Recently, there has been a surging interest in meaning-making processes in Chinese social media. However, most studies focus on the current contexts, ignoring the process of entextualization, that is, how semiotic resources have been decontextualized from the original/prior contexts and recontextualized in the current Chinese social media context. This paper intends to fill this gap by studying an example of entextualization coupled with modal shift. Specifically, an oral warning - initially a line frequently uttered in a Japanese animation serial - has been reshaped and reused to type comments in various video scenes and for various meanings on a Chinese video-sharing website Bilibili (https://www.bilibili.com/). Drawing on selective notions in computer-mediated communication and social semiotic multimodality, this paper analyses a set of these comments, which we call high energy texts, focusing on their multimodal quality and the meaning-making processes involved. We suggest that new possibilities are generated during the modal shift (i.e. from speech to writing); in particular, warning acts carried out through high energy texts have been transformed into a collective social action that is recognized and understood by Bilibili users, the participation of which enables the users to build alignment and enact different identities. Our paper showcases how a social semiotic multimodal perspective can enrich our understanding of the semiosis in Chinese social media.