Executive brief
Paperclip, a platform for managing issues and agent workflows, is vulnerable to a security flaw where malicious links can be embedded in documents and comments. If a user clicks one of these links, an attacker can execute code in the user's browser to steal their session and take over their account. This affects almost all text areas in the application, including chat threads and approvals.
Technical details
A stored cross-site scripting (XSS) vulnerability exists in the 'MarkdownBody' component of Paperclip. The component explicitly overrides the 'react-markdown' library's 'urlTransform' property with an identity function, effectively disabling the built-in protection against 'javascript:', 'vbscript:', and 'data:' URI schemes. Because the backend persists Markdown bodies for issues and comments without content validation, an authenticated attacker can inject malicious payloads. When a victim clicks the resulting link, the script executes within the application's origin. The impact is exacerbated by the lack of a global Content Security Policy (CSP) and the reuse of the vulnerable component across multiple UI surfaces.
Affected products
- paperclipai paperclip < 2026.416.0
Timeline
- 2026-04-16: advisory: GHSA-fpw4-p57j-hqmq published
- 2026-04-16: patched: Fixed in version 2026.416.0