{"skill":{"slug":"deep-topic-research","displayName":"深度话题调研工作流","summary":"对用户提出的具体话题/课题/行业进行深度调研，输出结构化报告。结合多智能体并行搜索（借鉴女娲Nuwa）和克制调用原则（借鉴AutoGLM DeepResearch），先展示中间发现再决策。","tags":{"deep-research":"1.0.0","industry-analysis":"1.0.0","latest":"1.0.0","research":"1.0.0","topic":"1.0.0","workflow":"1.0.0"},"stats":{"comments":0,"downloads":162,"installsAllTime":0,"installsCurrent":0,"stars":0,"versions":1},"createdAt":1777895678174,"updatedAt":1778492846193},"latestVersion":{"version":"1.0.0","createdAt":1777895678174,"changelog":"deep-topic-research v1.0.0\n\n- Initial release of a structured deep topic/industry research workflow.\n- Implements parallel multi-agent search, controlled information collection, and staged discovery-feedback loops.\n- Focuses on 2-4 key research dimensions, prioritizing credible sources and avoiding unreliable platforms (e.g., Zhihu, Baidu Knows).\n- Requires presentation of interim findings before proceeding, allowing user involvement in research direction.\n- Enforces rigorous source citation, balanced perspectives, and actionable recommendations in final reports.\n- Built-in self-check for data support, coverage, and efficiency before report delivery.","license":"MIT-0"},"metadata":null,"owner":{"handle":"realpda","userId":"s17aspeya65fa14fe9v237r5qd863x8n","displayName":"RealGhost","image":"https://avatars.githubusercontent.com/u/60616432?v=4"},"moderation":null}