An example of a planetary mass object or planemo that is almost airless is the Moon.
It does have an atmosphere, but very, very thin.
The Moon has an atmosphere so tenuous as to be nearly vacuum, with a total mass of less than 10 tonnes (9.8 long tons; 11 short tons). The surface pressure of this small mass is around 3 × 10−15 atm (0.3 nPa); it varies with the lunar day. Its sources include outgassing and sputtering, a product of the bombardment of lunar soil by solar wind ions.
So the atmosphere of the Moon is almost zero, and a spacecraft can orbit the Moon just above the highest peak in its path without significant air resistance. Air resistance will probably take millions of years to affect the orbit of an object in orbit around the Moon.
But no object in space is perfectly spherical. All space objects rotate faster or slower, and so all space objects are at least slighty oblate, wider at their rotational equators. So if a spacecraft orbits a world in a non equatorial orbit, it will be at least a little closer to the ground over the equator than over higher latitudes. So the planet's gravitational attraction will increase and decrease slightly over the course of an orbit.
If the spacecraft is in orbit around the equator of a planemo it can stay the same distance from the surface, if its orbit is perfectly circular. And of course a perfectly circular orbit is impossible; all orbits are at least slightly ellipticall and thus change their distances from the center of hte object at least slightly.
And of course the Sun and the Moon produce tides on Earth which raise and lower the levels of the sea, and even the levels of dry land. So even if a satellite passes over a location on the Earth's surface at the same distance from the center of the Earth every time the distanc eto the surface will be slightly different each time. And the tidal forces of the Sun and the Moon will also pull on the satellite and change its orbit.
And of course the Moon also experiences tides from the Sun and Earth, the Earth tides being many times stronger on the Moon than the Moon tides on Earth. So the lunar surface will rise and fall with the tides, and the Sun and Earth tides will slowly tug a lunar satellite out of its original orbit.
And different types of materials tend to have different densities, and different types of materials tend to be concentrated in different regions on an object in space. Thus a satellite of an object will pass over regions of greater density, called mass concentrations or mascons, and be attracted stronger by them, and will pass over regions of lesser density and will be less attracted by them. That will change its orbit.
And early in the space program it was discovered that the Moon has very strong lunar mascons which will distort the orbit of a spacecraft more and more will every orbit. It is fine to orbit the Moon a few times over a few days. but if a spacecraft has to orbit the moon for months or years or decades the masscons will perturb the orbit more and more over time.
So a lunar satellite has to carry enough propellant to adjust its orbit from time to time to stay in the proper orbit until its mission is completed and it now longer has to stay in the proper orbit.
The Moon is the most gravitationally "lumpy" major body known in the solar system. Its largest mascons can cause a plumb bob to hang about a third of a degree off vertical, pointing toward the mascon, and increase the force of gravity by one-half percent.2
Typical examples of mascon basins on the Moon are the Imbrium, Serenitatis, Crisium and Orientale impact basins, all of which exhibit significant topographic depressions and positive gravitational anomalies. Examples of mascon basins on Mars include the Argyre, Isidis, and Utopia basins. Theoretical considerations imply that a topographic low in isostatic equilibrium would exhibit a slight negative gravitational anomaly. Thus, the positive gravitational anomalies associated with these impact basins indicate that some form of positive density anomaly must exist within the crust or upper mantle that is currently supported by the lithosphere. One possibility is that these anomalies are due to dense mare basaltic lavas, which might reach up to 6 kilometers in thickness for the Moon. While these lavas certainly contribute to the observed gravitational anomalies, uplift of the crust-mantle interface is also required to account for their magnitude. Indeed, some mascon basins on the Moon do not appear to be associated with any signs of volcanic activity. Theoretical considerations in either case indicate that all the lunar mascons are super-isostatic (that is, supported above their isostatic positions). The huge expanse of mare basaltic volcanism associated with Oceanus Procellarum does not possess a positive gravitational anomaly.
Lunar mascons alter the local gravity above and around them sufficiently that low and uncorrected satellite orbits around the Moon are unstable on a timescale of months or years. The small perturbations in the orbits accumulate and eventually distort the orbit enough that the satellite impacts the surface.
Because of its mascons, the Moon has only four "frozen orbit" inclination zones where a lunar satellite can stay in a low orbit indefinitely. Lunar subsatellites were released on two of the last three Apollo manned lunar landing missions in 1971 and 1972; the subsatellite PFS-2 released from Apollo 16 was expected to stay in orbit for one and a half years, but lasted only 35 days before crashing into the lunar surface. It was only in 2001 that the mascons were mapped and the frozen orbits were discovered.2
The Luna-10 orbiter was the first artificial object to orbit the Moon and it returned tracking data indicating that the lunar gravitational field caused larger than expected perturbations presumably due to 'roughness' of the lunar gravitational field. The Lunar mascons were discovered by Paul M. Muller and William L. Sjogren of the NASA Jet Propulsion Laboratory (JPL) in 1968 from a new analytic method applied to the highly precise navigation data from the unmanned pre-Apollo Lunar Orbiter spacecraft. This discovery observed the consistent 1:1 correlation between very large positive gravity anomalies and depressed circular basins on the Moon. This fact places key limits on models attempting to follow the history of the Moon's geological development and explain the current lunar internal structures.
So if someone wants a lunar satellite to have a very low but long lasting orbit, they should choose one of the four "frozen orbits" which don't pass over any mascons. Presumably they should choose the orbit with the lowest terrain measured from the center of the Moon. With the extremely thin "atmosphere" of the Moon, an orbit passing only slightly over the highest point of ground on the orbit might last for thousands or millions of years.
Of course there are mascons in many other worlds in the solar system. The Moon has the strongest mascons of any major object in the solar system. But the mascons on other major solar system objects will have to be calculated for when putting an object into the lowest possible orbit around them.
There are thousands and millions of minor solar system objects, small moons, asteroids, comets, etc. which are not massive enough to pull their matter into spheriodal shape. Thus they have somewhat irregular shapes. And that should make choosing the proper orbit to orbit one at the lowest possible altitude rather difficult.
The more irregular a minor solar system body is, the more it will be like having strong mascons to perturb the orbit of any satellite. And many minor solar system bodies are believed to have formed by smaller bodies gently coming together, leaving large interior voids between the pieces. A small object with interior voids seems like it would have the same effect on saellites as very strong masscons.
So an airless object with a very homogenous composition and very minor mascons, and made of very light and weak material with a very low topography, floating in interstellar or intergalactic space, would seem to be the ideal object to have a very low satellite.