In the context of multiple security protocols running in the same environment, we propose a syntactical approach for identifying multi-protocol attacks. The proposed approach uses a canonical security protocol model, where terms that can be verified by protocol participants are denoted by canonical terms. In order to enable the identification of subtle "type-flaw" attacks, where terms can be substituted with other types of terms, we introduce a canonical identifier. The approach is validated by analyzing several security protocol pairs. The attacks discovered by our approach are also discovered by existing security protocol verification tools.