Large Language Models (LLMs) are increasingly utilized to automate software engineering tasks, such as generating repository-level documentation. While Multi-Agent Systems (MAS) are often adopted under the premise that task decomposition improves performance, the impact of architectural complexity on practical efficiency remains under-examined. This study empirically evaluates Retrieval-Augmented Generation (RAG) dependent architectures for the generation of README files for GitHub repositories. A systematic comparison is conducted between a Single-Agent pipeline, a specialized MAS, and a developer-guided planning (Dev-Plan) variant, benchmarked against LARCH, a state-of-the-art baseline, and the original ground truth. Results indicate that our proposed RAG architectures significantly outperform LARCH. A critical architectural trade-off is identified: the Single-Agent pipeline achieves lexical quality comparable to MAS while reducing token consumption by 86% and operating at twice the speed. In contrast, manual taxonomy analysis demonstrates that MAS achieves high structural consistency (98%), resolving the formatting issues observed in single-agent approaches. Autonomous planning is identified as the primary pipeline bottleneck; incorporating lightweight developer-guided plans (Dev-Plan) eliminates structural failures and produces the highest overall documentation quality, surpassing both autonomous AI, baseline, and ground-truth.