We present a novel approach to multiagent planning for self-interested agents. The main idea behind our approach is that multiagent planning systems should be built upon (single-agent) plan repair systems. In our system agents can exchange goals and subgoals through an auction, using their own (planning) heuristics and utility functions to determine when to auction and what to bid. Some experimental results for a logistics domain show that this system can be used to support the coordination of self-interested agents. Categories and Subject Descriptors I.2.8 [Problem Solving, Control Methods, and Search]: Plan execution, formation, and generation; I.2.11 [Distributed Articial Intelligence]: Intelligent agents, Multiagent Systems General Terms Algorithms, Design Keywords Multiagent Planning