Imagine you built a combination lock where you carefully checked that no single digit alone opens it, but forgot to check whether certain pairs of digits together leak the combination. That is essentially what happened to a compartmented secret sharing scheme published by Yang, Zhu, Fu, and Xia at ISIT 2026, built on the Chinese Remainder Theorem (CRT) for polynomial rings. Ding, Guo, Li, and Man exhibit three distinct families of unauthorized subsets — groups of participants who should not be able to reconstruct the secret — that can in fact reconstruct it completely. The root cause is almost embarrassingly straightforward: the original security analysis failed to account for all published polynomials available to participants. When you hand out shares as polynomial evaluations, every polynomial you publish is a linear equation in the secret. If you forget to count some of those equations, your proof that the system of equations is underdetermined for unauthorized sets is simply wrong. The damage is total in every non-degenerate parameter setting. The only configuration that survives is one where the sole authorized subset is the entire participant set — which is useless as a secret sharing scheme, since the whole point is to allow reconstruction by proper subsets. An "ideal" scheme that works only when everyone participates is not a scheme at all. The paper further notes that the same flaw — incomplete accounting of published information — appears in the same first author's hierarchical secret sharing schemes presented at ISIT 2024 and ISIT 2026. This is not an isolated bug but a recurring pattern in the proof methodology: the security argument undercounts the attacker's available information. From a cryptanalysis standpoint, this is clean, decisive work. There is no ambiguity in the break — three explicit unauthorized-subset families, each fully reconstructing the secret. The proof technique is algebraic and verifiable by anyone with a polynomial ring background. No computational experiments are needed; the attack is information-theoretic. The broader lesson for the secret sharing community is a familiar one: CRT-based constructions for polynomial rings are attractive because they promise ideality (shares the same size as the secret), but achieving ideality under complex access structures like compartmented or hierarchical ones requires the security proof to track every piece of public information. The gap between "intuitively secure" and "provably secure" is exactly where these three families of attacks live. This paper does not propose a fix. It is pure cryptanalysis — a demolition job on a published construction. Whether the CRT-for-polynomial-rings approach can be salvaged for compartmented access structures with lower bounds remains open.
