CVE-2026-55618: GOVCERT-LU eml_parser URL extraction bypass via HTML entities

Executive brief

eml_parser is a library used by email security systems and security operations centers to extract URLs from email messages for threat analysis. The vulnerability allows attackers to craft emails with URLs that use HTML entity encoding (e.g., `https://phishing.com`) to evade detection, causing malicious URLs to be missed by security scanning systems that rely on this library.

Technical details

The vulnerability is a validation bypass caused by improper encoding/escaping (CWE-116). eml_parser validates URL strings to filter out bogus values, but this validation occurs before HTML entity decoding. URLs using HTML entity encoding for critical characters (`:`, `/`, `.`) are incorrectly rejected, preventing both the URL and its host from appearing in the extraction results. For example, `https://phishing.example.com` decodes to `https://phishing.example.com` but is rejected before decoding. The fix in version 3.0.2 unescape HTML entities before validation. The attack requires user interaction (opening an email) but no authentication or special privileges, and can be delivered across the network via standard email protocols.

Affected products

• GOVCERT-LU eml_parser < 3.0.2

Timeline

• 2026-06-15: disclosed: Published to GitHub Advisory Database
• 2026-08-25: patched: Version 3.0.2 released with fix

References

• https://api.github.com/users/Sebasteuo
• https://github.com/Sebasteuo
• https://api.github.com/users/Sebasteuo/gists%7B/gist_id%7D
• https://api.github.com/users/Sebasteuo/repos
• https://avatars.githubusercontent.com/u/37353257?v=4
• https://api.github.com/users/Sebasteuo/events%7B/privacy%7D