The object of this study is a blog as a form of digital communication that often violate the linguistic security. The subject of the study was the comments of Yandex-Zen blogs about cooking and beauty as a tool for the implementation of verbal abuse. The purpose of the study is to highlight the linguo-conflictogenic components of discourse that violate the linguistic security of participants in digital communication. To achieve this goal, a descriptive method and the method of discursive analysis were used. In the course of linguistic analysis, it was revealed that online media content is characterized by an increased emotions, expressiveness, aggressiveness, sarcastic tone and conflict potential. Markers of aggression are most often invective; pejoratives with semantic dominants of aggression, hostility and irony; graphic selection of words, etc. Within the framework of hating, trolling and elfing, some linguo-conflictogenic elements of discourse that can be avoided in network communication were identified and described. It is concluded that the use of non-normative language units poses a threat to network communication from the linguistic security point of view. Meanwhile, the authors propose some measures to prevent the potential for conflict in digital communication, and outline the prospect of further research in the field of linguistic security of a person on the Internet.