Проектирование многоагентных AI-процессов

★ 6.9 · marketing

multi-agent-workflow is a Claude Code skill that designs and documents a multi-agent workflow end-to-end: it reads project context from memory.md and knowledge-base.md, selects the most suitable framework — AI Readiness Assessment, Human-in-the-Loop Design, Agent Orchestration Patterns, RAG Architecture, or Responsible AI Framework — and produces a structured document complete with executive summary, implementation table, risk matrix, and success metrics. A built-in quality checklist verifies that recommendations are specific and actionable, figures are realistic, and the output matches the required format. Core KPIs tracked include Time Saved Per Task, Automation Rate, Error Reduction %, and Cost Per AI Operation. The skill is aimed at marketing and automation teams that need a ready-to-review workflow document for stakeholder alignment and repeatable AI task automation.