invAIriant — evidence-based multi-lens architecture audits for AI-era codebases
Copyright 2026 mindicator & silicon bags quartet

Licensed under the Apache License, Version 2.0 (see LICENSE).

This work distills a body of architecture-audit practice (the audit types, the
0–10 lens scale, the score-to-severity mapping, the anti-averaging rules, and
the network-persistence domain lens) developed while auditing AI-assisted
development of complex, high-load systems.
Lens names (Cormen, Parnas, Turing, Lamport, Leveson, and others) are used as
mnemonic devices for schools of engineering questioning and imply no
endorsement by the individuals named.
