Piscina EUVD-2026-90364

Exploitation requires that the host application (1) uses a vulnerable Piscina version (before 4.9.4, 5.3.2, or 6.0.0-rc.5) AND (2) already contains a separate prototype-pollution primitive allowing the attacker to set inherited properties on Object.prototype. The abused options are those WITHOUT an own default on ThreadPool.options - specifically execArgv (preloads attacker code into workers via the Node Worker constructor), loadBalancer (an inherited function executed during task scheduling), and env (alters worker environment). The critical LIMITING factor is that Piscina in isolation is not vulnerable: absent a pre-existing prototype-pollution bug in the application, none of these inherited values can be injected. There is no user interaction required and no authentication to Piscina itself, but the AT:P/AC:H reflect the mandatory secondary primitive.

This is a prototype-pollution 'gadget' rather than a standalone remotely-exploitable flaw. The vendor CVSS 4.0 score of 9.2 (Critical) reflects the worst-case technical impact - code execution in worker threads via an inherited execArgv preload (VC:H/VI:H/VA:H) - but the same vector honestly encodes the key limiter: AC:H and AT:P (Attack Requirements: Present). Both signal that exploitation is NOT self-contained: the description explicitly states it requires 'a separate prototype-pollution primitive' already present in the consuming application. Without that pre-existing primitive, Piscina alone is not exploitable, so the effective real-world risk is materially lower than the 9.2 headline for most deployments. No EPSS score, no CISA KEV listing, and no public POC were provided in the input, so there is no evidence of active exploitation - treat this as 'no public exploit identified at time of analysis.' The PR:N in the vector reflects that no authentication to Piscina itself is needed, but this should not be read as 'trivially remotely exploitable' given the AT:P dependency. The 'Information Disclosure' tag understates the true impact: the primary consequence per the description is arbitrary code preloading (code execution) in worker threads via execArgv and loadBalancer, plus environment tampering via env. Net: a genuine, well-scoped supply-chain hardening fix for apps that also carry a prototype-pollution bug, but not a drop-everything remote 0-day for the average deployment.

An application uses Piscina for its worker pool and also contains a separate prototype-pollution vulnerability (e.g. an unsafe recursive merge of attacker JSON). The attacker pollutes Object.prototype.execArgv (or loadBalancer/env); when Piscina constructs its ThreadPool options as a plain object without an own default for that key, the polluted inherited value flows into the Node.js Worker constructor and preloads attacker-controlled code in a worker thread, yielding code execution. No public exploit was identified at time of analysis.