Clever, careful and lethally equipped with razor-sharp sight, smell and hearing.
Stoats are the tiny terminators among New Zealand's hit-list of pest predators.
Dubbed public enemy number one by the Department of Conservation, the wily assassins have taken a frightful toll on threatened species like wrybills, the New Zealand dotterel, black-fronted terns and young kiwi.
For birds that nest in holes and tree trunks - like the mohua, kākā and yellow-crowned kākāriki - a quick visit from a stoat can wipe out eggs, chicks and incubating adults in one attack.
But even heftier birds like a 3kg takahe or a 2kg kakapo can be felled by the slim and sleek little mustelids, which employ a brutal strategy of killing all in sight, and saving the left-overs for later.
They've even been known to swim across water gaps of up to 1.5km - putting them within reach of some of our island conservation strongholds
They were brought here from in the 1880s, in a colossally ill-considered attempt to control rabbits, and quickly set about laying waste to our wealth of native birdlife.
"Many of our most threatened species are now found only on islands or in sanctuaries completely free of stoats - they simply cannot survive where stoats are present," Department of Conservation scientific officer Kerry Weston said.
Most years, stoats were found in low densities and are sparsely distributed.
But that didn't stop them wreaking lethal and disproportionate havoc.
Just two or three that managed to break into Dunedin's fenced Orokonui Ecosanctuary were enough to eliminate a recently established population of saddlebacks.
So cunning were they that our best monitoring techniques weren't sensitive enough to spot them in every situation, such as when they were dispersed in low numbers at certain times of the year.
As it had been long assumed predators weren't a big problem above the treeline, the large focus had been on those lurking in forest ecosystems, where mass seeding events, or masting, triggered pest explosions of plague proportions.
But recent studies had since indicated stoats were in fact common in alpine areas, which wasn't surprising given they were cold-adapted animals - their coats even turned white in winter - and the bulk of their global range was associated with chillier climes.
Researchers were now trying to understand whether periodic seeding within alpine tussock could make stoats more common above the treeline, given early indications that one of their main prey species, mice, became more abundant during these events.
Around our mountains, they pursued the tiny rock wren, takahe, Hutton's shearwaters, lizards and alpine invertebrates like weta.
Detecting them - let alone trapping them - has been a painful headache for scientists like Weston.
In a recent study in Fiordland, she and colleagues looked at different ways to try to keep tabs on stoat numbers in the alpine environment, something that was crucial for working out population changes and how management programmes were faring.
"While we have a good understanding of the threat that introduced mammals pose in forested environments, the impacts of introduced predators in alpine environments are less well understood," she said.
"Climate change may create additional conservation challenges for species above the treeline, if invasive mammals become able to expand their ranges into higher altitudes as temperatures increase over time."
The Fiordland project came about because it was obvious something was driving the decline of the rock wren - and no one had ever designed a study to pinpoint what that was.
Motion-activated cameras set up at 20 rock wren nests in a habitat near the Homer Tunnel soon revealed the culprit - stoats were seen attacking all of them.
To Weston and her team, who weren't picking up any traces of stoats in their specially built, footprint-tracking tunnels, the footage came as a nasty surprise.
"If it wasn't for the cameras we had set up on the nests, we wouldn't have known the stoats were there."
That prompted the researchers to compare how effective these tunnels - the most widely used method for monitoring stoats in New Zealand - were against two fresh approaches, camera traps and artificial nests.
Often used by hunters to detect game animals, camera traps consisted of a motion-activated trail camera, baited with a meat lure, that could spot stoats across kilometre-scale grids.
Artificial nests were much simpler, but perhaps much more innovative - they comprised a fake nest fashioned out of tussock, housing a hen's egg and another fake plasticine egg that was staked into the ground with wire.
When a stoat smelt a hen egg, it stopped by the nest, investigated the fake egg, and left teeth marks behind when it couldn't carry it away.
Interestingly, these nests hadn't been designed for this, but for use as surrogates in nest survival studies.
"We found that in spring and early summer, when rock wrens and other nesting birds are at their most susceptible, cameras and artificial nests could detect stoat activity almost seven weeks earlier than the conventional tracking-tunnels," Weston said.
"Out study indicates that tracking tunnels may not be the most effective way to monitor stoats in all situations, and that artificial nests and camera traps provide an important alternative to the more conventional footprint tracking monitoring method."
The reasons why the tunnels didn't work as well had much to do with stoats' neophobic nature, which caused them to avoid or retreat from an unfamiliar object or situation.
They'd already been shown to actively avoid traps and baited tunnels within their home ranges for considerable periods, especially during their spring breeding season when they become even more cautious, but also when their natural prey were in plentiful supply.
"It is therefore possible that tracking tunnels, which require the stoat to enter them, may evoke neophobic responses and this may be more pronounced in certain habitats or during certain times of the year."
Camera traps and artificial nests could be less likely to cause a neophobic response, which was probably why the team were seeing a higher detection rate with these devices, compared with the tunnels, in spring and early summer.
"It is assumed that artificial nests give the appearance of an abandoned or unattended nest, and therefore, may be less likely to evoke neophobia," Weston said.
"For cameras traps, the approach is also passive, with the lure placed on open ground and representing carrion."
Ultimately, the work suggested camera traps were ideal candidates for situations like predator-free wildlife sanctuaries, where wildlife managers needed to effectively be able to assess pest control and quickly pick up any invaders.
But the fake nests, which still yielded the highest detection rates in spring, when stoat detectability was low, were the most cost-effective option, especially for community groups operating on a limited budget, Weston said.
Stoats are among the "big three" targeted by New Zealand's bold mission to be rid of pest predators by 2050 - the others are rats and possums - and for conservationists, the fight against them couldn't be more pressing.
Climate models show that both beech and tussock mast seeding can be predicted by the difference in summer temperature in the preceding two years.
Consequently, Weston said, we could predict the likely seedfall production next year by examining climate data from 2017 and 2018.
"Climate data indicates that a significant seedfall event is very likely next year - and the extent of the area where seedfall is likely and the strength of the temperature difference is greater than anything since 1972."