The document summarizes previous research on attacking white-box cryptography implementations and proposes new research directions. It discusses how white-box cryptography aims to protect secret keys by obscuring them in code, but all academic solutions have been broken. Previous attacks included power analysis, fault injection, code visualization and analysis. The document proposes using taint analysis to trace plaintext/ciphertext through code as an alternative to visualization. It also proposes differential fault analysis and code analysis to attack white-box implementations.