This paper proposes a decentralized two-stage optimization model that enables strategic and risk-averse participation of renewable generation companies (RGCs) while ensuring data privacy and minimizing system balancing costs. In the first stage, each RGC independently determines its optimal participation schedule in the day-ahead market, utilizing a Conditional Value-at-Risk (CVaR) formulation to manage risks arising from generation uncertainty. In the second stage, the distribution system operator (DSO) clears the day-ahead market to maximize social welfare under DistFlow-based network security constraints, followed by the execution of an intraday balancing market to compensate for real-time deviations. In this intraday market, stochastic fluctuations in generation and demand trigger the activation of distributed flexibility resources including energy storage systems, flexible loads, and vehicle-to-grid (V2G) services to maintain network balance. Additionally, the proposed model incorporates an insurance-based financial mechanism designed to cover a portion of the RGCs' imbalance penalties. Case studies on a modified IEEE 69-bus network demonstrate that the proposed framework can reduce RGC imbalance penalties by up to 54.37% and increase their net profits by 25.35%. Furthermore, the application of the proposed framework can lead to a reduction in the DSO’s intraday balancing costs by up to 19.53% in certain operating scenarios, while the average cost reduction across all studied scenarios is estimated to exceed 8.62%. These results indicate that the proposed framework integrates risk management, decentralized coordination, and demand-side flexibility into a unified structure, providing an efficient platform for the integration of distributed renewable resources.