Results for “llm-evaluation”

Production-tested commands, skills, and workflow patterns for Claude Code. Developed through 6+ months of daily use. Inc

一个生产级的深度研究 Agent 系统，从零构建多智能体编排、Red-Blue 对抗降噪、 语义级上下文压缩、跨 Agent 共享记忆四大核心能力，配套 165 次独立实验 + Bootstrap 统计显著性检验的完整评测体系。