Crowdsourcing is a new trend that uses the wisdom of crowds on the Internet to solve certain problems that need vast amounts of human resources. There have been a number of crowdsourcing platforms developed for various domains. However, the landscape of crowdsourcing platforms is widely dispersed and most tasks remain hidden. Finding out the tasks closely matching contributors' personal preference and capabilities is difficult. In this research, we aim to design a social mechanism for task-oriented crowdsourcing recommendations, which the requesters can easily use to find suitable contributors who are also very willing to finish their tasks. Our experimental results show that the proposed mechanism is effective in identifying appropriate contributors with respect to different types of task. The proposed mechanism can help to establish a long-term partnership that brings value to the participants of task-oriented crowdsourcing.