The growing automation of transport and the development of Vehicular Ad Hoc Networks (VANETs) demand intelligent solutions for the secure and efficient management of vehicular context. This work presents a hybrid architecture that combines the Model Context Protocol (MCP) with Hyperledger Fabric blockchain technology, integrating semantic representation capabilities with distributed mechanisms for validation and persistence. The proposal is implemented in a simulation environment based on NS-3 and SUMO, and its performance is evaluated in urban scenarios with different traffic densities. Context inconsistency detection algorithms based on fuzzy logic are introduced to optimize decision-making. Results show significant improvements in data integrity, traceability, and resilience to malicious attacks, with a controlled impact on latency and communication overhead. This approach opens new avenues for integrating artificial intelligence into vehicular context management, with potential applications in smart cities and autonomous driving. The article is submitted to DiTTEt 2025, under the theme of Artificial Intelligence and Disruptive Technologies applied to intelligent transportation.