MediumVulnerabilityLLM-specific

CVE-2026-27740: Discourse cross-site scripting in Review Queue via AI LLM output

Source: NVD(opens in a new tab)
Identifier: CVE-2026-27740
Published: March 19, 2026
Record updated: October 9, 2026

Known exploitation: Not listed in the CISA Known Exploited Vulnerabilities catalog at the last check.
EPSS: 0.3%

Summary

Discourse versions before 2026.3.0-latest.1, 2026.2.1, and 2026.1.2 contain a cross-site scripting flaw in the Review Queue. The platform renders raw output from an AI LLM with htmlSafe without adequate sanitization. An attacker can use prompt injection to make the AI return a malicious payload, which executes when a Staff member (Admin or Moderator) views the flagged post.

Mitigation

Fixed in 2026.3.0-latest.1, 2026.2.1, and 2026.1.2. As a workaround, temporarily disable AI triage automation scripts.

Details

Scores and exploitation

EPSS: 0.3%Estimated probability that this CVE is exploited in the next 30 days, from the FIRST Exploit Prediction Scoring System. Read on November 4, 2026.
Disclosure date: March 19, 2026

Weaknesses and techniques

Weakness (CWE): CWE-79
Attack pattern (CAPEC): CAPEC-198
CAPEC-86
AI threat technique (MITRE ATLAS): AML.T0051

Classification

Attack type: Prompt injection,
Other
Attack sophistication: Moderate
AI component targeted: Plugin
Impact: Integrity,
Confidentiality
Labels: security,
safety

How this record was produced

Summaries, mitigations and labels on this page are generated automatically from the source. The linked source is authoritative. The source gives this advisory a description and no title, so the title above is a headline generated from the source text.

Original source: https://nvd.nist.gov/vuln/detail/CVE-2026-27740

First tracked: March 19, 2026 at 06:07 PM ET, from NVD/CVE Database

Classifier: claude-haiku-5-5, prompt v4

Classifier confidence: 85% (self-assessed)

Summary

Mitigation

Related items