Vibe Researching as Wolf Coming: Can AI Agents with Skills Replace or Augment Social Scientists?

Researcher Yongjun Zhang has published a provocative paper titled 'Vibe Researching as Wolf Coming: Can AI Agents with Skills Replace or Augment Social Scientists?' on arXiv, introducing the concept of 'vibe researching' as the research parallel to 'vibe coding.' The paper argues that modern AI agents—systems with persistent state, tool access, and specialist skills—represent a fundamental shift from prior automation. Unlike simple chatbots, these agents can autonomously execute multi-step research workflows, reading files, running code, querying databases, and searching the web. Zhang uses the 'scholar-skill' plugin, a suite of 21 specialized skills for Claude Code that covers the entire research pipeline from idea generation to paper submission, as a concrete case study to illustrate this new capability.

Zhang develops a cognitive task framework that classifies research activities along dimensions of codifiability and tacit knowledge, identifying a delegation boundary that cuts *through* every research stage, not between them. The analysis concludes that while AI agents excel at speed, coverage, and methodological scaffolding, they currently struggle with theoretical originality and leveraging tacit field knowledge. The paper warns of three major implications for the social science profession: augmentation with fragile conditions, stratification risk, and a pedagogical crisis. It concludes by proposing five principles for responsible 'vibe researching' to guide the integration of these powerful AI agents into academic work.