<p><?tk 4?>Developing computational thinking (CT) competencies, comprising both CT skills and CT attitudes, is essential for individuals to understand, navigate, and effectively engage with technologies in today’s digital society. Moreover, K–12 is critical in fostering CT competencies, as students develop foundational cognitive and affective skills at this stage that influence their ability to engage with technology and problem-solving abilities. Although both CT components matter, previous studies have often focused on assessing general CT competency in K–12 education, limited attention has been paid to assessing specific CT skills, including CT concepts (e.g., control structures, parallelism, operators, representation, and functions), and CT practices (e.g., testing/debugging, abstraction, decomposition, algorithmic thinking, generalization, pattern recognition, and spatial reasoning). To address this gap, the present study systematically reviewed 58 CT skill assessment studies in K–12 education published between 2020 and 2024, following the PRISMA protocol. The reviewed studies were identified from four databases: Education Research Information Center (ERIC), PsycINFO, Scopus, and Web of Science. Results revealed that most studies: (1) were conducted in the USA, (2) primarily sampled elementary students, (3) focused on programming and computer science disciplines, (4) targeted decomposition, (5) found that pattern recognition frequently co-occurred with abstraction, decomposition, and algorithmic thinking, (6) employed traditional knowledge tests, and (7) reported reliability and validity evidence. This review not only synthesizes recent trends in CT skill assessments but also provides actionable guidance for researchers and policymakers in designing, implementing, and validating CT skill assessments to enhance teaching and learning practices in K–12 education.</p>