meta-agent-generator

You can automate agent creation by orchestrating the end-to-end lifecycle through multi-source interviews, agent deliberation, and triangulated critique to generate production-ready configurations. Initiate a domain request to trigger requirements gathering, initial generation, and iterative refinements.

Triangulated critique resolves conflicts and builds robust consensus during agent generation by selecting critics via Jaccard similarity. This ensures diverse, high-quality critiques across domain, design, and evaluation perspectives to refine agent configurations.

Distillation and token-budget management keep prompts efficient without sacrificing capability coverage during agent generation. This process refines the initial configuration generated from multi-source interviews into a versioned, auditable output.

Yes, multi-source interviews gather triangulated requirements from domain, design, and evaluation perspectives. This approach applies across domains requiring rapid generation, refinement, and evaluation of agent configurations.

Yes, the agent generator produces versioned, auditable configurations with human approvals. It orchestrates the end-to-end lifecycle from requirements gathering through distillation and evaluation to ensure production readiness.

The best way to select critics is using Jaccard similarity to ensure diverse, high-quality critiques. This automatic selection mechanism supports the deliberation and triangulated critique needed to build robust consensus.