The Violence Detection Lexicon is a carefully curated resource designed to support efficient detection of violent content online. The lexicon contains 140 terms, each assigned a severity weight on a scale of 1-3, capturing different levels of violent intent. Unlike existing approaches that favor exhaustive wordlists, this resource demonstrates that effective violence detection may require fewer, more precisely chosen terms. Through systematic evaluation against a corpus of 27,000 annotated YouTube comments, the lexicon achieved 65% coverage of violent content vocabulary while maintaining computational efficiency. This repository includes: -The core lexicon with weighted terms (CSV) -Validation data including inter-rater reliability assessments -Complete documentation of development methodology and implementation guidelines -Example code for lexicon implementation The resource is designed to support applications ranging from content moderation systems to academic research on violent rhetoric.