UNKNOWN is not PASS.

The most dangerous mistake in a research audit is not a FAIL; it is treating uncertainty as a PASS.

Keep the three states distinct

PASSmeans that the defined verification has passed. FAILmeans that the verification contract has been violated. UNKNOWNmeans that it cannot be confirmed with the current evidence. UNKNOWN does not mean that the strategy is wrong, nor does it mean that it is safe.

Why record conservatively?

For example, a PASS in the current status file does not prove that an independent SQL re-audit also passed. Likewise, a correction report is not enough if fault injection or end-to-end verification is still incomplete; the correct state is “correction reported, re-audit not completed.”

The cost of this principle

This can make progress look slower, but it prevents uncertainty from being quietly erased once the results become attractive. Research documentation is not a success story; it defines the boundary of what the evidence actually supports.