This is the first study aiming to test two universality claims of attachment theory within a rural Andean sample from Cusco, Peru. A total of 69 mothers and their children (6 to 36 months) participated. Child attachment security was assessed with the Attachment Q-set (AQS), maternal sensitivity was measured during three naturalistic episodes (free interaction, bathing, and feeding) with the Ainsworth sensitivity scale and the Maternal Behavior Q-sort (MBQS), and a cumulative maternal risk variable was calculated. Results revealed that most children displayed less characteristic secure base behaviors in the interactions with their mothers, compared to other reference samples. Furthermore, an association between maternal sensitivity and child attachment security was found, and a negative relation between maternal sensitivity and the cumulative risk variable. These results support some of the attachment theory’s universality claims, and suggest new avenues for research on assessment issues in rural samples in the Global South.