ICFR Sampling IPE Evidence Check

Check a SOX 404 / ICFR control test in three stages, plus a session receipt. First, size an attribute sample from caller-declared confidence level, tolerable deviation rate and expected deviation rate, using the standard zero-EDR Poisson attribute-sampling formula with an expansion factor for nonzero EDR, with deterministic interval sampling over the caller-declared population hash so any auditor can replay the same sample (ART-458). Then verify Information-Produced-by-Entity completeness and accuracy for the control test, reconciling a caller-declared source-extract hash, row count and control total against the same facts for the report built from it (ART-460). Then compose the control-test evidence artifact, reconciling a caller-declared attribute sample against a per-item pass/fail test-result set into one test-conclusion record: sample coverage, exception count against a caller-set tolerable-deviation threshold, and a conclusion of operating effectively, exception noted or incomplete, with reviewer sign-off held as a human gate (ART-461). Finally, a session receipt aggregating the run's execution hashes (CRY-05). Each stage uses its own caller-supplied inputs; no stage's output is an input to another stage, and the order is the testing cycle order: plan the sample, prove the data, then conclude on the test. This is a new chain under its own slug; the existing icfr-control-test-cycle chain, which contains art-461 alone, is left untouched. Findings are arithmetic findings, never audit opinions. Not legal advice.