GHSA-J5CX-PH8G-95V3
Vulnerability from github – Published: 2026-09-17 18:48 – Updated: 2026-09-17 18:48fulgur converts untrusted HTML/CSS into PDF, commonly on a server that
processes input supplied by many tenants. In versions prior to 0.19.0, a
body-direct child whose CSS-resolved height greatly exceeds the page height was
sliced into one fragment per page with no upper bound.
The height is taken directly from attacker-controlled HTML/CSS (height, vh
units), so a few bytes such as:
<div style="height:99999999px"></div>
forced on the order of 125,000 page fragments. The pagination code then
allocates vec![Vec::new(); page_count] and runs a per-page render loop,
resulting in CPU and memory exhaustion. A non-finite height (one that resolves
to +inf) additionally made the slicing loop's remaining -= last_slice_h
decrement never terminate, causing an infinite loop.
An attacker able to submit HTML/CSS to a fulgur-based conversion service can trigger this with a trivially small payload, denying service to the host and any co-tenants.
Patches
Fixed in 0.19.0. A MAX_PAGES cap bounds the slice loop — halting it even
for a +inf height — and non-finite layout heights are sanitized so they can no
longer drive the loop.
Workarounds
Upgrade to 0.19.0 or later. If upgrading is not immediately possible, validate
or constrain untrusted CSS (in particular height / vh on body-level
elements) before passing HTML to fulgur.
Attack Vector rationale
fulgur performs no network I/O of its own; it renders HTML/CSS handed to it by
the embedding application. This advisory scores the crate independent of any
specific adopting program, so per the CVSS v3.1 User Guide §3.7 (scoring library
vulnerabilities related to incoming data) the Attack Vector is assessed as
Network for the reasonable worst-case deployment — a network-facing service
that renders untrusted, attacker-supplied HTML without user interaction. A
concrete system that receives the HTML in one component and passes it to fulgur
in a separate component may assess a lower environmental Attack Vector (Local,
per §3.10).
References
- Fix: https://github.com/fulgur-rs/fulgur/pull/501
{
"affected": [
{
"package": {
"ecosystem": "crates.io",
"name": "fulgur"
},
"ranges": [
{
"events": [
{
"introduced": "0"
},
{
"fixed": "0.19.0"
}
],
"type": "ECOSYSTEM"
}
]
}
],
"aliases": [
"CVE-2026-68523"
],
"database_specific": {
"cwe_ids": [
"CWE-400",
"CWE-835"
],
"github_reviewed": true,
"github_reviewed_at": "2026-09-17T18:48:55Z",
"nvd_published_at": null,
"severity": "HIGH"
},
"details": "`fulgur` converts untrusted HTML/CSS into PDF, commonly on a server that\nprocesses input supplied by many tenants. In versions prior to 0.19.0, a\nbody-direct child whose CSS-resolved height greatly exceeds the page height was\nsliced into one fragment per page with **no upper bound**.\n\nThe height is taken directly from attacker-controlled HTML/CSS (`height`, `vh`\nunits), so a few bytes such as:\n\n```html\n\u003cdiv style=\"height:99999999px\"\u003e\u003c/div\u003e\n```\n\nforced on the order of 125,000 page fragments. The pagination code then\nallocates `vec![Vec::new(); page_count]` and runs a per-page render loop,\nresulting in CPU and memory exhaustion. A non-finite height (one that resolves\nto `+inf`) additionally made the slicing loop\u0027s `remaining -= last_slice_h`\ndecrement never terminate, causing an infinite loop.\n\nAn attacker able to submit HTML/CSS to a fulgur-based conversion service can\ntrigger this with a trivially small payload, denying service to the host and\nany co-tenants.\n\n## Patches\n\nFixed in **0.19.0**. A `MAX_PAGES` cap bounds the slice loop \u2014 halting it even\nfor a `+inf` height \u2014 and non-finite layout heights are sanitized so they can no\nlonger drive the loop.\n\n## Workarounds\n\nUpgrade to 0.19.0 or later. If upgrading is not immediately possible, validate\nor constrain untrusted CSS (in particular `height` / `vh` on body-level\nelements) before passing HTML to fulgur.\n\n## Attack Vector rationale\n\n`fulgur` performs no network I/O of its own; it renders HTML/CSS handed to it by\nthe embedding application. This advisory scores the crate independent of any\nspecific adopting program, so per the CVSS v3.1 User Guide \u00a73.7 (scoring library\nvulnerabilities related to incoming data) the Attack Vector is assessed as\n**Network** for the reasonable worst-case deployment \u2014 a network-facing service\nthat renders untrusted, attacker-supplied HTML without user interaction. A\nconcrete system that receives the HTML in one component and passes it to fulgur\nin a separate component may assess a lower environmental Attack Vector (Local,\nper \u00a73.10).\n\n## References\n\n- Fix: https://github.com/fulgur-rs/fulgur/pull/501",
"id": "GHSA-j5cx-ph8g-95v3",
"modified": "2026-09-17T18:48:55Z",
"published": "2026-09-17T18:48:55Z",
"references": [
{
"type": "WEB",
"url": "https://github.com/fulgur-rs/fulgur/security/advisories/GHSA-j5cx-ph8g-95v3"
},
{
"type": "WEB",
"url": "https://github.com/fulgur-rs/fulgur/pull/501"
},
{
"type": "PACKAGE",
"url": "https://github.com/fulgur-rs/fulgur"
},
{
"type": "WEB",
"url": "https://rustsec.org/advisories/RUSTSEC-2026-0200.html"
}
],
"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": "Fulgur: Unbounded page slicing from attacker-controlled CSS height causes denial of service"
}
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.