Summary
A Server-Side Request Forgery (SSRF) vulnerability in SoftVision webPDF's PDF converter allows attacker-controlled documents to trigger outbound requests using protocols such as http:// and file:///.
When a malicious XML or HTML file is uploaded and rendered to PDF, the server may access internal network services and local files, enabling internal port scanning and Local File Inclusion (LFI).
Affected context
This vulnerability affects deployments of SoftVision webPDF (prior to version 10.0.2) that:
- Accept user-uploaded XML, HTML, or similar rich content for conversion to PDF
- Allow the converter engine to automatically resolve embedded URLs, images, or external entities
- Run the conversion process with network access to internal services and read access to local files
The issue becomes more severe when the conversion service is deployed within a trusted network segment or on hosts with sensitive local data.
Exploit preconditions
- Attack surface: The webPDF file upload and PDF conversion endpoint.
- Required access: Ability to upload and submit XML/HTML (or similar) content for PDF conversion; this may be anonymous or authenticated depending on deployment.
- Trigger point: Rendering of the attacker-supplied document, which contains references using http://, file:///, or other supported URI schemes.
- Why exploitation is possible: The converter does not distinguish between internal and external resources and follows attacker-controlled URIs, allowing the server to act as a proxy to internal hosts and local files.
Why this happens
The root cause aligns with CWE-918: Server-Side Request Forgery (SSRF).
The PDF converter trusts and automatically dereferences resource references (e.g., image sources, external entities, embedded URLs) from user-supplied documents without validating destination schemes, hostnames, or network locations.
Because the conversion logic runs server-side with broader network and filesystem access than the attacker has directly, the application effectively exposes its internal network and local filesystem as an oracle to remote users.
This insecure pattern stems from treating untrusted document content as safe to resolve, instead of enforcing strict allowlists and access controls for outbound requests.
Potential impact
- Internal port scanning of services not exposed to the public internet
- Discovery of internal service endpoints and basic network topology
- Local File Inclusion (LFI) where contents of readable files may be embedded in the generated PDF
- Exposure of configuration files, credentials, or keys present on the server filesystem
- Pivoting opportunities for further attacks against internal systems once services are identified
- Increased blast radius if the conversion service has high-privilege filesystem or network access
General mitigation guidance
Effective mitigation focuses on restricting what the PDF converter can access and how it processes untrusted content:
- Apply strict URL and protocol allowlists, limiting resolved resources to known-safe, external domains and approved schemes only (e.g., disallow file:///, internal IP ranges, and metadata endpoints).
- Disable or heavily constrain automatic external resource fetching during document rendering where business requirements permit.
- Introduce network-level egress controls (firewalls, egress proxies) to prevent the converter from reaching internal-only networks or sensitive infrastructure.
- Run the conversion service under least-privilege accounts with minimal filesystem access, isolating it in a dedicated host, container, or sandbox with constrained network routes.
- Normalize and treat uploaded documents as untrusted input, avoiding features that implicitly resolve external entities or embedded resources unless explicitly needed.
- Use defense-in-depth: combine application-level checks (URL validation, scheme restrictions) with infrastructure controls so a single misconfiguration does not reintroduce full SSRF risk.
How teams detect issues like this
Identifying and monitoring SSRF risks in document conversion pipelines can involve:
- Static analysis rules or code scanning to locate:
- Unrestricted HTTP client usage in conversion or rendering code
- Support for file:// or other local-resource schemes driven by user-supplied content
- Lack of hostname/IP validation or network boundary checks before outbound requests
- Configuration review of PDF/rendering engines to ensure features that auto-fetch external entities, images, or URLs are either disabled or tightly scoped.
- Security-focused code review aimed at:
- Any functionality that dereferences URLs from uploaded content
- Network utility libraries used by rendering components
- Error handling that may surface internal endpoint details back in responses or PDFs
- Runtime monitoring and logging:
- Centralized logging of outbound requests from the conversion service, with alerts for connections to internal IP ranges, loopback (127.0.0.1), or unexpected ports.
- Network IDS/IPS or firewall logs to detect unusual egress patterns from the conversion host.
- Optional automated tools can be configured in CI/CD pipelines to flag high-risk URL handling and SSRF patterns before deployment.

