Hallucination
When an LLM emits text that sounds plausible but is factually incorrect or unsupported by its inputs. In coding, hallucinations show up as invented function signatures, fictitious library APIs, or non-existent CLI flags. The single biggest reliability risk in vibe-coding workflows.
Background
Hallucinations result from the model's training objective (predict plausible text) not aligning perfectly with truth. They worsen at long context, with under-specified prompts, and on niche libraries. Mitigations: RAG to ground the model in real source, tool use so the model can verify by running code, structured output so the model can't fabricate types, and human-in-the-loop review for high-stakes changes. Vibe coding is most reliable when paired with fast feedback loops (tests, type checkers) that surface hallucinations within seconds.