2c1acda2f3
- Update DQN trainer with gradient collapse detection warmup - Add portfolio tracker improvements - Include hyperopt trial results (multiple Sharpe ratio experiments) - Add new test files for action/position sign convention, early stopping, cash reserve bugs, and portfolio execution - Update trained model files - Add Claude Code configuration and skills 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude <noreply@anthropic.com>
1.5 KiB
1.5 KiB
Automatic Topology Selection
Purpose
Automatically select the optimal swarm topology based on task complexity analysis.
How It Works
1. Task Analysis
The system analyzes your task description to determine:
- Complexity level (simple/medium/complex)
- Required agent types
- Estimated duration
- Resource requirements
2. Topology Selection
Based on analysis, it selects:
- Star: For simple, centralized tasks
- Mesh: For medium complexity with flexibility needs
- Hierarchical: For complex tasks requiring structure
- Ring: For sequential processing workflows
3. Example Usage
Simple Task:
Tool: mcp__claude-flow__task_orchestrate
Parameters: {"task": "Fix typo in README.md"}
Result: Automatically uses star topology with single agent
Complex Task:
Tool: mcp__claude-flow__task_orchestrate
Parameters: {"task": "Refactor authentication system with JWT, add tests, update documentation"}
Result: Automatically uses hierarchical topology with architect, coder, and tester agents
Benefits
- 🎯 Optimal performance for each task type
- 🤖 Automatic agent assignment
- ⚡ Reduced setup time
- 📊 Better resource utilization
Hook Configuration
The pre-task hook automatically handles topology selection:
{
"command": "npx claude-flow hook pre-task --optimize-topology"
}
Direct Optimization
Tool: mcp__claude-flow__topology_optimize
Parameters: {"swarmId": "current"}
CLI Usage
# Auto-optimize topology via CLI
npx claude-flow optimize topology