Did the Cyrillic letters for the sounds Greek lacked come from Glagolitic, from a Semitic script, or from neither?

Roughly a dozen Early Cyrillic letters have no Greek model. The majority reading derives them from Glagolitic, which is older and already fitted to Slavonic. A second reading compares sha, tsy and zhivete with Hebrew or Samaritan letters, an argument made since the nineteenth century. A third holds that shapes this simple cannot support either claim. No exemplar of either alphabet survives from the ninth century, when the choices were made, so every argument runs on shape comparison against later manuscripts.