allows code supplied to the interpreter to shadow built‑in guard hooks via positional‑only arguments. This bypasses the embedding application's access policy, permitting execution of arbitrary code within the trusted sandbox. The flaw arises from the check_function_argument_names() function failing to reject forbidden hook names when they appear as positional‑only arguments, a weakness identified as CWE‑184. Consequently, any user input that defines such functions can override.

Affected versions are all releases of the Zope Foundation RestrictedPython product older than 8.3, including 8.2 and earlier deployed in a variety of web applications and data‑processing pipelines that rely on RestrictedPython for sandboxing third‑party code. The problem is not confined to a particular operating system; any environment executing RestrictedPython may be impacted.

The CVSS score is 8.3, indicating high severity. The EPSS score is < 1%, so the current estimated exploit likelihood is very low, but the absence of a KEV listing means no widespread exploitation has been reported to CISA yet. Based on the description, the attack vector is inferred to be local, requiring an adversary to supply crafted Python code to the RestrictedPython interpreter. If the application accepts untrusted input, the threat is significant, as the attacker can bypass the sandbox and run native or arbitrary code within the host process.