Skip to main content
1 of 2

It depends on our technology and how long they've been out there.

Our telescopes are getting good

No one who's answered so far considered microlensing effects from the stealth system, but it looks like that gives a detection range of around 550 AU - well within the Oort clouds mentioned in the two current Answers. All-sky surveys of microlensing effects to pick up faint stars, wandering planets, and the like are current research projects. If your stealth system comes within a few hundred AU while one of the all-sky microlensing surveys is active, it will be detected.

Similarly, the James Webb Space Telescope is really good at seeing in infrared - the region of the spectrum which would be hardest to hide a star with current understanding of physics. To decrease the surface temperature of your stealthed star, you'd want to increase the radius of the Dyson sphere - probably to also enclose the habitable planets as well, but that means a larger patch of sky would be anomalously warm - even though it wouldn't be as warm as if there were no sphere. JWST is able to distinguish planets' temperatures from background for many nearby stars, and able to resolve the distances between the planets and their host stars - so the thermal signature from a Dyson sphere might be detectable out to tens of light years. The weakness of JWST is that it's not made to do rapid all-sky surveys, so a fast-moving Dyson sphere or one which happens to be in a direction where JWST doesn't look would be able to pass unnoticed.

How long is the stealth system in the vicinity of ours?

Comets kicked out of the Oort cloud could take centuries to fall close enough to the Sun for us to detect. If the stealthed system is just passing by, we might not notice any effect until the comet rate increases - and by then it might be very difficult to figure out the trajectory of the triggering mass.

...

If you want a "drive-by" interaction, you can probably pass within 1000 AU safely, provided your steal system is moving very quickly compared to the sun's orbital speeds at that distance. There will be lots of disturbance in the Oort cloud, but if folks are invading / surveying Earth during the brief interaction, a Kardashev Type II ought to be able to get up close well before we notice the disturbance. If you want a hidden presence for thousands of years, you're going to have a hard time harding a star within 100 LY that won't be found by JWST, Roman, Hubble, or one of the big survey programs.