The part has been on the car a while and the welds have been clean. Then a brown or dark halo shows up at one joint, set back from the bead rather than on it, and it returns after every wash. Nothing was done to the car in between. The corrosion did not begin when you first noticed it.
What usually happened is that the joint was borderline when it was made. The weld looked acceptable outside and the part passed a visual check, and the metal survived early service on the chrome still available at the surface. Then a winter or a coastal trip tipped it over.
Stainless resists rust because of chromium, which forms a passive oxide on the surface. That protection depends on chromium being available in the metal rather than locked up in a compound. Weld stainless in the wrong temperature range for too long and carbon and chromium combine at the grain boundaries, leaving the metal around them short of chromium.
The result is a narrow band of depleted metal beside the weld, still stainless by grade and no longer stainless in behaviour. Chlorides find it first. That is why the stain shows as a line parallel to the weld rather than a random patch, and why it appears on the side that took the most heat.
Free iron is the common one. Grinding dust, a carbon steel wire brush, a mild steel bench, a file used on something else: any of these leave particles that rust while the stainless beneath them is untouched. This staining shows as specks or smears over a wide area and comes off with a little effort.
Heat tint is the second. The colour left by welding is not rust and will not turn brown on its own, but it is porous and holds moisture and salt against the metal, so it is a place for corrosion to start rather than a corrosion itself. Chemical attack is the third: an acid or strong alkaline cleaner left to dwell etches the joint and leaves a dull mark with a soft edge.
Location is the first test. Sensitisation sits in a band set back from the fusion line. Free iron sits wherever it landed and pays no attention to the weld. Heat tint is exactly where the arc went. A pinhole leak produces a sooty streak downstream, and that mark grows in the direction of gas flow.
Behaviour is the second. Clean the stain off gently and watch what happens next. Free iron does not return once removed, because its source came away with it. Sensitisation returns, in the same narrow band, within a short period of normal use. That one observation separates the two most common causes without any instrument.
Use a bright light and a magnifier and follow the toe of the weld all the way round, not just across the top. A brown band at a consistent distance from the bead is metallurgical. Rust in pinpricks that straddle bead and parent metal is contamination. Check the underside too, because condensate collects there.
Then check for a leak. With the engine idling from cold, hold a white card close to the joint and move it around the circumference. A dark mark means gas is escaping and carrying soot with it, so the stain is exhaust deposit rather than corrosion. A leak also draws air in on overrun, which overheats the metal beside it.
Surface contamination and heat tint are recoverable. Mechanical cleaning followed by a proper passivation treatment restores the passive layer, provided the cleaning itself deposits no iron. Sensitisation is not recoverable: the chromium is still in the steel but bound up as carbide, and no surface treatment puts it back.
Which makes it a welding decision rather than a cleaning one. Lower heat input, faster travel and an effective purge inside all cut the time the metal spends in the damaging range, and that discipline belongs in the TIG stainless steel welding work rather than the cleaning cupboard. On thin-wall tube the difference is decided before the arc starts, and it shows later in a stainless steel exhaust system.
