Most violence risk assessment scales were originally developed for use in forensic settings at the time of discharge or release of patients into the community after long-term treatment. However, there is a considerable need for specialized, brief and structured risk assessment tools to inform risk decisions in short-term psychiatric treatment. The present study reports on research findings from the development and implementation of the violence risk screening-10 (V-RISK-10) in two acute psychiatric settings in Norway. The 10-item screen is easy to use, time-saving and may be used for screening of violence risk during hospital stay and after discharge into the community. Prospective validation studies of the screen concerning inpatient and post-release community violence have been conducted. Although data analyses are not yet complete, preliminary findings indicate that the screen has good predictive validity. This suggests that the screen is a promising tool in short-term acute psychiatric settings. However, the importance of reliability in mental health data and tests is well recognized, and a screen with good predictive validity is not worth much if clinicians are unable to agree on the scoring of one and the same patient. In this article we report results from a naturalistic interrater reliability investigation that involved 25 mental health professionals and 73 acute psychiatric patients. V-RISK-10 scoring was accomplished by two raters for each patient. The interrater reliability value for total scores was acceptable. Variations pertaining to the individual V-RISK-10 item, patient characteristics and rater characteristics are discussed.