Automatic deep research skill for Claude Code and AI agents. Multi-round search with multi-source verification, outputs structured reports. 自动深度研究技能。给 Claude Code 等 AI Agent 用的研究助手，通过多轮搜索和多源验证，输出结构化报告。

Production-grade Claude Code skills, auto-research validated. Includes /deep-research with 89-run benchmark.

Deep research skill using Kimi CLI with structured reports, token tracking, and auto-retry