feat(ownts): real-world FP benchmark + срезано 17/28 ложных срабатываний #908

Job	Run time
corpus benchmark (real C# recall + specificity)	2m 46s
tests (py3.13)	6s
own-check SARIF -> GitHub code scanning (dog-food)	36s
own-check repo scan (github + msbuild) + composite action	45s
C# leak extractor (Roslyn) -> OwnIR -> core	41s
lint (ruff + mypy --strict)	13s
audit aggregation selftests	7s
tests (py3.11)	9s
golden C# compiles & runs (.NET)	19s
tests (py3.12)	6s
extended codegen fuzz	36s
P-014 Tier B — external reference resolution (--ref-dir)	29s
OwnTS (React useEffect) -> OwnIR -> core	6s
	6m 59s

Navigation Menu