Comparatively speaking, asteroids and/or ice ages being more frequent than expected seems more plausible.
(The alternative is extreme plasticity, allowing lifeforms to adapt to otherwise hostile atmospheres...)
We meet that threshold, though. We are extremely adaptive to otherwise hostile environments by virtue of technology, particularly wearable one - from winter jackets to SCUBA gear.
There used to be iron deposits etc. on the surface. They were recovered by pre-industrial societies. Now we are digging deeper and deeper for resources, but once we lose that ability, it's possible we won't be able to regain it.
What do you do once the nuclear power plant's core has melted, the mine is full of water, and there is nobody around who knows what these things are even there for because all the people who did died years ago of old age. You can't read it up either once we've moved our stored knowledge from paper to machines in what the ancient texts refer to as the mythical "US-EAST-2 region".
Of those, it feels like the only irreconcilable one is scale: e.g. something you need so much of, repeatedly, that no less-efficient substitute is viable.
And of those, geologically-created, carbon-based energy sources (coal, oil) are the primaries that come to mind.
(Also, nitrogen, but I imagine the Haber–Bosch process would survive)
Also, the iron was only an example. Coal and other fossil fuels we can only burn once, then they are gone.
A species at a similar development level in some other solar system with slightly better societal goals or a need for a specific resource could already be colonizing other planets or asteroids.
Any process which requires scaling intelligence from competitively local and tribal (or any equivalents) to integrated and planetary will do.
You either pass that test or you fail it. If you fail it, stagnation and eventually devolution beckon.
What was critical in parent's comment was the irrecoverability of the civilization, after such an action. Scaling intelligence is more of a repeatable "trial and error" step (as the article puts it, in a biology context).