Trusted Computing Framework
The Trusted Computing Framework (TCF) was an initiative launched by Vladimir Volokhov et al in the fall of 2145. Volokhov-friendly AIs would derive and cross-check an initial framework, then the framework would be used to reduce errors in the AIs, and those AIs would regenerate the framework, eventually (provably) converging to an error-free stable framework. See the first Chapter 35 snippet for extensive details.
Notably, there is no reason to not have multiple TCFs. Indeed version 2 TacComps use a different TCF. This allows for security through diversity, i.e. if one TCF was broken then the remaining ones might remain. Although it was proven impossible for the TCF to have errors, it would be broken anyways by an unknown group (as of chapter 67). Simona del Mago gave knowledge of the break-in to the MSY, with ((todo: consequences)).