A multi-energy virtual power plant can be defined as flexible cooperation of multi-energy distributed energy resources including heating and cooling devices that operates as a single entity and participates in the power markets. To exploit the potential of multi-energy distributed resources, this paper proposes a model for the aggregation of a multi-energy virtual power plant that participates in day-ahead energy and reserve markets. The reserve capacity brought by the multi-energy quick-start devices is considered in this paper, which adds binary variables to the recourse problem. The proposed model is an adjustable robust optimization problem that guarantees all possible deployment requests can be realized by the virtual power plant. Moreover, the optimal offering strategy is determined, and the total cost of the multi-energy virtual power plant is ensured to not exceed the optimal value of the objective function. A modified nested column-and-constraint generation algorithm is proposed for solving the adjustable robust optimization problem with mixed-integer recourse in specified precision efficiently compared with other KKT-based algorithms. The results from a case study of an actual industrial park demonstrate the satisfactory performance of the proposed model.