The traditional account is straightforward. Wild palm civets eat ripe coffee cherries, digestive enzymes act on the beans during passage, and the beans are recovered from droppings, washed and roasted. The animal's preference for ripe fruit did part of the sorting, and supply was limited by how much civets happened to eat.
Commercial demand broke that constraint. Investigations by welfare organisations and journalists in Indonesia and the Philippines have documented wild-caught civets kept in small barren cages on a near-exclusive coffee cherry diet, showing stereotypic behaviour, injuries and poor condition. Under caging the animals eat what they are given, so the ripeness selection that supposedly justified the product no longer operates. Certification schemes exist, but the supply chain is hard to verify and beans of unclear origin are widely sold as wild-collected.
Whether the digestion improves the coffee at all is a separate and unsettled question. Some chemical work reports differences in protein breakdown and volatile profile, and blind tastings have not consistently favoured the passed beans. The species remains common and is listed as Least Concern, so the issue is welfare and consumer fraud, not extinction risk.