Climate change poses major challenges to Mediterranean dairy farming, where adaptation depends largely on peer-to-peer information exchange among farmers. Yet the structural conditions governing climate adaptation information diffusion remain poorly understood in North African livestock systems. We analyzed the climate adaptation information network of 239 actors reconstructed from interviews with 60 dairy farmers in the Bizerte Governorate, northern Tunisia, using a dual analytical framework that distinguishes nomination asymmetry/ (directed metrics) from social proximity (undirected metrics and bidirectional diffusion analysis). The network was extremely sparse (density = 0.00482), highly unequal in information access (in-degree Gini = 0.797), and strongly geographically segmented. Although the degree distribution was heavy-tailed, formal goodness-of-fit testing did not support a scale-free model. Community detection and four convergent geographic association statistics confirmed that adaptation knowledge is largely confined within administrative zones (81.4% within-zone ties; Cramér’s V = 0.729; p < 0.001). The dual framework revealed only 30% overlap between structural bridge and nomination intermediary rankings. Broker-targeted seeding reached 53.6% of the network versus 31.2% for random seeds (p < 0.001), confirmed by cascade simulations. Brokerage was unrelated to socio-demographic characteristics but associated with greater adaptation information exposure and lower digital dissemination capacity — a structural double bind. Our findings show that the principal obstacle to climate adaptation diffusion is not information scarcity but structural lock-in arising from sparse cross-community connectivity, geographic fragmentation, and a shortage of boundary-spanning actors. We recommend identifying structural brokers through network mapping, equipping them with communication infrastructure, and replacing demographic targeting with network-informed extension strategies.