agents/pool/integrated/ - Merged agents
• .claude/agents/pool/elite/ - High-performing agents

For each agent found, read its definition (.md file).

Step 3: Read Skill Sheets

Read corresponding skill sheets from `.claude/agents/manifests/` to understand:

• Agent capabilities
• Constraints
• Usage metrics
• Synergy hints

Step 4: Calculate Coverage

Compare task requirements against available agent capabilities:

• List required skills for this task
• Calculate match percentage for each agent
• Identify potential synergies between agents

Step 5: Decision Matrix

Coverage Rate	Action
90%+	Use existing agent directly
60-90%	Create integrated agent by merging sources
Below 60%	Create new specialized agent

Step 6: Agent Creation (if needed)

If creating new/integrated agent:

1. Use .claude/agents/_template.md as base
2. Save agent definition to appropriate pool/ subdirectory
3. Create skill sheet in manifests/

Step 7: Execute Task

Execute the following task using the selected/created agent:

$ARGUMENTS

Step 8: Update Metrics

After task completion, update the agent's skill sheet in `manifests/`:

• Increment usage_count
• Update success_rate
• Set last_used to today's date
• Add entry to task_history

Step 9: Elite Promotion Check

If agent meets criteria (usage_count >= 5 AND success_rate >= 0.8):

• Move agent from current location to pool/elite/
• Update skill sheet tier: elite

Step 10: Report

Report:

• Agent used (name, tier, location)
• Action taken (used existing / created new / integrated)
• Coverage achieved
• Any agents created or promoted