The five consistent superstring theories are UV-complete and consistent quantum theories of gravity; that's demonstrably true. Adding M-theory (the eleven-dimensional supergravity) to that set of maximally uncompactified theories, you can obtain a huge landscape of effective field theories obtained as compactifications of those string/M-theories.
Question: Is the set of all EFTs obtained from string theory a complete list of all the possible EFTs that can be consistently coupled to gravity?
The phrase "string universality" is the statement that the answer to the aforementioned question is affirmative.
String universality examples:
32 supercharges in d=11 and d=10. The only consistent (anomaly-free) supergravities are the eleven-dimensional supergravity, one non-chiral theory (the IIA one) and another chiral (2,0) one. All of them arise as low energy limits of string/M-theories.
16 supercharges in d=10. Anomaly cancellation does only allow four possibilities: $E_{8} \times E_{8}$, $SO(32)$ , $E_{8} \times U(1)^{248}$ and $U(1)^{496}$. Heterotic strings realize the first two ones and amazingly, an anomaly-inflow argument can be invoked to show that the other two options are inconsistent. Again, string theory exactly covers the space of all posibilities.
See String Universality in Six Dimensions and the talk The Swampland in d larger than 6 for amazing dicussions for (2,0) theories in 6d.
I recommend the talk Supersymmetric Quantum Field Theories and the Swampland for more examples that exhibit the amazing match between string theory and all the consistent theories that can be coupled to gravity (specially in six dimensions).
Nowadays the common jargon for "string universality" is the "string lampost principle": The space of all possible EFTs that can be consistently coupled to quantum gravity are realizable as EFTs derived from string theory.