GHSA-QFX6-Q44V-9QQW
Vulnerability from github – Published: 2026-11-13 23:19 – Updated: 2026-11-13 23:19Summary
readPsd() commits the image/layer buffers from the declared header geometry (width x height x channels x depth, and per-layer rectangles) before validating those values against the actual data. A tiny crafted PSD declaring huge geometry forces a multi-gigabyte allocation.
Impact
On ag-psd 30.2.0, Node 24, a 52-byte PSD makes readPsd(buf) commit ~3.17 GB (measured peak RSS) in ~0.45 s before throwing a catchable RangeError ("Offset is outside the bounds of the DataView") — about 61,000,000x amplification. Because the buffer is committed before validation, the allocation OOM-kills the process before the error on essentially any host (uncatchable), and concurrent requests exhaust memory. Any service reading untrusted PSDs with ag-psd is exposed to a remote DoS.
This is distinct from #39 and #206, which involve genuinely large PSDs (memory proportional to real data). Here the input is 52 bytes: the file declares enormous geometry that is never present, so the allocation is driven by an unvalidated header field, not by actual content, an uncontrolled-allocation denial-of-service vulnerability, not proportional memory use.
Steps to reproduce
The PoC is a 52-byte PSD (sha256 a1146a6fabdfef289472f99a56851249e7a46e8d19d4d8bf85fe833e2f0158b6). Reconstruct with:
base64 -d > poc.psd <<'EOF'
OEJQUwABAAAAAAAAAA0AADcuAAA58wAgAAMAAAAAAAAAAAAAAAAAAdEgNZv+i9bc9ygd2A==
EOF
Then (commits ~3.17 GB before throwing a catchable RangeError):
const { readPsd } = require('ag-psd');
const buf = require('fs').readFileSync('poc.psd');
readPsd(buf.buffer.slice(buf.byteOffset, buf.byteOffset + buf.byteLength));
Root cause
psdReader.js sizes the off-heap typed-array buffers from the declared header dimensions (and per-layer rectangles, a second independent lever) before validating against the available bytes.
Suggested fix
Validate the declared dimensions, channel count, depth, and per-layer rectangles against the actual input length (and a sane maximum) before allocating.
{
"affected": [
{
"package": {
"ecosystem": "npm",
"name": "ag-psd"
},
"ranges": [
{
"events": [
{
"introduced": "0"
},
{
"fixed": "31.0.0"
}
],
"type": "ECOSYSTEM"
}
]
}
],
"aliases": [
"CVE-2026-62297"
],
"database_specific": {
"cwe_ids": [
"CWE-789"
],
"github_reviewed": true,
"github_reviewed_at": "2026-11-13T23:19:38Z",
"nvd_published_at": "2026-10-30T21:17:13Z",
"severity": "HIGH"
},
"details": "### Summary\nreadPsd() commits the image/layer buffers from the declared header geometry (width x height x channels x depth, and per-layer rectangles) before validating those values against the actual data. A tiny crafted PSD declaring huge geometry forces a multi-gigabyte allocation.\n\n### Impact\nOn ag-psd 30.2.0, Node 24, a 52-byte PSD makes readPsd(buf) commit ~3.17 GB (measured peak RSS) in ~0.45 s before throwing a catchable RangeError (\"Offset is outside the bounds of the DataView\") \u2014 about 61,000,000x amplification. Because the buffer is committed before validation, the allocation OOM-kills the process before the error on essentially any host (uncatchable), and concurrent requests exhaust memory. Any service reading untrusted PSDs with ag-psd is exposed to a remote DoS.\n\nThis is distinct from #39 and #206, which involve genuinely large PSDs (memory proportional to real data). Here the input is 52 bytes: the file declares enormous geometry that is never present, so the allocation is driven by an unvalidated header field, not by actual content, an uncontrolled-allocation denial-of-service vulnerability, not proportional memory use.\n\n### Steps to reproduce\nThe PoC is a 52-byte PSD (sha256 a1146a6fabdfef289472f99a56851249e7a46e8d19d4d8bf85fe833e2f0158b6). Reconstruct with:\n\n base64 -d \u003e poc.psd \u003c\u003c\u0027EOF\u0027\n OEJQUwABAAAAAAAAAA0AADcuAAA58wAgAAMAAAAAAAAAAAAAAAAAAdEgNZv+i9bc9ygd2A==\n EOF\n\nThen (commits ~3.17 GB before throwing a catchable RangeError):\n\n const { readPsd } = require(\u0027ag-psd\u0027);\n const buf = require(\u0027fs\u0027).readFileSync(\u0027poc.psd\u0027);\n readPsd(buf.buffer.slice(buf.byteOffset, buf.byteOffset + buf.byteLength));\n\n### Root cause\npsdReader.js sizes the off-heap typed-array buffers from the declared header dimensions (and per-layer rectangles, a second independent lever) before validating against the available bytes.\n\n### Suggested fix\nValidate the declared dimensions, channel count, depth, and per-layer rectangles against the actual input length (and a sane maximum) before allocating.",
"id": "GHSA-qfx6-q44v-9qqw",
"modified": "2026-11-13T23:19:38Z",
"published": "2026-11-13T23:19:38Z",
"references": [
{
"type": "WEB",
"url": "https://github.com/Agamnentzar/ag-psd/security/advisories/GHSA-qfx6-q44v-9qqw"
},
{
"type": "WEB",
"url": "https://github.com/Agamnentzar/ag-psd/commit/2d7da68400990bae8ba248a065eb45021fa8000a"
},
{
"type": "PACKAGE",
"url": "https://github.com/Agamnentzar/ag-psd"
},
{
"type": "WEB",
"url": "https://github.com/Agamnentzar/ag-psd/blob/2d7da68400990bae8ba248a065eb45021fa8000a/CHANGELOG.md"
},
{
"type": "WEB",
"url": "https://github.com/Agamnentzar/ag-psd/tree/31.0.0"
},
{
"type": "ADVISORY",
"url": "https://nvd.nist.gov/vuln/detail/cve-2026-62297"
}
],
"schema_version": "1.4.0",
"severity": [
{
"score": "CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H",
"type": "CVSS_V3"
}
],
"summary": "ag-psd can exhaust memory when parsing crafted PSD dimensions"
}
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.
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.