CVE-2026-54025: LibreChat is an enhanced ChatGPT clone that supports multiple AI providers. Prior to 0.8.4-rc1, there is a…

Summary

LibreChat versions before 0.8.4-rc1 are affected by a flaw in the markdown artifact preview pipeline. When the custom image renderer in generateMarkdownHtml (client/src/utils/markdown.ts) falls back to marked v15.0.12's built-in renderer, double-quote characters in image alt text are not HTML-escaped. An attacker can craft alt text that breaks out of the alt attribute and injects an event handler, which runs in the victim's browser when the HTML is assigned to innerHTML in the Sandpack preview iframe.

Mitigation

Fixed in 0.8.4-rc1.

Related items

• MediumRecent AI agent experiments reveal a risky future of machine 'teamwork' inside corporationsSimilar attack · CNBC Technology
• MediumClaude Code's “Don't Ask Again” File Exposes Passwords and Keys in 1,122 Public Copies on GitHubSimilar attack · The Hacker News
• MediumCVE-2026-118046: JS Help Desk plugin stored DOM-based cross-site scripting via jsticket_messageSimilar attack · NVD/CVE Database
• MediumARTEX abuse expands as cybercriminals target financial sectorSimilar attack · Microsoft Security Blog
• MediumCVE-2026-102610: fast-agent serve unauthenticated remote command execution via /mcp endpointSimilar attack · NVD/CVE Database