GHSA-H7WH-Q2QR-458C
Vulnerability from github – Published: 2026-11-13 18:09 – Updated: 2026-11-13 18:09Summary
Rails::HTML::SafeListSanitizer (both the HTML4 and HTML5 variants) shares one Rails::HTML::PermitScrubber across every sanitize call on the same instance, assigning its tags and attributes on each call without synchronization. Two concurrent sanitize calls with different tags: or attributes: options can interleave, so one call is scrubbed with the other call's allowlist.
- Versions affected:
< 1.7.2 - Not affected: none
- Fixed versions:
1.7.2
Impact
Action View memoizes a single SafeListSanitizer per helper module and uses it for every sanitize call in the process, so in a multi-threaded server (for example Puma) all requests share the same PermitScrubber. If any code path calls sanitize with a more permissive allowlist, an attacker who can trigger that path concurrently with another request can cause the other request to be sanitized with the permissive allowlist. If the permissive allowlist admits script or an event-handler attribute, the strict call emits it unchanged, which is a cross-site scripting vector when the strict output is stored or rendered to other users.
Applications are affected only if they run more than one thread per process and some sanitize call passes tags: or attributes: options that differ from another call's, including a call that passes no options and so uses the defaults. Applications that use a single configuration everywhere, pass their own scrubber:, or run single-threaded are not affected.
Workarounds
Use the same allowlist for every sanitize call, or pass a scrubber: built per call (Rails::HTML::PermitScrubber.new with tags= and attributes= set) instead of tags: and attributes: options.
Credit
Reported by hiumee via HackerOne.
{
"affected": [
{
"package": {
"ecosystem": "RubyGems",
"name": "rails-html-sanitizer"
},
"ranges": [
{
"events": [
{
"introduced": "0"
},
{
"fixed": "1.7.2"
}
],
"type": "ECOSYSTEM"
}
]
}
],
"aliases": [
"CVE-2026-128069"
],
"database_specific": {
"cwe_ids": [
"CWE-362",
"CWE-79"
],
"github_reviewed": true,
"github_reviewed_at": "2026-11-13T18:09:16Z",
"nvd_published_at": null,
"severity": "MODERATE"
},
"details": "## Summary\n\n`Rails::HTML::SafeListSanitizer` (both the HTML4 and HTML5 variants) shares one `Rails::HTML::PermitScrubber` across every `sanitize` call on the same instance, assigning its `tags` and `attributes` on each call without synchronization. Two concurrent `sanitize` calls with different `tags:` or `attributes:` options can interleave, so one call is scrubbed with the other call\u0027s allowlist.\n\n- Versions affected: `\u003c 1.7.2`\n- Not affected: none\n- Fixed versions: `1.7.2`\n\n## Impact\n\nAction View memoizes a single `SafeListSanitizer` per helper module and uses it for every `sanitize` call in the process, so in a multi-threaded server (for example Puma) all requests share the same `PermitScrubber`. If any code path calls `sanitize` with a more permissive allowlist, an attacker who can trigger that path concurrently with another request can cause the other request to be sanitized with the permissive allowlist. If the permissive allowlist admits `script` or an event-handler attribute, the strict call emits it unchanged, which is a cross-site scripting vector when the strict output is stored or rendered to other users.\n\nApplications are affected only if they run more than one thread per process and some `sanitize` call passes `tags:` or `attributes:` options that differ from another call\u0027s, including a call that passes no options and so uses the defaults. Applications that use a single configuration everywhere, pass their own `scrubber:`, or run single-threaded are not affected.\n\n## Workarounds\n\nUse the same allowlist for every `sanitize` call, or pass a `scrubber:` built per call (`Rails::HTML::PermitScrubber.new` with `tags=` and `attributes=` set) instead of `tags:` and `attributes:` options.\n\n## Credit\n\nReported by hiumee via HackerOne.",
"id": "GHSA-h7wh-q2qr-458c",
"modified": "2026-11-13T18:09:16Z",
"published": "2026-11-13T18:09:16Z",
"references": [
{
"type": "WEB",
"url": "https://github.com/rails/rails-html-sanitizer/security/advisories/GHSA-h7wh-q2qr-458c"
},
{
"type": "WEB",
"url": "https://github.com/rails/rails-html-sanitizer/issues/228"
},
{
"type": "WEB",
"url": "https://github.com/rails/rails-html-sanitizer/commit/15f764c3563d5dd9b97ad32eaa7454577ccec4b9"
},
{
"type": "PACKAGE",
"url": "https://github.com/rails/rails-html-sanitizer"
},
{
"type": "WEB",
"url": "https://github.com/rails/rails-html-sanitizer/releases/tag/v1.7.2"
},
{
"type": "WEB",
"url": "https://github.com/rubysec/ruby-advisory-db/blob/master/gems/rails-html-sanitizer/GHSA-h7wh-q2qr-458c.yml"
}
],
"schema_version": "1.4.0",
"severity": [
{
"score": "CVSS:3.1/AV:N/AC:H/PR:N/UI:R/S:C/C:L/I:L/A:N",
"type": "CVSS_V3"
}
],
"summary": "rails-html-sanitizer: Possible XSS via shared scrubber state in concurrent calls"
}
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.