Increased predation where ground cover is reduced after severe wildfire is increasingly implicated as a factor causing decline of vulnerable prey populations. In arid central Australia, one species detrimentally affected by repeated wildfire is the great desert skink or tjakura (Liopholis kintorei), a distinctive lizard of the central Australian arid zone that constructs and inhabits multi-entranced communal burrows. We aimed to test whether tjakura or predator activity at burrow entrances varied with cover and how tjakura respond to predator presence. Using time-lapse photography, we monitored tjakura and predator activity at the largest entrance of 12 burrows ranging from high (>70%) to low (<50%) cover and at multiple entrances of two other burrows. Overall activity did not vary between burrows with high and low cover. Within burrow systems tjakura were more active at sparsely vegetated entrances, often sitting wholly or partly inside the burrow. However, consistent between and within burrow systems, skinks spent proportionally more time fully outside where cover was higher. Predators—mostly native—were detected at most burrows, with no apparent relationship between predator activity and cover. Skinks also did not appear to modify their activity in response to predator visits. Our results indicate that tjakura may spend more time outside burrow entrances when cover is higher but there was no direct evidence that this related to perceived or real predation risk. Differences in food availability, thermoregulatory opportunities and opportunities for ambush foraging associated with differences in vegetation cover or composition are other factors likely to be important in determining the activity of tjakura at burrows. Our research demonstrates the usefulness of camera traps for behavioural studies of ectothermic burrowing animals. The complex relationships between tjakura activity and vegetation cover thereby revealed, suggest outcomes of fire-mediated habitat change on predator–prey interactions are not easily predictable.