CVE-2026-72703: Rocq Prover 8.20 before 9.2.0 Guard Checker Accepts Non-Terminating Fixpoint via Unchecked Cross-Calls
The guard checker in Rocq Prover treats a parameter of a nested mutual fixpoint as uniform without examining calls between the different bodies of that fixpoint. finduniformparameters in kernel/inductive.ml inspects only self-recursive calls, so when no body calls itself the function concludes that every parameter is uniform. A parameter that grows through a cross-call from one body to another therefore keeps the subterm specification it inherited from the enclosing fixpoint, and a recursive call guarded by that specification is accepted although the argument is not structurally smaller. A non-terminating definition is admitted as structurally decreasing, which yields a term whose value equals its own successor and so a proof of False, from which any proposition follows. The proof requires no axioms, plugins or unsafe flags and Print Assumptions reports it as closed under the global context. Introduced in Coq 8.20 and fixed in Rocq 9.2.0.
Affected Software
Remediation
Recommended actions to resolve this vulnerability, in priority order.
- Upgrade
Upgrade
Rocq Proverto a version that resolves this vulnerability.Fixed in 9.2.0 - Upgrade
Upgrade
Coqto a version that resolves this vulnerability.Fixed in 8.20
Event History
Frequently Asked Questions
Which installations are affected?
The issue was introduced in Coq 8.20 and affects Rocq Prover releases before 9.2.0. Rocq 9.2.0 contains the fix.
What must an attacker or untrusted proof author be able to do?
They need to get a crafted proof term or definition checked by the affected Rocq kernel, typically by having a target compile or validate untrusted Rocq source. Exploitation requires no axioms, plugins, or unsafe flags, although it does require user interaction according to the supplied CVSS vector.
Are default safety checks sufficient to prevent exploitation?
No. The malformed nested mutual fixpoint is accepted as structurally decreasing by the normal guard checker, without enabling unsafe options or adding axioms.
What is the impact of accepting a malicious definition?
A non-terminating definition can be admitted and used to construct a proof of False. Any proposition can then be derived, undermining the soundness of proofs checked by the affected kernel.
Can Print Assumptions identify proofs exploiting this issue?
No. The supplied proof can appear closed under the global context in Print Assumptions, despite relying on the guard-checker flaw.