Smart contracts, due to their decentralized nature and immutability, have demonstrated significant potential in various sectors such as finance and supply chains. However, as their applications have expanded, their security vulnerabilities have increasingly come to light. In response to the frequent occurrences of smart contract attacks, researchers have undertaken a series of studies, including the development of automated vulnerability detection tools, dynamic monitoring techniques, and vulnerability remediation strategies. However, these detection tools often rely on static analysis and fail to capture dynamic vulnerabilities that occur during runtime. Additionally, dynamic monitoring is limited by the ability to acquire and process real-time data, and it lacks the capability to globally analyze security risks during phased attacks. Similarly, mechanisms for updating vulnerabilities might inadvertently introduce new risks by leaving backdoors in smart contracts. To address these issues, this paper proposes a new method for assessing the reputation and evaluating the risks of smart contracts, aimed at enhancing the security of the blockchain ecosystem through the analysis of smart contract invocation trajectories. The effectiveness and practicality of this method have been validated by evaluating real attack cases that have occurred on-chain. Finally, the paper summarizes the research findings and explores potential future research directions, aiming to provide new insights and solutions for the field of smart contract security.