Autonomous supervisor → worker → critic loop for Claude Code — delegate one goal, an independent critic grades each round against a rubric, re-dispatch until done. 主Agent监督子Agent+独立评委，定验收标准反复迭代到完美，全程不问你。
10 hands-on demos covering every major agentic workflow design pattern — built with the OpenAI Agents SDK
Multi-agent content pipeline that loops generator, evaluator, and refiner agents until output passes your voice rules. Built with Claude Agent SDK. Forkable for any voice.
A reusable, domain-agnostic self-improving loop for Claude Code — an adversarially-gated evaluator–optimizer that runs unattended without fooling itself or wrecking the machine.
A production grade financial terminal interface designed for natural language climate intelligence, bond model interrogation, automated underwriting audits, and portfolio hedging generation. Built with a Bloomberg inspired design system optimized for risk analysis workflows.
AI multi-agent system that auto-fixes data quality using an evaluator-optimizer pattern. Arabic-first. Built with AI-assisted development.
Add a description, image, and links to the evaluator-optimizer topic page so that developers can more easily learn about it.
To associate your repository with the evaluator-optimizer topic, visit your repo's landing page and select "manage topics."