GHSA-7Q3F-WX44-378M
Vulnerability from github – Published: 2026-10-01 15:26 – Updated: 2026-10-01 15:26Summary
isPathAllowedForModule decides whether a resolved path belongs to an allowlisted external module using a raw string prefix test. node_modules/foo2 starts with node_modules/foo, so a package whose name merely shares a prefix with an allowlisted one is treated as being inside it, and a relative require from the allowlisted package reaches it even with transitive loading disabled.
Where it is
lib/resolver-compat.js, lines 122 to 132, quoted from HEAD 7a1f5100b96f48d34e0fe104ab37c0acc5944f92:
isPathAllowedForModule(path, mod) {
if (!super.isPathAllowed(path)) return false;
if (mod) {
if (mod.allowTransitive) return true;
if (path.startsWith(mod.path)) {
const rem = path.slice(mod.path.length);
if (!/(?:^|[\\/])node_modules(?:$|[\\/])/.test(rem)) return true;
}
}
return this.externals.some(regex => regex.test(path));
}
With mod.path of .../node_modules/foo and a resolved path of .../node_modules/foo2/index.js, startsWith is true and rem is 2/index.js, which contains no node_modules segment, so the function returns true.
The node_modules test in rem is what stops a genuine transitive dependency from slipping through. It does not stop a sibling, because a sibling's remainder never contains that segment.
Impact
Code running in NodeVM under an external module allowlist with transitive: false can reach a package that was not allowlisted, provided an allowlisted package performs a relative require to a prefix-sharing sibling.
Two preconditions are worth stating plainly rather than leaving implicit. The deployment must already have such a package layout, and an allowlisted package must have a reachable code path that does the relative require. This is not something the attacker creates; it is something they find. That narrows it considerably, and it is why I have not scored it higher.
Reachability
NodeVM.run at lib/nodevm.js:506 executes the script. require comes from createRequireForModule at lib/setup-node-sandbox.js:168-172 and reaches the resolver callback at lib/nodevm.js:380-384. LegacyResolver.resolveFull at lib/resolver-compat.js:145-160 sets currMod for direct requires, the relative specifier resolves through DefaultResolver.resolveFull and tryFile at lib/resolver.js:327-330, and the authorization decision lands on the function above.
Suggested fix
Require a separator after the prefix, so a sibling cannot match:
if (path === mod.path || path.startsWith(mod.path + path.sep)) {
That is the same anchoring the rem regex already applies to node_modules, applied one level earlier.
{
"affected": [
{
"database_specific": {
"last_known_affected_version_range": "\u003c= 3.11.6"
},
"package": {
"ecosystem": "npm",
"name": "vm2"
},
"ranges": [
{
"events": [
{
"introduced": "0"
},
{
"fixed": "3.11.7"
}
],
"type": "ECOSYSTEM"
}
]
}
],
"aliases": [
"CVE-2026-92945"
],
"database_specific": {
"cwe_ids": [
"CWE-22",
"CWE-863"
],
"github_reviewed": true,
"github_reviewed_at": "2026-10-01T15:26:45Z",
"nvd_published_at": null,
"severity": "MODERATE"
},
"details": "## Summary\n\n`isPathAllowedForModule` decides whether a resolved path belongs to an allowlisted external module using a raw string prefix test. `node_modules/foo2` starts with `node_modules/foo`, so a package whose name merely shares a prefix with an allowlisted one is treated as being inside it, and a relative require from the allowlisted package reaches it even with transitive loading disabled.\n\n## Where it is\n\n`lib/resolver-compat.js`, lines 122 to 132, quoted from HEAD `7a1f5100b96f48d34e0fe104ab37c0acc5944f92`:\n\n```js\nisPathAllowedForModule(path, mod) {\n if (!super.isPathAllowed(path)) return false;\n if (mod) {\n if (mod.allowTransitive) return true;\n if (path.startsWith(mod.path)) {\n const rem = path.slice(mod.path.length);\n if (!/(?:^|[\\\\/])node_modules(?:$|[\\\\/])/.test(rem)) return true;\n }\n }\n return this.externals.some(regex =\u003e regex.test(path));\n}\n```\n\nWith `mod.path` of `.../node_modules/foo` and a resolved path of `.../node_modules/foo2/index.js`, `startsWith` is true and `rem` is `2/index.js`, which contains no `node_modules` segment, so the function returns true.\n\nThe `node_modules` test in `rem` is what stops a genuine transitive dependency from slipping through. It does not stop a sibling, because a sibling\u0027s remainder never contains that segment.\n\n## Impact\n\nCode running in `NodeVM` under an external module allowlist with `transitive: false` can reach a package that was not allowlisted, provided an allowlisted package performs a relative require to a prefix-sharing sibling.\n\nTwo preconditions are worth stating plainly rather than leaving implicit. The deployment must already have such a package layout, and an allowlisted package must have a reachable code path that does the relative require. This is not something the attacker creates; it is something they find. That narrows it considerably, and it is why I have not scored it higher.\n\n## Reachability\n\n`NodeVM.run` at `lib/nodevm.js:506` executes the script. `require` comes from `createRequireForModule` at `lib/setup-node-sandbox.js:168-172` and reaches the resolver callback at `lib/nodevm.js:380-384`. `LegacyResolver.resolveFull` at `lib/resolver-compat.js:145-160` sets `currMod` for direct requires, the relative specifier resolves through `DefaultResolver.resolveFull` and `tryFile` at `lib/resolver.js:327-330`, and the authorization decision lands on the function above.\n\n## Suggested fix\n\nRequire a separator after the prefix, so a sibling cannot match:\n\n```js\nif (path === mod.path || path.startsWith(mod.path + path.sep)) {\n```\n\nThat is the same anchoring the `rem` regex already applies to `node_modules`, applied one level earlier.",
"id": "GHSA-7q3f-wx44-378m",
"modified": "2026-10-01T15:26:45Z",
"published": "2026-10-01T15:26:45Z",
"references": [
{
"type": "WEB",
"url": "https://github.com/patriksimek/vm2/security/advisories/GHSA-7q3f-wx44-378m"
},
{
"type": "ADVISORY",
"url": "https://nvd.nist.gov/vuln/detail/CVE-2026-92945"
},
{
"type": "WEB",
"url": "https://github.com/patriksimek/vm2/commit/6ac3916da84e060c403e407b6b6318fcc66b0e72"
},
{
"type": "PACKAGE",
"url": "https://github.com/patriksimek/vm2"
},
{
"type": "WEB",
"url": "https://github.com/patriksimek/vm2/releases/tag/v3.11.7"
},
{
"type": "WEB",
"url": "https://www.vulncheck.com/advisories/vm2-before-3.11.7-module-allowlist-bypass-via-prefix-matching"
}
],
"schema_version": "1.4.0",
"severity": [
{
"score": "CVSS:3.1/AV:N/AC:H/PR:L/UI:N/S:U/C:L/I:L/A:N",
"type": "CVSS_V3"
}
],
"summary": "vm2: External module allowlist uses a raw prefix test, so a prefix-sharing sibling package is treated as allowlisted"
}
Drag nodes to arrange them. Double-click a related vulnerability to open it, or a source to visit its link. The Physics button on the left tunes the layout and the simulation; the legend at the bottom left hides or shows nodes by kind and by sighting type.
Sightings
| Author | Source | Type | Date | Other |
|---|
Nomenclature
- Seen: The vulnerability was mentioned, discussed, or observed by the user.
- Confirmed: The vulnerability has been validated from an analyst's perspective.
- Published Proof of Concept: A public proof of concept is available for this vulnerability.
- Exploited: The vulnerability was observed as exploited by the user who reported the sighting.
- Patched: The vulnerability was observed as successfully patched by the user who reported the sighting.
- Not exploited: The vulnerability was not observed as exploited by the user who reported the sighting.
- Not confirmed: The user expressed doubt about the validity of the vulnerability.
- Not patched: The vulnerability was not observed as successfully patched by the user who reported the sighting.
Vulnerability timeline
Sightings and recorded dates, grouped by day · UTC
How durations are calculated
Following the VTEM framework (Black Hat USA 2015), time to patch starts when the vendor or a coordinator was informed, total time to patch at discovery, and disclosure to patch at public disclosure; vendor response time runs from the report to its acknowledgement. Time to exploit and time unpatched start at the earliest discovery or report, falling back to public disclosure. Unpatched time ends at the first recorded patch, or the calculation time when no patch is recorded. Exploitation to patch shows the interval between those two events; a published proof of concept is a public exploit, not exploitation.
Each result shows its source dates. A negative interval means the end event came first; when date-only inputs leave the order open, the result says so. Date-only inputs are approximate. The earliest evidence wins: KEV report dates and sightings are labelled as such. Records describing another vulnerability are listed but never calculated from, and a PoC alone does not establish exploitation. Missing patch evidence does not prove that no patch exists. Filters below change the event list only.
Loading timeline…
The approach is described in our paper Mapping CVEs to MITRE ATT&CK Techniques: A Curated Gold-Set Classifier and the Limits of LLM-Assisted Label Expansion.
Browse all ATT&CK techniques and the vulnerabilities related to each.
Related by attack behaviour
Vulnerabilities whose description is nearest to this one in the vector space of the CIRCL/vulnerability-attack-technique-biencoder model. This is a similarity search over the bi-encoder space (plain cosine), not a classification, and it has no measured accuracy.