Virtual collaborations bring together people who must work together despite having varied access to and understanding of the work at hand. In many cases, the collaborations are technology supported, meaning that the work is done through shared documents. We develop a framework articulating the characteristics of documents supporting collaborators with access to asymmetric knowledge versus those with access to symmetric knowledge. Drawing on theories about document genre, boundary objects, and provenance, we hypothesize that documents supporting asymmetric collaborators are likely to articulate or prescribe their own (a) purpose, (b) context of use, (c) content and form, and (d) provenance in greater detail than documents supporting symmetric collaborators. We explore these hypotheses through content analysis of documents and instructions for documents from a variety of free/libre open‐source projects (FLOSS). We present findings consistent with the hypotheses developed as well as results extending beyond our theory‐derived assumptions. When participants have access to the same knowledge, the study suggests that prescriptions about the content of documents become less important compared with prescriptions about the context, provenance, and process of work. The study contributes with a dynamic perspective on communicative practices that consider an often‐uneven distribution of knowledge in virtual collaborations.