GHSA-GJJ5-9665-RWRC
Vulnerability from github – Published: 2026-10-02 23:18 – Updated: 2026-10-02 23:18Overview
probe-image-size scans the SVG header with a searching regular expression, /<[-_.:a-zA-Z0-9][^>]*>/. On input that contains many < characters but no >, the engine restarts the [^>]* scan at every < position and runs to end of input each time, giving quadratic time complexity.
Both the synchronous and the streaming parser are affected.
Impact
Every entry point that reaches the SVG parser is affected: probe.sync(), probe(stream) and probe(url). The URL form is the most exposed one — the input is fetched from a remote host, so an attacker only needs to supply a link.
Processing a crafted buffer blocks the Node.js event loop at 100% CPU for the whole duration. In production environments such as upload validators, image proxies or link unfurl services, a small number of concurrent requests is enough to deny service.
Root Cause Analysis
Two independent problems.
-
Absence of input size cap in the sync path.
lib/parse_sync/svg.jscopied the entire buffer into a string and matched against it. There was no size limit at all, so cost scaled with the size of the attacker-supplied buffer. -
Repeated rescanning in the stream path.
lib/parse_stream/svg.jsdid cap accumulated data at 64 KB, but calledparseSvg(str)on the whole accumulated string on every chunk, givingO(chunks × N²). The cap does not help here: the more chunks the input is split into, the more times the quadratic scan is repeated.
Chunk size is influenced by the sender. highWaterMark (16 KB) is a buffering threshold, not a lower bound — a socket read returns whatever has arrived. A server that writes one byte at a time produces one-byte chunks; this was confirmed against the real needle pipeline with default options.
The original report identified (1) only, and stated that the 64 KB cap mitigates the streaming path. It does not.
Proof of Concept (PoC)
Synchronous:
const probe = require('probe-image-size')
// ~200 KB of '<a' — contains '<' but never '>'
probe.sync(Buffer.from('<a'.repeat(100000), 'latin1'))
Streaming — the same payload split into chunks, slower per byte than the synchronous form:
const { Readable } = require('stream')
const probe = require('probe-image-size')
const payload = Buffer.from('<a'.repeat(32768), 'latin1')
const chunks = []
for (let i = 0; i < payload.length; i += 4096) chunks.push(payload.subarray(i, i + 4096))
await probe(Readable.from(chunks))
Measurements on the maintainer's machine:
| path | input | time |
|---|---|---|
probe.sync() |
25 KB | 0.9 s |
probe.sync() |
50 KB | 5.5 s |
probe.sync() |
100 KB | 18 s |
probe.sync() |
200 KB | 54 s |
probe(stream) |
64 KB, 1 chunk | 1.6 s |
probe(stream) |
64 KB, 4 chunks | 2.9 s |
probe(stream) |
64 KB, 16 chunks | 9.6 s |
{
"affected": [
{
"database_specific": {
"last_known_affected_version_range": "\u003c= 7.3.0"
},
"package": {
"ecosystem": "npm",
"name": "probe-image-size"
},
"ranges": [
{
"events": [
{
"introduced": "0"
},
{
"fixed": "7.4.0"
}
],
"type": "ECOSYSTEM"
}
]
}
],
"aliases": [
"CVE-2026-104861"
],
"database_specific": {
"cwe_ids": [
"CWE-1333",
"CWE-400"
],
"github_reviewed": true,
"github_reviewed_at": "2026-10-02T23:18:02Z",
"nvd_published_at": "2026-10-02T18:17:02Z",
"severity": "HIGH"
},
"details": "## Overview\n\n`probe-image-size` scans the SVG header with a searching regular expression, `/\u003c[-_.:a-zA-Z0-9][^\u003e]*\u003e/`. On input that contains many `\u003c` characters but no `\u003e`, the engine restarts the `[^\u003e]*` scan at every `\u003c` position and runs to end of input each time, giving quadratic time complexity.\n\nBoth the synchronous and the streaming parser are affected.\n\n## Impact\n\nEvery entry point that reaches the SVG parser is affected: `probe.sync()`, `probe(stream)` and `probe(url)`. The URL form is the most exposed one \u2014 the input is fetched from a remote host, so an attacker only needs to supply a link.\n\nProcessing a crafted buffer blocks the Node.js event loop at 100% CPU for the whole duration. In production environments such as upload validators, image proxies or link unfurl services, a small number of concurrent requests is enough to deny service.\n\n## Root Cause Analysis\n\nTwo independent problems.\n\n1. **Absence of input size cap in the sync path.** `lib/parse_sync/svg.js` copied the entire buffer into a string and matched against it. There was no size limit at all, so cost scaled with the size of the attacker-supplied buffer.\n\n2. **Repeated rescanning in the stream path.** `lib/parse_stream/svg.js` did cap accumulated data at 64 KB, but called `parseSvg(str)` on the whole accumulated string on *every* chunk, giving `O(chunks \u00d7 N\u00b2)`. The cap does not help here: the more chunks the input is split into, the more times the quadratic scan is repeated.\n\n Chunk size is influenced by the sender. `highWaterMark` (16 KB) is a buffering threshold, not a lower bound \u2014 a socket read returns whatever has arrived. A server that writes one byte at a time produces one-byte chunks; this was confirmed against the real `needle` pipeline with default options.\n\nThe original report identified (1) only, and stated that the 64 KB cap mitigates the streaming path. It does not.\n\n## Proof of Concept (PoC)\n\nSynchronous:\n\n```js\nconst probe = require(\u0027probe-image-size\u0027)\n\n// ~200 KB of \u0027\u003ca\u0027 \u2014 contains \u0027\u003c\u0027 but never \u0027\u003e\u0027\nprobe.sync(Buffer.from(\u0027\u003ca\u0027.repeat(100000), \u0027latin1\u0027))\n```\n\nStreaming \u2014 the same payload split into chunks, slower per byte than the synchronous form:\n\n```js\nconst { Readable } = require(\u0027stream\u0027)\nconst probe = require(\u0027probe-image-size\u0027)\n\nconst payload = Buffer.from(\u0027\u003ca\u0027.repeat(32768), \u0027latin1\u0027)\nconst chunks = []\nfor (let i = 0; i \u003c payload.length; i += 4096) chunks.push(payload.subarray(i, i + 4096))\n\nawait probe(Readable.from(chunks))\n```\n\nMeasurements on the maintainer\u0027s machine:\n\n| path | input | time |\n| --- | --- | --- |\n| `probe.sync()` | 25 KB | 0.9 s |\n| `probe.sync()` | 50 KB | 5.5 s |\n| `probe.sync()` | 100 KB | 18 s |\n| `probe.sync()` | 200 KB | 54 s |\n| `probe(stream)` | 64 KB, 1 chunk | 1.6 s |\n| `probe(stream)` | 64 KB, 4 chunks | 2.9 s |\n| `probe(stream)` | 64 KB, 16 chunks | 9.6 s |",
"id": "GHSA-gjj5-9665-rwrc",
"modified": "2026-10-02T23:18:02Z",
"published": "2026-10-02T23:18:02Z",
"references": [
{
"type": "WEB",
"url": "https://github.com/nodeca/probe-image-size/security/advisories/GHSA-gjj5-9665-rwrc"
},
{
"type": "ADVISORY",
"url": "https://nvd.nist.gov/vuln/detail/CVE-2026-104861"
},
{
"type": "WEB",
"url": "https://github.com/nodeca/probe-image-size/commit/60cc96ac0b671e79e328213d0a8e831312b09e84"
},
{
"type": "WEB",
"url": "https://github.com/nodeca/probe-image-size/commit/9b74656d6f973cc59ea2ab1375c0d88390a402ad"
},
{
"type": "WEB",
"url": "https://github.com/nodeca/probe-image-size/commit/c032aefabdecf5cb50548ab9ba175db56353078f"
},
{
"type": "PACKAGE",
"url": "https://github.com/nodeca/probe-image-size"
},
{
"type": "WEB",
"url": "https://github.com/nodeca/probe-image-size/releases/tag/7.4.0"
}
],
"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": "probe-image-size: Quadratic-time Denial of Service in the SVG Parser"
}
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.