Traditional craftsmanship education relies heavily on face-to-face, hands-on experience transfer between master and apprentice, yet the aging of instructors and complexity of the craft make this "demonstration–observation–correction" teaching model difficult to scale. This study takes Chao embroidery as a case and proposes a "Digital Master" human-machine collaborative teaching system, exploring how technology can mediate collaborative relationships and social learning in craftsmanship inheritance. The system transforms master–apprentice teaching behaviors into a "projected guidance–visual detection–instant feedback" framework, integrating constructivist learning theory with multimodal interaction. Step-by-step projection, height detection, and path correction form a three-in-one teaching interaction model, while operation recording and a learner community connect individual practice with peer support. This approach lowers the entry threshold for Chao embroidery and offers a digital solution for intangible cultural heritage inheritance that emphasizes social interaction and collaborative mechanisms.