Skip to main content

⚙️ Settings

The Settings module provides configuration functionality for the Wegent system, including management of Agents, Models, Shells, and Skills.

📋 Documentation in This Module​

DocumentDescription
Agent SettingsConfigure Agents, Bots, prompts, and collaboration modes
Configuring ModelsConfigure AI models (Anthropic Claude, OpenAI GPT, etc.)
Configuring ShellsConfigure runtime environments (ClaudeCode, Agno, Dify)
Managing SkillsUpload, manage, and use Skill capability extension packages

🎯 Core Configuration​

Agent Settings​

Configure the core components of AI Agents:

Agent (Team) = Bot(s) + Collaboration Mode
Bot = Shell + Model + Prompt + MCP Tools + Skills

Collaboration Modes:

  • Solo: Single bot working independently
  • Pipeline: Sequential execution, forming a processing pipeline
  • Route: Leader routes content to appropriate experts
  • Coordinate: Leader coordinates parallel execution and aggregates results
  • Collaborate: All members freely discuss and share context

Model Configuration​

Supports multiple AI model providers:

ProviderSupported Models
AnthropicClaude Haiku 4, Claude Sonnet 4, Claude Opus
OpenAIGPT-4, GPT-4 Turbo, GPT-3.5 Turbo

Shell Configuration​

Supports multiple runtime environments:

ShellDescriptionUse Case
ClaudeCodeClaude Code SDK, supports code execution and file operationsCode development, file processing
AgnoAgno framework, supports multiple collaboration modesConversations, multi-agent collaboration
DifyExternal Dify API proxyDify workflow integration
ChatDirect LLM API (no Docker)Lightweight conversations

Skills Management​

Skills are Claude Code capability extension packages:

  • Upload Skills: Package as ZIP file and upload
  • Manage Skills: View, download, update, delete
  • Use Skills: Reference Skills in Bots

🚀 Configuration Workflow​

Recommended configuration order:

  1. Configure Models → Set up AI models and API keys
  2. Configure Shells → Select runtime environment
  3. Upload Skills → Add capability extensions (optional)
  4. Create Bots → Combine models, shells, and prompts
  5. Create Agents → Combine bots and collaboration modes

🔗 Related Resources​