Confidence & unknown
Confidence describes how strongly the evidence establishes the primitive and its parameters.Unknown is an honest classification when evidence is incomplete.
Why this matters: A wrong classification is worse than no classification. Teams that trust inflated confidence make bad migration decisions.
Confidence levels
Unknown
Unrecognized primitives, unparseable files, and missing parameters must not be silently dropped. Prefer:- Emit
Unknownwith evidence attached - Surface a warning for unparseable files
- Keep findings-so-far in the inventory