Security context

What an agent needs to avoid regressing past fixes and find the next vuln in this repo.

docling-project/docling

              main @ d2283ad
            

Fixes

CVEs

HIGH

Peak severity

27.3%

Coverage

Highlights

Path Traversal: 3 prior fixes. Scrutinize any change in this area.

docling/backend/html_backend.py: most-fixed (2 issues). Treat as high-risk during review.

8 high-severity fixes in this history; regressions here are high-impact.

Recurring patterns

The bug types that recur here, drawn from past fixes, not open vulnerabilities.

XXE: XML parsers historically enabled external entities and DTD loading, allowing file disclosure and SSRF via crafted XML. Multiple backends share the same insecure default pattern.

Path Traversal: LaTeX macros and dependency staging allowed arbitrary file reads via \input, \include, \includegraphics paths. Fixes add canonicalization and prefix checks, but sibling macro handlers may not enforce all.

SSRF: Remote image fetching and artifact downloads lacked redirect validation, allowing SSRF to private/loopback addresses. Guards are per-component and may not cover all fetch paths.