Agents

You can create custom AI agents by merging base traits from a configuration file with user customizations to generate unique agent prompts. The system then produces a prepared prompt, voice mapping, and model preference for spawning agents in parallel.

Trait-driven agent composition dynamically merges base traits and user customizations to generate unique agent personalities. It assigns specific voices and styles to each agent, enabling diverse collaboration without manual prompt engineering for every agent instance.

Yes, you can spawn multiple custom agents in parallel to tackle complex tasks from different perspectives. For example, you can launch three agents with distinct voices to analyze security posture simultaneously for scalable task execution.

Yes, the agent orchestration framework relies on YAML and Handlebars as core dependencies. YAML manages the base traits configuration file, while Handlebars supports the dynamic template composition for generating customized agent prompts and voice mappings.

The best way to assign distinct personalities is through trait-driven voice mapping. By defining traits in a base YAML file and layering user customizations, the system automatically generates prepared prompts and voice mappings without manual prompt engineering.

You should avoid parallel agent orchestration when tasks require strict sequential dependencies or when a single agent perspective is sufficient. The framework is designed for complex tasks needing diverse viewpoints and scalable parallel execution rather than simple linear workflows.