Skip to content

GHSA-gmc2-2x9w-cgh9 CVE-2026-47683

vm2's bufferAllocLimit cap bypassed by Buffer.concat and Buffer.from arrayLike

High npm
Severity
0.0 / 10
High · CVSS 4.0
Exploit chance (EPSS)
0.54%
Chance of exploitation in the next 30 days. Higher than 45.30% of scored vulnerabilities.
Known exploited
No
Not in CISA's catalog
Weakness
CWE-770 Allocation of Resources Without Limits or Throttling
Fix
3.11.6
Upgrade vm2 to this version or later

Description

Summary

vm2 bufferAllocLimit cap bypassed by Buffer.concat and Buffer.from arrayLike

The bufferAllocLimit option introduced in 3.11.0 (GHSA-6785-pvv7-mvg7) caps host-side Buffer allocations driven by sandbox code, the way embedders opt into timeout. The cap wraps Buffer.alloc, Buffer.allocUnsafe, Buffer.allocUnsafeSlow, and the deprecated Buffer(N) / new Buffer(N) forms. Two other API paths reach the same host C++ allocator with an attacker-controlled size and are not capped: Buffer.concat(list, totalLength) and Buffer.from(arrayLike) with a fake length. Sandbox code can use either to allocate an arbitrary number of host external bytes in a single call, defeating the explicit DoS mitigation the embedder configured.

Details

lib/setup-sandbox.js installs checkBufferAllocLimit at every wrapped entry to host Buffer allocation:

  • alloc() at lib/setup-sandbox.js:474 and the connect(alloc, host.Buffer.alloc) at line 480.
  • allocUnsafe() at line 488 and connect(allocUnsafe, host.Buffer.allocUnsafe) at line 496.
  • allocUnsafeSlow() at line 504 and connect(allocUnsafeSlow, host.Buffer.allocUnsafeSlow) at line 510.
  • BufferHandler.apply at line 424 and BufferHandler.construct at line 433 for the deprecated Buffer(N) / new Buffer(N) numeric-first-arg paths.

Buffer.concat is not wrapped. The sandbox-visible Buffer.concat is therefore the bridge proxy of the host Buffer.concat, which calls into Node's Buffer.allocUnsafe(totalLength) internally without going through the sandbox-side allocUnsafe wrapper. Same for Buffer.from when the argument is array-like ({length: N}): Node's fromArrayLike allocates a buffer of size N before the iteration that fills it. Neither of those allocator paths consult localBufferAllocLimit.

The mitigation rationale documented in docs/ATTACKS.md Category 23 explicitly enumerates the surfaces that were considered and either capped (Buffer.alloc family) or punted to follow-up (new Uint8Array(N), new ArrayBuffer(N), String.prototype.repeat). Buffer.concat(list, totalLength) is not listed in either group, and Buffer.from(arrayLike) is mentioned only as "bounded by source array size which had to be allocated through some other path first" -- which is not true for the {length: N} form, because no array of length N actually exists.

A single call from sandbox to Buffer.concat([Buffer.from('a')], 50 * 1024 * 1024) allocates 50 MiB of host external memory. The allocation itself is a single synchronous host C++ call that timeout cannot interrupt, exactly like the original advisory. The zero-fill that follows is interruptible, but the memory is already committed by the time the interrupt could fire, so the embedder's container memory budget is the only ceiling. The same pattern in a loop, or with a larger totalLength, drives RSS up by hundreds of megabytes per call.

The fix uses the existing checkBufferAllocLimit(size) helper and a sandbox-side wrapper installed via connect(...) -- one for Buffer.concat that sums the totalLength (or falls back to summing list lengths) and one for Buffer.from that recognises the array-like-with-numeric-length branch.

PoC

'use strict';
const { VM, NodeVM } = require('vm2');

function ext() { return Math.round(process.memoryUsage().external / 1024 / 1024); }
function tryBypass(label, code) {
    const ext0 = ext();
    let buf;
    try { buf = code(); }
    catch (e) {
        console.log(`[${label}] CAPPED -- ${String(e).split('\n')[0]}`);
        return;
    }
    console.log(`[${label}] BYPASSED -- got ${buf && buf.length} bytes (external +${ext() - ext0} MB)`);
}

console.log('Cap is configured at 1024 bytes.\n');

const vm1 = new VM({ bufferAllocLimit: 1024 });
tryBypass('VM Buffer.alloc(50MB)        ',
    () => vm1.run('Buffer.alloc(50 * 1024 * 1024)'));

const vm2 = new VM({ bufferAllocLimit: 1024 });
tryBypass('VM Buffer.concat 50MB        ',
    () => vm2.run('Buffer.concat([Buffer.from("a")], 50 * 1024 * 1024)'));

const vm3 = new NodeVM({ bufferAllocLimit: 1024 });
tryBypass('NodeVM Buffer.concat 50MB    ',
    () => vm3.run('module.exports = Buffer.concat([Buffer.from("a")], 50 * 1024 * 1024);'));

const vm4 = new VM({ bufferAllocLimit: 1024 });
tryBypass('VM Buffer.from({length: 8MB})',
    () => vm4.run('Buffer.from({length: 8 * 1024 * 1024})'));

Run with node poc.js against vm2@3.11.3:

Cap is configured at 1024 bytes.

[VM Buffer.alloc(50MB)        ] CAPPED -- RangeError: Buffer allocation size 52428800 exceeds bufferAllocLimit 1024
[VM Buffer.concat 50MB        ] BYPASSED -- got 52428800 bytes (external +50 MB)
[NodeVM Buffer.concat 50MB    ] BYPASSED -- got 52428800 bytes (external +50 MB)
[VM Buffer.from({length: 8MB})] BYPASSED -- got 8388608 bytes (external +8 MB)

Process RSS climbs by the same amount each call, confirming a real host C++ allocation rather than a sandbox-realm-only effect.

Impact

This is the same DoS class GHSA-6785-pvv7-mvg7 was filed for: untrusted sandbox code amplifying a small payload into a large synchronous host external-memory allocation that V8's timeout cannot preempt. In the environments the advisory cites -- Docker memory limits, Kubernetes pods, AWS Lambda -- a single 200-byte sandbox payload can drive a multi-hundred-megabyte RSS jump and OOM the host process.

The Category 23 fix was specifically scoped to "cap host Buffer external allocation" and embedders are documented to opt into bufferAllocLimit as their layered defense against this class. The two paths above are uncapped, so an embedder that has configured bufferAllocLimit: 32 * 1024 * 1024 (the value recommended in the README's Hardening recommendations) is still vulnerable to the exact attack the option was designed to prevent. The mitigation invariant -- "every Buffer external allocation driven by sandbox code is capped by bufferAllocLimit" -- does not hold.

No sandbox escape; pure DoS.

Affected versions

npm

PackageVulnerableFixed in
vm2
< 3.11.6
3.11.6

How to fix

Move to the fixed version or later with your package manager, update the lockfile, and run your tests. If the package only arrives through another dependency, upgrade that dependency first. If none of its releases pulls a fixed version yet, force the version with your package manager's override.

vm2 npm → 3.11.6

npm
npm install vm2@3.11.6
yarn
yarn add vm2@3.11.6
pnpm
pnpm add vm2@3.11.6

Severity breakdown

CVSS 4.0 0.0

CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:N/VC:N/VI:N/VA:H/SC:N/SI:N/SA:N

Attack vector
Network
Attack complexity
Low
Attack requirements
None
Privileges required
None
User interaction
None
Confidentiality (vulnerable system)
None
Integrity (vulnerable system)
None
Availability (vulnerable system)
High
Confidentiality (subsequent systems)
None
Integrity (subsequent systems)
None
Availability (subsequent systems)
None

Weakness

CWE-770: Allocation of Resources Without Limits or Throttling

The product allocates a reusable resource or group of resources on behalf of an actor without imposing any intended restrictions on the size or number of resources that can be allocated.

How to prevent it
  • Requirements: Clearly specify the minimum and maximum expectations for capabilities, and dictate which behaviors are acceptable when resource allocation reaches limits.
  • Architecture and Design: Limit the amount of resources that are accessible to unprivileged users. Set per-user limits for resources. Allow the system administrator to define these limits. Be careful to avoid CWE-410.
  • Architecture and Design: Design throttling mechanisms into the system architecture. The best protection is to limit the amount of resources that an unauthorized user can cause to be expended. A strong authentication and access control model will help prevent such attacks from occurring in the first place, and it will help the administrator to identify who is committing the abuse. The login application should be protected against DoS attacks as much as possible. Limiting the database access, perhaps by caching result sets, can help minimize the resources expended. To further limit the potential for a DoS attack, consider tracking the rate of requests received from users and blocking requests that exceed a defined rate threshold.
  • Implementation: Assume all input is malicious. Use an "accept known good" input validation strategy, i.e., use a list of acceptable inputs that strictly conform to specifications. Reject any input that does not strictly conform to specifications, or transform it into something that does. When performing input validation, consider all potentially relevant properties, including length, type of input, the full range of acceptable values, missing or extra inputs, syntax, consistency across related fields, and conformance to business rules. As an example of business rule logic, "boat" may be syntactically valid because it only contains alphanumeric characters, but it is not valid if the input is only expected to contain colors such as "red" or "blue." Do not rely exclusively on looking for malicious or malformed inputs. This is likely to miss at least one undesirable input, especially if the code's environment changes. This can give attackers enough room to bypass the intended validation. However, denylists can be useful for detecting potential attacks or determining which inputs are so malformed that they should be rejected outright.
  • Architecture and Design: For any security checks that are performed on the client side, ensure that these checks are duplicated on the server side, in order to avoid CWE-602. Attackers can bypass the client-side checks by modifying values after the checks have been performed, or by changing the client to remove the client-side checks entirely. Then, these modified values would be submitted to the server.
  • Architecture and Design: Mitigation of resource exhaustion attacks requires that the target system either: recognizes the attack and denies that user further access for a given amount of time, typically by using increasing time delays uniformly throttles all requests in order to make it more difficult to consume resources more quickly than they can again be freed. The first of these solutions is an issue in itself though, since it may allow attackers to prevent the use of the system by a particular valid user. If the attacker impersonates the valid user, they may be able to prevent the user from accessing the server in question. The second solution can be difficult to effectively institute -- and even when properly done, it does not provide a full solution. It simply requires more resources on the part of the attacker.
  • Architecture and Design: Ensure that protocols have specific limits of scale placed on them.
  • Architecture and Design, Implementation: If the program must fail, ensure that it fails gracefully (fails closed). There may be a temptation to simply let the program fail poorly in cases such as low memory conditions, but an attacker may be able to assert control before the software has fully exited. Alternately, an uncontrolled failure could cause cascading problems with other downstream components; for example, the program could send a signal to a downstream process so the process immediately knows that a problem has occurred and has a better chance of recovery. Ensure that all failures in resource allocation place the system into a safe posture.
  • Operation, Architecture and Design: Use quotas or other resource-limiting settings provided by the operating system or environment. For example, when managing system resources in POSIX, setrlimit() can be used to set limits for certain types of resources, and getrlimit() can determine how many resources are available. However, these functions are not available on all operating systems. When the current levels get close to the maximum that is defined for the application (see CWE-770), then limit the allocation of further resources to privileged users; alternately, begin releasing resources for less-privileged users. While this mitigation may protect the system from attack, it will not necessarily stop attackers from adversely impacting other users. Ensure that the application performs the appropriate error checks and error handling in case resources become unavailable (CWE-703).

References

Timeline

  1. Published
  2. Last updated by the source
  3. Latest EPSS score
AdvisorySeverityWhy it's related
GHSA-489w-w794-jq94
vm2: NodeVM zlib Buffers expose pooled host memory across the VM boundary
High Also affects vm2
GHSA-2v2p-6j97-cjg9
vm2: Host Promise rejection from an exposed constructor can terminate the vm2 host process
Medium Also affects vm2
GHSA-5h3f-q97h-ccvc
vm2: NodeVM custom resolution bypasses external path boundaries
Critical Also affects vm2
GHSA-pq68-rvw4-xp4r
vm2 contains a sandbox escape vulnerability
Critical Also affects vm2
GHSA-f8gf-w286-fmq2
vm2: `allowAsync: false` can be bypassed through Promise thenable assimilation in VM and NodeVM
High Also affects vm2
GHSA-6rh5-qq4q-97xh
vm2: NodeVM builtin denylist bypass via fs/promises despite -fs, allowing host filesystem writes
High Also affects vm2
GHSA-8686-vhfx-7r3j
vm2: NodeVM node:-prefixed negative builtin deny bypass exposes child_process
Critical Also affects vm2
GHSA-wjwh-qqvp-g4p4
vm2 sandbox escape via WebAssembly.compileStreaming Promise species bypass
Critical Also affects vm2
Is this in your code?

MergeHarbor checks every repository's dependencies against this database, ranks what matters by exploitability and reachability, routes it to the owning team and opens the fix pull request.

Check your repos with MergeHarbor