Executive brief
Penpot is an open-source design and prototyping platform used by teams to collaborate on visual designs. A flaw in how custom font names are processed allows an attacker to inject malicious code that executes when team members open any file referencing that font. The attack is passive—no user interaction required—and could lead to theft of session credentials, unauthorized actions on behalf of victims, and unauthorized access to their files and projects.
Technical details
The vulnerability is a stored cross-site scripting (XSS) flaw in the font family naming functionality. The backend accepts arbitrary font family names in the create-font-variant and update-font RPC endpoints with no validation, and the frontend interpolates the user-controlled font name directly into a @font-face CSS rule, then injects it via innerHTML without sanitization. An attacker can craft a font name containing markup such as </style><script> to break out of the style element and execute arbitrary JavaScript on the Penpot origin. The attack is passive: any team member who opens a file referencing the malicious font triggers script execution simply by rendering the page. The vulnerability requires low privileges (ability to create or rename fonts within a team) and affects versions up to and including 2.14.3. The fix, released in version 2.15.3, implements a whitelist validator for font family names that restricts characters to letters, digits, spaces, hyphens, underscores, and dots.
Affected products
- Penpot Penpot up to and including 2.14.3
Timeline
- 2026-08-26: disclosed
- 2026-05-14: patched: Fix committed on 2026-05-14; patch released in version 2.15.3