By injecting malicious bytecode into interpreters for VBScript, Python, and Lua, researchers found they can circumvent malicious code detection.