What problem does it solve? Choosing the right multi-agent architecture, writing valid tool schemas, and diagnosing agent system bottlenecks are error-prone when done by hand. This Skill replaces guesswork with three deterministic scripts that score architecture patterns from requirements, validate tool schemas for Anthropic and OpenAI formats, and analyze execution logs for cost, latency, and failure bottlenecks. ## Core Features & Use Cases - Architecture planning: Feed a requirements JSON (goal, tasks, constraints, team size) into the planner to get a scored pattern choice (supervisor, swarm, pipeline, hierarchical), agent roles, communication links, a mermaid diagram, and an implementation roadmap. - Tool schema generation: Convert plain-language tool descriptions into validated provider-ready schemas, emitting both Anthropic and OpenAI formats with a hard gate that every schema must pass validation. - Execution log evaluation: Analyze agent run logs for success rate, latency distribution, cost breakdown, SLA compliance, bottleneck analysis, and prioritized optimization recommendations, with a verification loop that requires zero critical issues. - Use Case: You need to build a research automation system with several cooperating agents. Run the planner on your requirements to pick a supervisor pattern, generate and validate tool schemas for each agent, then evaluate a pilot run's logs and iterate until no critical issues remain. ## Quick Start Ask the agent to design a multi-agent architecture from your requirements, then generate validated tool schemas and evaluate a pilot run's execution logs for bottlenecks.
