What happens if you try building stealth solutions in theories where the coupling function changes sign?

Your question touches on a genuinely subtle theoretical problem, but I need to be direct: the preprints provided don't address "stealth solutions" in coupling-sign-reversal theories.

That said, the archive does illuminate why sign-flipped couplings are trickier than they first appear.

When a coupling function changes sign—say, from attractive to repulsive—you're inverting the potential landscape. Naively, this breaks stability: an upside-down bowl has no minimum. But quantum mechanics can rescue classical instability. Negative coupling scalar field theory [AG-2026.03-1028], for instance, has no classically stable ground state yet exhibits stable and unitary quantum evolution, potentially offering an interacting continuum description of the Higgs that evades triviality.

However, sign reversals introduce a second danger: superluminality and chronology violation. When higher-dimensional operators carry negative coefficients, they can produce faster-than-light signal propagation on certain backgrounds, which in turn seeds closed timelike curves—time machines [AG-2024.06-694]. This looks catastrophic. But quantum backreaction from low-energy modes sabotages the classical geometry before paradoxes crystallize, potentially invalidating the threat without forbidding the operators outright.

A third complication emerges from coupling-constant anomalies: topological obstructions in the space of bare parameters themselves [AG-2024.04-797]. These can prevent entire families of gapped phases from existing, forcing theories into gapless or symmetry-broken regimes—a constraint that tightens further when you modify couplings.

For a concrete "stealth" scenario (perhaps a theory whose light content remains hidden at low energy despite sign-flipped couplings), you'd need preprints specifically addressing screening, decoupling, or emergent weakly-coupled sectors in such theories—which aren't in this list.